See also also bike pic for other bike-related technical discussions and more failed parts.

See also also The Bicycle Museum of Bad Ideas for more wAcKy bicycle designs.

Centerless/Hubless Wheels

TL;DR:

• The advantage of centerless/hubless wheels is: they look nice.
• Centerless/hubless wheels have more drag, cost more, weigh more, and fail faster.
• Every few years, somebody introduces a new version and uses big words to say they fixed all the problems. They did not.
• Here is some analysis of why it is hard to build good centerless/hubless wheels.

           
From [https://www.bikeradar.com/features/the-weirdest-bike-on-ebay-right-now-15000-of-retro-aero-tech as of 2025-03] [https://uncrate.com/reevo-hubless-e-bike as of 2025-03] [https://mikeshouts.com/reevo-hubless-electric-bicycle-indiegogo as of 2025-03] [https://www.designboom.com/technology/reevo-hubless-electric-bike-09-23-2020 as of 2025-03]

The idea of centerless/hubless bicycle wheels has been around for over 100 years [Thorpe]. Yet every few years it seems somebody proposes them as a “new” idea.

I know of two centerless/hubless wheels that made it to production and were sold to the general public:

• In 1993, the company Wear and Tear started seling the “Black Hole” centerless/hubless front wheel for racing.

  Production was under 100 wheels. After some racing wins, the UCI banned it, and production stopped.

  Online comments said it was loud, and the rim support rollers wore much faster than ordinary hub wheel bearings. Both would limit its appeal outside of racing.

• Around 2022, the brand Reevo made a production e-bike with front and rear centerless/hubless wheels.

  Reevo claimed over 1,800 bikes shipped to customers. Others say it was more like 100-200.

  Online comments say it was unreliable. Also loud, slow, heavy, and uncomfortable. But poor reliability may make the other problems much less important. Shipments stopped by 2024.

Some folks call these “spokeless” wheels, but a disk wheel is also spokeless. So I use “centerless/hubless” here.

One view is that there is no such thing as a “hubless” wheel: somebody enlarged a hub to almost the size of the wheel, then used a very large hollow axle. Put that way, one question is: does a very large hub have advantages over a standard-size hub?

Centerless/hubless wheels are:

• An idea that shows up often, over many decades.
• A particularly bad idea. At least for most common uses.
• Often hyped as “ground-breaking” or “innovative”, when they are not.

Given that, it seems like a (bad) idea worth discussing.

What about them is bad? Compared to ordinary wheels with hubs, centerless/hubless wheels:

• have more drag/friction, so they slow you down;
• cost more, weigh more, and wear out faster;
• have no way to mount ordinary disk brakes; and
• have few benefits to “pay” for all their problems.

It is okay if something has problems. For example, pneumatic tires get flats, but are used widely because never-flat airless tires have more drag and less comfort. A “good idea” can have problems, but should have benefits. That is, at least some riders find the benefits are a win, despite the problems.

What are the benefits of centerless/hubless wheels? Or, at least what are the claimed benefits? Some claims I have seen are they:

• look cool;
• weigh less because there are no spokes;
• can carry cargo in the middle of the wheel;
• can hold a folded folding bike, so it is more compact when folded;
• have no air drag from moving spokes, so less air drag overall.

Sections below:

• Problems and claimed benefits
• Problems
• Claimed beneifts
• Examples: Products, Prototypes, Proposals
• Background
• Summary

Problems and claimed benefits

• Problems
• Claimed beneifts

Problems

• Drag/Friction
• Cost, weight, and durability
• You cannot use standard disk brakes
• Summary of Problems

More on claimed benefits is below. First, though, more detail on problems: when we discuss benefits, it is important to know how big are the benefits, compared to how big are the problems. Consider two wheels, both are lighter, but both have more drag:

• If a wheel is 1 gram lighter but has 100x the drag, it is a lose.
• If a wheel half the weight and has 0.001% more drag, it is a win.

Some problems with centerless/hublesss wheels:

Drag/Friction

The tire and rim need to turn, so they need a bearing. A bearing has some drag. When you coast, it takes force at the tire/road to overcome the bearing drag — that is a part of “rolling drag”, one of the things that slows you down.

Many hubs use 6001 bearings with an outside diameter of 28 mm. An MTB tire is about 560 mm, or about 20 times larger. When the tire rolls, the turning force at the tire has about 20:1 leverage on the hub bearing. In other words, because there is leverage, drag at the tire — what slows you down — is about 1/20 of the bearing’s drag.

For example, in the following diagram if F_tension is at half the radius of F_friction, then F_friction is also half as big as F_tension. (This is just a random picture from the internet, it is not really designed to show bearing drag. Until I find or make a better picture, you need to guess a little to match the picture to the meaning of the text above.)

  
[https://torquenitup.weebly.com/free-body-diagram-involving-torque.html as of 2025-03-05]

A centerless/hubless wheel also needs to turn, so also has a bearing. However, the bearing is almost the same diameter as the tire, so the tire only has low leverage to overcome bearing drag. For simplicity, I use “no leverage” or 1:1 in this dicussion. A real centerless/hubless wheel might be 1.3:1 or 1.6:1, etc., but the results and conclusions are similar to 1:1.

Suppose hub-type bearings and centerless/hubless wheel bearings both have similar bearing drag. The centerless/hubless wheel gives up 20:1 leverage, so drag at the tire is 20x higher. In other words, if hub-type wheels have 1 Watt of bearing drag, then centerless/hubless wheels have about 20 Watts of bearing drag — with similar bearing types, and used in similar ways.

Worse, parts in the hubless bearing run at about road speed. That is, if you ride along and put your finger on the outer race of the bearing on the hubless/centerless wheel, the bearing race will rub your finger at about the same speed as if you put your finger on the tire, or put your finger on the road. In a hub-type bearing, the outer race runs at about 1/20 of road speed. “Moving faster” can increase bearing drag (that is: force). For example, moving grease or oil faster tends to increase the force needed to move it.

Even worse, when a bearing is 20x as big, the seals are also 20x as big. A seal spans between the rotating outer race and non-rotating inner race. The seal touches both, in order to seal. That means it rubs on one of them, so has drag. A 20x bigger seal has 20x more area that rubs. All else equal, that means 20x more drag. You have the same “leverage” problem above, but also 20x more seal drag. So the seal drag force at the tire is 20 × 20 = 400x compared to a 6001's seal drag. Yes, seal drag for a 6001 is small. But make it 400 times larger and it could be big

The Reevo uses a bunch of little wheels to support the rim, rather than a conventional bearing. However, it is still a “bearing”, and has the same problems. With a bearing in a hub, you get 20:1 leverage; with the same bearing in a little wheel out at the rim, you lose most of that leverage. If you then use 20x little wheels, you have seal drag for all 20 of them.

              
Screen captures from [https://www.youtube.com/watch?v=MgPUpccQ_mw as of 2025-03]

With conventional wheels, drag from hub bearings is only a minor part of the rider’s effort — maybe 1% of your energy goes to hub drag. But if bearing drag gets 20x larger, then 20% of your energy goes just to wheel bearing drag! That means you have about 20% less left over for moving through the air and climbing hills. So you ride a lot slower.

Drag problems are also hard to solve. Especially since we also want our bearings to be both cheap and durable. Cheap bearings tend to be less durable — and not so cheap if you have to keep replacing them. Seals help durability but add drag. Removing seals removes that drag, but riding in rain/dirt/etc. without seals lets junk in the bearing. Junk can hurt drag even before it damages the bearing. And then it also damages the bearing.

To summarize: it is hard to make a centerless/hubless wheel which turns with low drag. Low-drag bearings are possible, but tend to be more expensive than hub bearings, and can have poor durability in ordinary riding. Adding seals can help durability, but can add a lot of drag. For e-bikes, added drag is less of a problem, but can hurt pedal-only riding and also e-bike range.

Cost, weight, and durability

Are Reevo’s wheels just built badly? Or are they bad for some inherent hard-to-fix reason? It would be wrong to take a “built badly” problem and use it to claim all centerless/hubless wheels have the problem.

I claim: centerless/hubless wheels have a bunch of problems which are inherent — always hard to get right.

Below is a discussion of some basic design problems for hubless/centerless wheels. Drag is also a basic design issue, but is so important that I covered it separately above.

Also, high volume manufacturing usually improves cost, weight, and durability. The following considers the best case for centerless/hubless wheels: assume they are in high-volume mass production, with no more development problems. It just considers problems which remain, even in mass production.

• They put the bearings closer to the dirt.

  Hub bearings are far above the trail or road, so they stay fairly clean. A centerless/hubless wheel puts the bearings close to the ground, where they are easy to splash with mud and dirt, and easy to submerge just by rolling the wheel through a deep puddle.

  Even if the bearings have seals, “more dirt exposure” is a problem: more mud and dirt on the outside of the seals means the seals wear faster. Once seals wear enough, mud, dirty water, and other stuff gets inside. Then friction goes up and durability goes down. More: [http://pardo.net/bike/pic/fail-036/index.html as of 2025-03].

• Larger bearings are ... larger.

  Larger bearings have more metal (cost and weight), and take more work to make (cost). Instead of a conventional bearing, Reevo has the rim rotate on 17 rollers. But it has the same problems: there are many rollers, each with bearings. That is more weight and cost than a hub, since a hub has just a few parts and uses just two bearings.

• Centerless/hubless wheels carry loads indirectly.

  A beam or frame is usually lightest when it carries a load in a straight line. Going around corners or holding the load off to one side (“cantilevered”) tends to weigh more to carry the same load.

  Spokes, hubs and forks carry loads in a fairly straight line. That helps make a wheel light compared to the load it carries. A fork is also fairly light. In a centerless/hubless wheel like the Reevo, the wheel is supported by a “support hoop” that carries wheel loads in an indirect path — since a straight line would go through the center. The indirect path makes the centerless wheel heavier.

  Some centerless/hubless designs, such as the [Terpstra], have a structural rim and grab just part of it with rollers. Combining the support structure and rim can have advantages [Wheel structure types], but also problems:

  • A support hoop distributes the load; the Reevo has many rollers to support the rim, but each roller is light. Grabbing a structural rim only one place means the rest of the rim has high leverage on the “grab”. So the grab rollers and the structure to support them needs to be stronger — and thus heavier.

  • Forces on the rollers are higher, and higher forces increase drag. Both for the bearings in each roller, and also where the roller rolls on the rim.

  • A hub-type wheel carries loads directly through the center of the wheel. A structural rim carries load around the edge, whcih is an indirect load path. An indirect path adds weight.

• The centerless/hubless wheel fork is shorter, but not much lighter.

  A fork for a hubless wheel is shorter than a standard fork, and may be marginally lighter. But it only removes the lightest part of a conventional fork, so the savings is small. The rear wheel and frame are similar. Centerless/hubless wheels are heavier due to their indirect load path, larger bearings, etc.; so add more weight than is saved in the fork.

• Driving a centerless/hubless wheel often adds friction and/or weight.

  It is simple to drive a hub-type rear wheel: put a sprocket on the hub, and run a chain or belt to it from a sprocket on the cranks.

  Driving a centerless/hubless wheel is more difficult, due to driving at a larger size.

  • An MTB rim is around 560 mm diameter. A sprocket the same size has about 135 teeth. To go as fast as a 23-tooth sprocket, you either have to pedal 123/23 = 6 times as fast, or step up the chain speed. A chainring 6 times longer would hit the ground, so an extra step-up drive or gearbox gets used. That adds drag and weight compared to a plain chain or belt drive.

  • Chains and belts need to clear the tire, so are offset. For a centerless/hubless wheel, part of the sprocket and chain/belt runs close to the ground. If you ride close to things like rocks, curbs, etc., the chain and sprocket can hit them and get damaged. Sprockets can be built more robustly to withstand strikes, but that adds weight. A normal chain or belt is also offset to clear the tire, but it is up high and so clears most things — although long-cage derailleurs hang down and sometimes get damaged.

  • A big sprocket also needs a longer chain — about 3x as long, and 3x the weight. Belts weigh less, so the penalty is smaller. But it is still added weight.

  • Another choice is a rim-size gear, with a “spur” gear driving it. A rim-size gear with the teeth pointing in is called a “ring gear”. Gears can be located inside the wheel, making them harder to damage than a sprocket next to the wheel. Gears typically need to be more precise than sprockets, thus more expensive. There still needs to be something to reach around the tire, so a multi-stage drive is needed. Addig stages hurts drag, weight, and cost.

  • Driving at the rim means part of the drive is close to the ground. In turn, it is exposed to more dust, mud, and dirty water. Chains, belts, and gears can be shielded so they are harder to get dirty. However, large seals have a lot of drag. Removing the seal cuts drag, but then simply rolling the wheel through a deep puddle can cover it in water and dirt, speeding up wear of the drive.

          
  From [https://odditymall.com/diy-hubless-fat-tire-bicycle as of 2025-03]               
  Screen captures from [https://www.youtube.com/watch?v=MgPUpccQ_mw as of 2025-03]

Even in high-volume production, centerless/hubless wheels have the problems above.

Worse, even if a maker can solve some problems, just one significant problem can still be a deal-breaker: centerless/hubless wheels have few advantages to make it worth suffering with even one serious problem. You can often improve one thing at the expense of another, but it is harder to improve everything all at once.

Sometimes people argue that problems can be solved if you spend more. If you spend more on centerless/hubless wheels, you should compare to a spend-more conventional bike [A fair comparison].

To summarize: centerless/hubless wheel designs have problems which hurt cost/weight/durability comapred to conventional hub-type wheels. These are fundamental issues which are hard to solve for any centerless/hubless wheel. Beware of “fixes” that improve one thing at the expense of another: you may have one wheel which is low-drag and another wheel which is durable; yet it may be hard to make to one wheel which is both low-drag and durable.

You cannot use standard disk brakes

“Disk brakes” might be included in cost/weight/durability, but I called them out separately:

• Disk brakes for bicycles took a long time to develop. Disk brakes for centerless/hubless wheels will be at least somewhat different. Figuring out disk brakes for centerless/hubless wheels may be a big project on its own.

• At the same time, many riders are happy enough with rim brakes, and it is not important whether disk brakes are available. If you are happy with rim brakes, you do not care about disk brake issues.

The Reevo uses rim brakes. The Reevo’s brakes are bad, but there are good rim brakes. The Reevo’s rim wanders side-to-side, which makes it hard to fit a good rim brake. But there are centerless/hubless wheel designs with less wander.

Motorcycles sometimes mount brake rotors to the rim of a hub-type wheel, and it is possible to mount a disk brake to the rim of a standard or centerless bike wheel.

           
From [https://s1.cdn.autoevolution.com/images/news/a-motorcycle-braking-guide-part-3-46159_1.jpg as of 2025-03], [https://www.tuningblog.eu/en/E-Scooter-ATV-Quad/hubless-e-bike-top-secret-497637 as of 2025-03], [https://www.gadgetreview.com/21-game-changing-bikes-paving-the-way-for-a-greener-future as of 2025-03,] and [https://www.topsecretebike.com as of 2025-03]

But:

• A centerless/hubless wheel needs a very big brake rotor. That adds weight.

  A large rotor also has more leverage and better cooling (more surface area), which may allow use of brake designs that do not work with hub-mounted disk brakes. For example, an aluminum 560 mm rotor might give acceptable service but weigh less than a 203 mm steel rotor.

  This suggeests light centerless/hubless disk brakes are possible. But: finding out what is possible is additional development on top of other centerless/hubless work.

• Putting the brake rotor close to the road makes it easier to splash with dirt, mud, water, etc. In turn, this can make braking less predictable. One reason people like disk brakes is they are predictable in the mud and wet.

• A dirty rotor wears faster than a clean one, adding cost — both parts and labor to do the replacement.

• Existing disk brake calipers demand the rotor has very little side-to-side motion (lateral play). Otherwise, the brake rubs — it makes annoying sounds and slows you down. Floating rotors can tolerate more lateral play without much drag; but they still make rubbing sounds, which some riders find annoying.

  Some centerless/hubless wheels have a lot of lateral play — [Reevo], for example. Others have much less lateral play, but may cost a lot more, or have other down-sides. It may be hard to get all needed attributees at the same time.

To summarize rim-mounted brakes: you cannot use conventional disk brakes with the rotor mounted to the rim. It may be possible to develop good disk brakes to attach to the rims of hubless/centerless wheels, but is yet another project, in addition to all the other centerless/hubless wheel problems.

           
From [https://odditymall.com/diy-hubless-fat-tire-bicycle as of 2025-03]

An alternative is a disk brake mounted on the frame, then using chain or belt drive to connect it to the wheel.

• On the rear, you already have a drive, so adding a brake does not add cost/weight for a chain or belt drive.

• One caveat is that if the drive chain falls off, you also lose the brake. Hub mounted brakes can also fail, but “one more way to fail” could be an issue.

• On the front, there is not already a drive chain or belt, and adding one just for the brake may add quite a bit of weight.

  A front “brake-only” drivetrain always moves with the wheel, since you cannot freewheel the brake. Moving the drivetrain, even with no load, adds some drag. It may be enough to be significant.

  (The rear wheel’s drivetrain is normally moving anyway. So on the rear, “move the drivetrain” friction counts as pedaing drag, not brake drag.)

To summarize brakes:

• You cannot use standard disk brakes mounted to the rim of a centerless/hubless wheel.

• On the rear, you can use a standard brake which is separate from the wheel, and connected by the drivetrain. Be aware that this adds a new way for the brake to fail.

  On the front, a separate brake adds substantial weight and maybe also substantial drag.

• It may be possible to develop disk brakes for centerless/hubless wheels which are competitive in cost/weight/etc. with conventional hub-mounted disk brakes. However, that is an additional project beyond all the other projects/tasks for centerless/hubless wheels.

• If you are happy with rim brakes, issues with disk brakes are not a “problem”.

Summary of Problems

To summarize the main points above: centerless/hubless wheels introduce several hard-to-solve problems:

• more drag/friction, so they slow you down;
• they cost more, weigh more, and wear out faster;
• and they have no way to mount ordinary disk brakes.

Worse, solving one or a few of these problems does not help, if big problems remain. For example, even if you get down the cost and weight, something which still slows you down a lot will be unpopular.

At least, it will be unpopular unless it has some really big advantage over conentional wheels. That brings us to:

Claimed beneifts

Why would you want to use centerless/hubless wheels? That is, what improvements do they offer over ordinary hub-type wheels? And even where there is an improvement, is it big enough so it is worth dealing with their disadvantages?

Just to set up the discussion, here are are some other “problem vs. benefit” bike examples, where something has problems, but we use it because it has benefits that outweight those problems:

• Pneumatic tires get flats (problem!) and also cost more and wear out faster than airless tires (more problems!). But pneumatic tires “pay” for their problems by making the bike a lot faster and more comfortable.

• Derailleur gearing has more friction than a single-speed chain or belt drive; and derailleurs cost and weigh more and wear out faster. They “pay” for these problems by making the bike faster, more versatile, and more comfortable.

For both of these, the advantages are big enough advantages. If pneumatic tires made you 1% faster, the advantage would be too small for you to be interested in the down-sides.

That said, even though pneumatic tires and derailleurs have big advantages for many riders, they are not always used. Many hire bikes use airless tires. Many riders prefer single-speed chain or belt drive, so they can avoid some of the problems with derailleurs.

For most riders and riding, centerless/hubless wheels seem to lack benefits. Below are things I have seen suggested as advantages. One is just wrong. The others seem like real advantages, but the down-sides (drag, cost, weight, durability) mean that for most riders and most riding, ordinary hub-type wheels are the better choice. That is, the disadvantages of hubless/centerless wheels outweigh the advantages.

• Centerless wheels look cool

  But how much will most riders pay for art? People build and enjoy riding chopper/low-rider bikes which are expensive and are slower than regular bikes. A few people would ride centerless/hubless wheels for art and despite the problems. Most riders? No.

• Centerless wheels weigh less because there are no spokes

  This claim is obviously wrong: spokes carry your weight, so if you get rid of the spokes, you need something else to carry your weight. That “something else” adds back at least some of what you saved by removing spokes.

  In other words, it can be lighter if you can replace spokes with something lighter. Really: spokes and everything they imply: getting rid of spokes means you cannot use nice light low-drag hub bearings. But the wheel still needs to turn, so getting rid of spokes means you have to replace the bearings, too. Are the replacements just as light?

  So it is not lighter just because “it has no spokes”. Indeed, it seems many centerless/hublesss wheels (a) have no spokes; and (b) are much heavier than comparable hub-type wheels.

  If you think about this claim a bit, you can come up with a longer list of things which are a useful part of the wheel and so need to be replaced: axle, bearings, hub shell. You can get rid of XYZ, but XYZ was there for a reason. So whatever weight is saved by getting rid of XYZ turns in to weight added back for new parts needed to do XYZ’s job. When you add it all up, what is lighter?

  Maybe you can sometimes build a centerless/hubless wheel which is lighter than a comparable hub-type wheel. Great! But if so, it is way more complicated and tricky than “it has no spokes”.

  It seems notable that people keep making this claim and others like it. Indeed, it seems common across many kinds of gadgets that somebody claims “this is better because it is different: ...”. Many of these, all you need to do is think about it briefly, and you can find several reasons why the claim is bad.

• You can carry cargo in the middle of the wheel

  Carrying cargo in the middle of the wheel seems useful.

  • But what things are easy to carry in a centerless wheel, yet hard to carry with a conventional wheel?

  • Really, what things that you want to carry?

    If the only good way to carry greased water balloons on a bike is with a centerless/hubless wheel, well okay. But I rarely want to carry greased water balloons. So for me, it is not an advantage.

       
  Screen captures from [https://www.youtube.com/watch?v=khw8kpt0SGA as of 2025-03]

  To be an advantage, it has to be something that is hard to carry without a centerless wheel, and also something which many people want to carry.

  I have seen a picture of a bike trailer similar to the Extrawheel [https://extrawheel.com as of 2025-03], but with a centerless/hubless wheel; it was used to carry a pet dog. For grocery shopping, you could just use an Extrawheel. But it is hard to carry a dog with an Extrawheel. A one-wheel pet trailer may have advantages over two-wheel trailers: a one-wheel trailer is narrower, and the trailer tire follows the bike tires and so may give a smoother ride. And an Extrawheel-style trailer may track better and store more compactly than longer-wheelbase one-wheel trailers, such as the Goeland/Jack Taylor.

          
  From [https://fat-bike.com/2020/10/first-look-extrawheel-mate-fat-bike-trailer as of 2025-03], [https://www.bikequarterly.com/biketest/#lightbox[gallery-1]/93 as of 2025-03], and [http://www.blackbirdsf.org/taylor/images/goeland.jpg as of 2025-03]

  The dog trailer is the only use I recall where somebody built one, and when I saw it, I thought there might be many riders who find the advantages pay for the disadvantages. Lots of folks have dogs and take them (or want to take them) on rides. And for shorter trips, worse drag/cost/weight/durability may be an okay tradeoff to get better tracking and a trailer that stores small.

  So that may be an example of a “cargo” use for centerless/hubless wheels where the advantages outweigh the disadvantages. There may be others I do not know about, or uses where I do not see a value, but others do. But it seems noteworthy that folks flogging centerless/hubless wheels mention “cargo”, yet rarely seem to have an example which shows an advantage for centerless/hubless wheels.

  Also, a dog trailer based on a Goeland/Jack Taylor may be only a little worse. Maybe there are developments to the Goeland/Taylor so it stows more compactly? A dog trailer using a centerless/hubless wheel seemspromising, but maybe the same effort put in to a Goeland/Taylor gives something just as good.

• A folding bike which is more compact when folded

  The basic idea: spoked wheels with hubs have a lot of “wasted space” inside the wheel; cast wheels are thinner, but still use space in the middle of the wheel. A centerless/hubless wheel leaves that space open, and so when a folding bike is folded, it can be stowed in that space.

  Thus, the folded/stowed bicycle is basically no bigger than than the size of two tires. This could be good for folding smaller, and/or for folding faster than other bikes. Also, most of the outer profile of the folded bike is soft tire rubber, rather than hard metal. That could make the folded bike better for carrying and stowing.

  I can imagine many folding bike riders might want one: many folding bikes already sell well, even though they have lots of compromises (drag, cost, weight, durability) compared to regular bikes. Why? Many people commute, shop, or travel via bus, train, carpool, or other transit; and use a bike as “faster than walking” at the ends of transit. It may not be possible (or practical) to carry and stow a full-size bike. A folding bike may be easy to carry. And even with its limitations can be better than walking, hailing a taxi, or otherwise doing without a bike.

  Here, a centerless/hubless bike which folds better could “pay” for the disadvantages compared to conventional wheels. I have not seen where somebody built one, though. It may be there are too many practical problems with folding in to the middle of the wheels; or that centerless/hubless wheels add too many problems.

• No air drag from moving spokes, so less air drag overall.

  Above 20 km/h, most of your power goes to air drag. Most air drag is from pushing you (the rider) through the air. But a bicycle’s drag is also significant. Spokes swishing through air are a significant part of a bike’s air drag. So for at least some riding, wheel drag is enough that reducing it is an advantage.

  Centerless/hubless wheels do away with spokes, so they should have less air drag. The “Black Hole” centerless/hubless front wheel was used in some races where the rider won, suggesting they can provide a net advantage [Black Hole].

  There are at least five challenges for using centerless/hubless wheels to improve aerodynamics:

  • Centerless/hubless wheels have more bearing drag, which subtracts from any aerodynamic gains. You could have better aerodynamics yet still ride slower. Bearing drag needs to be small enough so the wheel is faster overall. But, it is hard to reduce centerless/hubless wheel bearing drag, especially if you also care about cost, weight, and durability.

  • There are other ways to reduce spoke drag, such as shielding the spokes. That can reduce the advantage of no-spoke wheels.

    Most of a wheel’s spoke drag is the spokes near the top of the wheel: that is where the spokes are moving fastest — they are stationary at the bottom of the wheel. Also, air drag goes as the square of the speed — e.g., if you double the speed, you get 2 × 2 = 4 times the drag. Light plastic panels (sometimes called “fairings” or “shrouds”) can cover spokes at the top part of a wheel, so the spokes and air inside the panels have similar speeds.

    An example is Nullwinds [https://nullwinds.com/pages/technology as of 2025-03]. An earlier version used a flat panel on either side of the wheel, where each panel “hides” in the aerodynamic profile of the tire. The current version is much wider but also covers the top of the tire, to reduce both spoke and tire drag. Both reduce spoke drag without using a centerless/hubless wheel.

            
    From [http://nullwinds.com/products-fairings.html as of 2017/02] [https://nullwinds.com/pages/aerodefender-wheel-fairings as of 2025-03] [https://nullwinds.com/pages/vehicle-aero-technical-discussion as of 2025-03]

    Spoke fairings reduce spoke drag, but still have more drag than “no spokes”. However fairings have low cost and weight, and have no bearings/etc. to add drag or wear out.

    The main disadvantages of fairings are (a) you still have some drag; and (b) fairings are not permitted for most racing, so few companies have made an effort to develop and sell them. But for ordinary riders, they could be a better aerodynamic solution than centerless/hubless wheels.

  • Another choice is “aero” wheels with fewer spokes. Areo wheels have problems: they tend to be heavier and more expensive, and they may be less durable. (“May be”: Aero wheels can be as durable as regular wheels, but they need to be designed and made with durability in mind. That may add cost and weight. Historically, some aero wheels have been less durable [http://pardo.net/bike/pic/fail-020/index.html as of 2025-03].) Wheels with just a few spokes may also be more prone to accidents from squirrels and other things getting in the spokes. However, aero wheels may have fewer/smaller down-sides than hubless/spokeless wheels. And even though an aero wheel has some spoke air drag, it may have enough less bearing drag that it is a net win over a centerless/hubless wheel.

  • A wheel’s total air drag also includes the tire. As with spokes, a tire’s aerodynamic drag is highest near the top of the wheel, because skin drag goes up as the square of the speed. On a centerless/hubless wheel, the tires still rotate, so they still have this drag. Getting rid of spokes can reduce air turbulance around some parts of the tire, so “no spokes” might also help a tire’s air drag. But the part of the tire with the highest air drag is at the top. That is above the air which is stirred by spokes, so a centerless/hubless wheel does not help that.

    As an aside: it appears that in ordinary riding, fenders do not slow you down. That may be surprising, since fenders need to be wider than the tires to be effective, so fenders present a bigger aerodynamic profile. However, fenders carry some air inside them. As with fairings, the tire speed and air speed are more similar. Thus, it seems the fender’s wider profile is offset by reduced tire drag. (No, this is not directly related to hubless/centerless wheels, just thrown in for interest.)

  • Ordinary riders go slower than racers, and so often have much less wheel drag than racers — so there is less potential benefit. Also, ordinary riders often have luggage and less-aerodynamic “street” clothes, which mean wheel drag is an even smaller percentage of total drag. In other words, an aerodynamic improvement may be big enough to matter for racing, but too small to matter for ordinary riders and riding. And an aero “improvement” may be a disadvantage if it adds signficant cost, weight, and maintenance.

  Note also that a wheel’s aerodynamic drag is not yet well-understood. This does not directly affect the actual benefits (or not) of a centerless/hubless wheel, but it does complicate measurement, analysis, and comparison. Notably, what we now think is the "better" wheel may turn out to have more drag than something we think is a "worse" option [Wheel aerodynamics are still incomplete].

    
  From [https://www.bikeradar.com/features/the-weirdest-bike-on-ebay-right-now-15000-of-retro-aero-tech as of 2025-03]

  The idea “no spokes have no air drag” is reasonable.

  But better aerodynamics may have only modest up-sides for ordinary riders, yet centerless/hubless wheels have many problems to solve — drag, cost, weight, maintenance — before ordinary riders can benefit. Put another way, ordinary riders may find aero wheels or fairings a better all-around solution, even if the total benefit is a little less.

To summarize: centerless/hubless look nice, but rarely have any other benefit. And they have many down-sides, so for ordinary riders and riding, they are usually worse than conventional hub-type wheels. Many problems with hubless/centerless are hard to solve, so it is also unlikely that next year’s centerless/hubless wheels will do better. E-bikes are less sensitive to added drag, cost, weight, so centerless/hubless wheels may be less of a bad choice. Still, it is hard to make enough improvements that they can meet or beat ordinary hub-type wheels.

Examples: Products, Prototypes, Proposals

• Arthur
• Black Hole
• Blood Falcons
• Cyclotron
• Cyclotron folding
• Gafoor et al.
• Hotard-Campbell-Collier
• Karthi et al.
• KOSMOS
• Nulla
• Reevo
• Ross
• Sada
• Terpstra
• The Q
• Thorpe
• Top Secret
• Twist
• Ujet
• Yale demo

Some examples of bicycles using centerless/hubless wheels, including products, prototypes, proposals, and patents:

Name           	   Structure  Bearing         Brake   Drive      Rack/fend
-------------  	   ---------  --------------  ------  ---------  ---------
Arthur         	   hoop       rim-on-rollers  rim     belt       Y
Black Hole     	   rod        rim-on-rollers  rim     N/A        N (race)
Blood Falcons       -          -               -       -         N (art)
Cyclotron      	   hoop        -               -       -         -
Cyclotron Folding  hoop        -               -       -         Y
Gafoor et al.  	   hoop       rim-on-rollers  rim     ring gear  N
Hotard et al.      hoop       rim-on-rollers  none    N/A        rack
Karthi et al.      hoop       rim-on-rollers  rim     ring gear  N
KOSMOS             hoop        -              rim     N/A        Y
Nulla          	   rim        rim-on-rollers   -      ring gear  N
Reevo          	   hoop       rollers-in-rim  rim     ring gear  N
Ross           	   hoop       bearing          -      ring gear  N
Sada           	   rim/rod    rim-on-rollers  none    roller     N
Terpsra        	   rim        rim-on-rollers  none    N/A        N (art)
The Q          	   hoop       rim-on-rollers  remote  chain      N (art)
Thorpe         	   rim        rim-on-rollers  none    roller     Y
Top Secret     	   hoop       bearing         disk    ring gear  N
Twist          	   hoop        -              none     -         N
Ujet           	   hoop       bearing         disk    N/A        Y
Yale demo      	   rim        rim-on-rollers  none    ring gear  N

• Structure: for more see Wheel structure types
• Bearing: for more see Bearing layouts
• Brake: rim brake, "disk" for a disk mounted to the rim, "remote" for a brake connected via the drivetrain, "none" for no brake on the centerless/hubless wheel; "—" if the brake type is not described.
• Drive: N/A when only the front wheel is centerless/hubless; "ring gear" for a ring/spur gear; "belt" or "chain" for driving a large sprocket or pulley; "roller" for friction drive to the rim; “—” if the drive type is not described. See also [Drivetrain types].
• Rack/fend: "Y" if there is at least some discussion or consideration; "N" if there is not one, or if there is but it seems non-functional. "N (art)" for bikes meant as art, so you would not expect racks/fenders anyway, and "N (race)" for race-oriented use, where rack and fender mounts are already uncommon [http://pardo.net/bike/pic/mobi/d.hidden-bad-fender-eyelets/000.html as of 2025-03].

Arthur

        
From [https://patents.google.com/patent/US3329444A/en] as of 2025-03

[https://patents.google.com/patent/US3329444A/en as of 2025-03]

Vehicle frame and spokeless wheel arrangement
Inventor Lidov Arthur
Current Assignee Individual
US Patent US3329444A
Granted 1967-07-04

Notes:

• The goals are (1) to provide suspension so solid rubber tires can be used; and (2) a design which can be made of plastic, so it can be made cheaply.

• It uses rollers to support the wheel. Near the ground, the rollers run on the inside surface of the rim. Away from the ground, the rollers run on the tire’s surface.

• The tire tread is two rings, with a gap in-between. On the rear wheel, a smooth (not toothed) belt runs in a groove between the parts.

• "Suitable brake means, not shown" — it does not further discuss how the brake works.

• The wheel supports serve as fenders. Neither fenders nor racks are discussed in the patent.

• These wheels would have much higher friction than normal wheels, due to solid tires and using a belt drive. Rollers running on the tire could have high drag, but if most weight is supported near the bottom of the wheel, the top rollers would have only light load and thus light drag.

• It seems likely that the top rollers would contact a lot of debris picked up by the tire. It could be hard to protect them, so they seem likely to sometimes jam and skid on the tire, rather than roll. Rollers near the bottom run on the inside of the rim, which should be cleaner. However, if the wheel passes through e.g., a deep puddle, they may get bathed in mud and dirty water. Thus, would also be prone to faster bearing failure than bearings in a hub.

• It uses a belt tensioner on the taut side of the belt, so pedalling would tend to slack the belt and let it slip. That is not directly related to “centerless/hubless wheels”, But it suggests the design was not thought out carefully nor tested before filing a patent.

• I included this as being more like current mainstream bikes than some of the older centerless/hubless wheel patents. Although the patent ignores drag, I note patents are meant to discuss benefits of an idea, not down-sides. But the design shown here seems likely to have high friction.

Black Hole

     
From [https://www.bikeradar.com/features/the-weirdest-bike-on-ebay-right-now-15000-of-retro-aero-tech as of 2025-03] [https://weightweenies.starbike.com/forum/viewtopic.php?t=724 as of 2025-03]

Notes:

• The goals of this wheel are (1) reduced drag through improved aerodynamics; and (2) reduced weight.

• The wheel structure is “structural rod” [Wheel structure types]. The half-open center looks hoop-like, but the rollers are only near the top and bottom of the wheel. Fore and aft portions have no load path between rim and “fork”.

• It uses three rollers near the bottom and one at the top. The rollers are roughly at 12 o’clock (top), 4:30, 6 (bottom), and 7:30. In other words top and bottom, plus 45 degrees ahead and behind of the normal road contact (approximately).

  The rollers are mounted to the "fork", and run on a track on the rim. The patent shows several options for rollers, including:

  • Separate lateral and radial rollers, like [The Q].
  • Pairs of rollers in a "V" configuration, which engage a V-shaped track on the rim.
  • Single inline rollers with a "V" or inverted-"V" which engage a matching track on the rim.

  It appears the product uses single rollers, probably like patent Figure 4.

• The V-shaped and single-inline roller configurations have some rubbing/scuffing compared to a radial roller. The rubbing adds drag.

  Lubrication can reduce drag from scuffing, but the track is exposed, so sticky lubricants are likely to attract dirt. The track is exposed, so all lubricants (sticky or not) are likely to get washed or worn away. For dry riding, frequent relubrication (e.g., daily) is probably fine, as this is a race product. However, there is a risk lubricant would get on the brake track.

  Radial plus lateral rollers are also shown in the patent, but are wider, which would hurt aerodynamics.

  There is a trade-off between roller configuration, aerodynamic drag, and rolling drag.

  • A V-arranged pair of rollers (patent Figure 11) are narrower than radial+lateral rollers, so should have less aerodynamic drag. However, the radial+lateral layout has the major loads (radial) carried on rollers which can turn without scuffing. Scuffing causes drag. A V-aranged pair of rollers has scuffing because the wheel’s track has different speeds depending on the radius where it touches a roller.

  • A V-arranged pair of rollers has less scuffing than single-row rollers. With a V-pair, the roller’s face has constant speed. Whereas a single roller’s face has speed which depends on the radius on the roller. Worse, the speed gradiant opposite the track on the rim. The roller is much smaller diameter than the rim track, so the speed gradiant is much larger. Thus, a V-pair probably has much lower rolling drag than single-row rollers.

  • A V-pair is wider than single-row rollers, so may contribue significant air drag. A slightly narrower V-pair alternative is V-configured bearings with a track like patent Figure 7, where the rollers are not quite opposite each other, so the left roller can run on the right track and the right roller runs on the left track. That would be a little bit narrower, but still not as narrow as a single row of rollers. Also the V-track would tend to trap grit, both from centrifugal force and gravity.

  • Yet another alternative is flanged rollers or a flanged (channel) track. That way, the main load (radial) is carried without scuffing. Side loads cause scuffing, but wheel side loads are usually much smaller than radial loads, so this might be a good trade. Beware that a channel track on the rim also tends to trap debris, so may be unsuitable for general use. Flanged rollers do not tend to trap debris.

• I have seen nothing specific about the bearings used in these. As this is a “lowest drag“ on-road or track (not MTB) race product, it probably uses shielded bearings (non-contact, see [Bearing seals]) for low drag.

  Non-sealed bearings are likely to fail faster in ordinary riding. Even before they fail, dirt can get in and increase the drag.

  Non-sealed bearings are entirely reasonable for a race product: regularly clean and lubricate the bearings, and especially after every rain ride. And/or switch to a different wheel when it may rain. Racers often use diffrent wheels depending on the course and weather anyway. So switching is a normal thing to do. The Black Hole uses an integrated fork, so switching wheels is somewhat more work. But Rinko riders often remove and replace their fork twice in one ride [https://www.rossmancycles.com/rinko as of 2025-03-13], so it is entirely reasonable to occasionally change forks before a race.

  Sealed bearings [Bearing seals] should last longer, but may have a lot more drag. If drag is big enough, another approach (aero wheels, fairings, etc.) has more overall benefit.

• It uses rim brakes.

• It is a front wheel only, so no drive mechanism is needed.

• No fenders or racks are considered, but this is a race product, where rack and fender mounts are already uncommon.

• Comments online say some riders used it races where they won, suggsting they can provide an advantage. But maybe the rider would have won by a larger margin using a conventional wheel? Still, races are often won on slim margins, so if the Black Hole had big down-sides, they likely would have lost.

  I do not know of controlled tests to say whether it was an advantage.

• It was banned by the UCI, so there was little motivation to keep developing it.

• The Black Hole was developed on a small budget and raced only a few years. It seems likely the aerodynamics and maybe rolling drag could be improved.

• The aerodynamic benefit may be big enough it matters for racing, yet too small to matter for ordinary riding.

  • At lower speeds, aero drag is less.

  • Most aero drag is the rider, not the bike. Ordinary riders often ride in a less-aerodynamic position, so make an even larger part of total air drag. In turn, a given bike drag improvement is a smaller fraction of the total.

    Also, ordinary riders often have less-aerodynamic clothes, luggage, racks, lights and other things which cause drag. In turn, a given spoke drag improvement is a smaller fraction of the total.

  • For ordinary riding, it hardly matters if you use an approach which is a few seconds slower in an hour. For racing, a few seconds in an hour is enough to win or lose a race. But for ordinary riding, it is too small to notice even in an all-day ride.

    In turn, a “second-best” approach (aero wheels, fairings) may be a better trade-off: you get some drag benefit, with little harm to cost, weight, or durability.

• It was competing against other aero wheels of the day. Modern wheels may be better. So even if it had an advantage in 1993, it might today be slower than today’s aero wheels.

• Wheel/spoke fairings may also compete against it for general use. However, wheel/spoke fairings are also banned by the UCI.

• Online comments said it was loud, and the rim support rollers wore much faster than ordinary hub wheel bearings. Both would limit its appeal outside of racing.

Links:

• [https://www.bikeradar.com/features/the-weirdest-bike-on-ebay-right-now-15000-of-retro-aero-tech as of 2025-03]

  Ride reports from the time claim that the ride quality of the Black Hole was remarkably normal, though was said to be incredibly noisy, with the whole structure reverberating and amplifying vibrations on rough surfaces.

• [https://bikerumor.com/how-to-break-into-the-cycling-industry-reynolds-aerodynamics-expert-paul-lew as of 2025-03]

  How to Break Into the Cycling Industry — Reynold’s Aerodynamics Expert Paul Lew
  Tyler Benedict
  2012-10-25

  We started a company in Indianapolis, IN, where I grew up, and started designing and manufacturing a spokeless, hubless front wheel for track cycling called the Black Hole…

  It took us about three years, and in 1993 we started selling them. It was exciting, but very radical. For most people, it was interesting, but not something they’d buy. In 1994, Bryan Walton, a former 7-11/Motorola/Saturn pro cyclist started working with us and started racing the wheel in the 4000m individual pursuit. Then Jurgen Zack, a triathlete that set the bike split record at Kona several times, was going to race it, too, but the UCI banned it calling it an unfair advantage. So, that was the end of the Black Hole wheel. We shut the company down. We sold less than 100 wheels, but people still have them.

• [https://rouesartisanales.over-blog.com/article-15522060.html as of 2025-03]

• [https://weightweenies.starbike.com/forum/viewtopic.php?t=724 as of 2025-03]

  Cyco 2003-11-25

  Seemed like a cool idea but.... The hole got smaller and smaller on every prototype [...]. It had 4 bearings, at about 12, 4:30, 6, and 7:30. The rim rolled on these.

  wally318, 2004-12-04

  Imagine the sound of the rim rotating on the 4 rollers and guitaring on the hollow fork/wheel. [...] If memory serve’ me there were 2 model between 1750 and 1900 grams. What does a weight weenie fork and aero front wheel weigh?

  Samu Ilonen 2004-12-04

  EA EC90 Aero weights ~370g and LW front weel is under 500g or HED H3C is ~ 680g. So It’ around 1kg [For a conventional fork and front wheel]

  Fying Wombat, 2004-12-06

  I believe the biggest problem was with the bearings for the “rim” they were subject to rapid wear and once the design vanished spares were no longer available.

    

• [https://patentimages.storage.googleapis.com/86/14/f4/61db3a94cc703c/US5419619.pdf 2025-03], [https://patents.google.com/patent/US5419619A/en as of 2025-03]

  Title: Hubless wheel
  Inventor: Paul E. Lew
  Current Assignee: Wear and Tear Inc
  Patent: 5,419,619
  Granted: 1993-05-30

  Abstract: A hubless wheel for a vehicle which provides advantageous weight and aerodynamic properties.

• [https://bicycledesign.net/2006/08/hubless-wheels as of 2025-04]

     Supposedly, the wheel rotated on three self-lubricating ball bearings in the monoblade fork. According to the company, the fork/wheel assembly weighed about a pound less than a conventional system.

• [https://bicycledesign.net/2006/08/the-amazing-shrinking-black-hole-and-a-lime as of 2025-04]

     [T]his picture of the wheel from a mid nineties issue of Road Bike Action.

• [https://www.designboom.com/design/hubless-spokeless-bicycle-wheel-compilation as of 2025-04]

    

  ‘blackhole’ hubless wheel system by wear and tear image © damon rinard

  hubless bicycle wheel by karl kammelzel

Blood Falcons

        
From [http://bloodfalcons.blogspot.com/2013/03/hubless.html] as of 2025-03

Notes:

• Art drawing, not meant to be a real thing.

• My main reason for including this is it uses “support rods’ [Wheel structure types] to support both front and rear wheels, showing more layout choices.

Links:

• [http://bloodfalcons.blogspot.com/2013/03/hubless.html as of 2025-03-13]

Cyclotron

           
From [https://www.kickstarter.com/projects/1989795590/the-cyclotron-bike-revolutionary-spokeless-smart-c/description] as of 2025-03, [https://www.designboom.com/technology/cyclotron-bike-spokes-tyres-18-07-2016] as of 2025-03 [https://www.renehersecycles.com/jobst_cornering] as of 2025-03

Notes:

• The Cyclotron was advertised as an e-bike with centerless/hubless wheels, improved aerodymics, and automation/computer features.

• Some people online suggest the Cyclotron was intended as a scam, and never meant to be a working prototype or usable product. For example, there are videos which appear to show a ridable prototype, but the rider’s pedaling motion is irregular in a way which suggests the pedals are not connected to the wheel, and instead the rider is just pedaling freely while coasting. Advertising claimed it was “on sale”, but I did not find anything suggesting any were delivered. Nor do I see test data which would suggest they had a prototype working well enough to test it.

  The wheels are enclosed in a housing with slots at the bottom.

  • Probably the housing would be damaged by rolling the bike down steps, riding off a curb, or over a sharp dip in the road, such as you sometimes find at driveways and intersections. They might also drag on off-camber curves (bike leans one way, road slopes the other).

  • Indeed, the housings are so close to the ground that simply going around a corner seems likely to drag the housing and maybe cause a crash. A photo above shows a rider on an ordinary bike at close to 45° lean, whereas the Cyclotron looks like it could drag the housing and lose tire contact at perhaps 20°.

  • You could build conventional wheels in housings and have the same problems. So (a) this is not a problem specific to centerless/hubless wheels; and (b) even a crude prototype would have shown the (presumed) problems, with no need to wait for a prototype using centerless/hubless wheels. It suggests that even though the sellers described it as “on sale”, they had no ridable prototypes. There are videos of somebody “riding” a Cyclotron, but as noted above it may have just been low-speed coasting.

  The wheel housings also seem likely to pick up mud and other debris and jam the wheels. That happens occasionally with ordinary bike fenders, even though they have much more open space for debris to escape. This again suggests they did not have a ridable prototype. Including: they could have built conventional wheels in housings and had about the same problems.

  This is all background to say we should be careful what we “learn” from the design: features may be mock-ups and not anything that works.

  I will discuss this as “the Cyclotron has ...” as-if it really existed and was workable. Best to read that as “the Cyclotron was claimed to have ...” or similar.

• The Cyclotron uses wheel space for storage. This seems to be the only reason they give for “why centerless/hubless’. There is some discussion of aerodynamics, but those claims focus on frame profile and hidden routing for cables/etc.

  The baskets look convenient. On conventional wheels, joined panniers that drape over a conventional rack would provide similar total capacity and ease of handling. Two panniers cannot carry all sizes/shapes well that you can carry in one large basket; but panniers still work well for many things.

  The baskets do not look at all weather resistant — whatever you put in them will get wet if you ride in the rain. Most panniers and trailers offer a rain cover or are made of waterproof material. A rain cover should be possible for a basket, but this is not discussed.

• There is no information on how the wheels work — structure, bearings, etc..

  Even for a legitimate product, potential buyers should be cautious about everything custom to a bike. First, if it breaks or works poorly, you can be stranded or left with an unridable bike. Some things break in ways that can hurt you. Second, you cannot simply go out and buy another maker’s replacement or upgrade parts.

  When there is little information on the novel parts of a bike, that should make you worry you will wind up with something expensive and useless.

• Cyclotron claims integrated brakes: “modified caliper brakes” — but give no details. Hiding the brakes is desirable for a clean look. But in actual riding, brakes need vigorouos air flow to keep them from overheating. Rim brakes sometimes melt the brake pads, even with the rim running in free air. Disk brakes on bicycles can glow dull red even with free air flow. E-bikes are even more demanding, and some disk brake makers have introduced up-sized brakes specifically for e-bikes.

  The lack of details about the Cyclotron’s brakes suggests they did not have a working/ridable/testable prototype.

  A lack of concrete details should make you suspicious of even a legitimate product: if the brakes do not work right, the bike is unridable. And since the brakes are (presumably) specific to the Cyclotron, you cannot simply go out and buy another maker’s replacement or upgrade parts.

  This is not specific to centerless/hubless wheels — hiding the brakes on hub-type wheels has the same issue.

• How is the rear wheel driven? Chain and ring-gear drives in other centerless/hubless wheels lead to various lumps in the housings. For example, Reevo’s drive needs space for a spur gear to drive a ring gear on the wheel. The Cyclotron’s housing does not appear to have enough space for a spur gear. It may be possible to build something which fits, but I did not find any information.

  Like the brakes, this also seems like the sort of thing that customers want to know, makers should be happy to describe, and a lack of details should make you suspicious. Even for a legitimate product, if the drivetrain is inadequate, the bike is unridable. And as with brakes, the drivetrain is presumably specific to the Cyclotron, so you cannot simply go out and buy another company’s replacement or upgrade parts.

Links:

• [https://www.kickstarter.com/projects/1989795590/the-cyclotron-bike-revolutionary-spokeless-smart-c/posts/2725702 as of 2025-03]

• [https://luxurylaunches.com/transport/cyclotron-the-worlds-first-hubless-smart-bike-is-on-sale-for-2000.php as of 2025-03]

• [https://www.reddit.com/r/shittykickstarters/comments/bfhwro/cyclotron_publishes_update_we_cant_afford_to as of 2025-03]

Cyclotron folding

           
From [https://www.frankebearings.co.uk/new-contract-with-award-winning-hub-less-wheel-electric-bike] as of 2025-03, [https://patents.google.com/patent/US378253A/en] as of 2025-03

Notes:

• It appears this is unrelated to the [Cyclotron].

• There is a picture which appears to show somebody riding a prototype, but it is likely an artistic rendering: the picture first appeared when they said they were still working on a ridable prototype, and I found no information suggesting a working version or any other prototype has been built.

• I have ridden bikes with a seat directly over the rear wheel. They tend to "wheelie" even when you do not want it. Likely, if Cyclotron built one — even a non-foling prototype with standard hub-type wheels — they would find few people want to ride this for ordinary (not stunt) riding. Since they are showing the "seat over wheel" version, that also suggests to me they have not built one.

  The front wheel is steered about an axis which is in line with the back of the front wheel. Other bikes have been built this way and are ridable. However, I suspect most people would dislike the handling. Again, it suggests to me they have not built one.

  It is however, being presented as an idea, not as a sellable product. The main focus of Cyclotron appears to be powered 4-wheel carts and cars, not bicycles. The bicycle may be a design exploration, or to get publicity.

• The purpose of the centerless wheels is to make it easy to fold the bike quickly.

• It uses a “support hoop’ [Wheel structure types] to support the wheel.

• The bearing layout [Bearing layouts] is not clear.

• No brake is shown or discussed.

• The drive is not described, but appears to be a ring gear [Drivetrain types]. Some of their non-bicycle designs are shown with a ring gear.

• There is no discussion of mouting fenders or racks. Some of the structure forms a partial fender, but it does not cover the front wheel near the ground, so the rider will still likely get somewhat wet with dirty road water. It could probably be extended to make a good fender.

• The general idea of folding in to the "wasted space" of a wheel seems reasonable, and can offer both large wheels for ride comfort, and also fast and compact folding.

  It seems like it should fold about as quickly as best-in-class fast-folding bikes, like the Bike Friday Tikit, which can fold and unfold in 15 seconds [https://www.bikeradar.com/reviews/bikes/folding-bikes/bike-friday-tikit-review as of 2025-03], but which uses smaller-than-standard wheels to get to its folded size.

  The Cyclotron folder looks to use a larger-than-standard front wheel. So even though the folded size is not much more than the front wheel, it is still about half the size of of a standard bike. Which suggests a careful redesign of a WW2 paratrooper bike, using moren custom parts (e.g., narrower hubs, single-sided fork) might yield a bike which folds as fast and to a similar size, which uses a standard riding position, and which avoids the need for hubless/centerless wheels.

          
  From [https://onlinebicyclemuseum.co.uk/1943-bsa-airborne-folding-paratrooper-bicycle] as of 2025-03, [https://onlinebicyclemuseum.co.uk/1943-ww2-bsa-airborne-folding-paratrooper-bicycle/1943-ww2-bsa-airborne-folding-paratrooper-bicycle-21] as of 2025-03

]

Links:

• [https://www.frankebearings.co.uk/new-contract-with-award-winning-hub-less-wheel-electric-bike as of 2025-03]

• [https://www.bikeforums.net/folding-bikes/1169928-worth-share-interesting-fold.html as of 2025-03]

• [https://www.ecyclopic.com as of 2025-03]

• [https://www.cyclopic.co.uk as of 2025-03] Main site, focused on centerless wheels for automotive (not bicycle) use.

• [https://www.standard.co.uk/news/tech/electric-penny-farthing-cyclopic-a4110151.html as of 2025-03]

  Electric penny farthings are coming to improve your London commute
  Mark Blunden
  2019-04-05

  The six-speed bike is powered by an electric drive system, which also controls the brakes and steering as there is no front axle.

  Prototyping, including trying to reduce the weight, is due for completion this summer.

[https://patents.google.com/patent/US378253A as of 2025-03]

Gafoor et al.

           

Notes:

• A school project; not meant to be a path to a product.

  It is not clear what was the goal for this project. The abstract of the paper says centerless/hubless wheels are not used widely in bicycles “So, considering it as the most objective of our venture, the hubless cycle is designed.” They both designed and built a hubless bicycle, but it does not seem to have been evaluated in any way.

  The paper appears in a journal of “Applied Science and Engineering Technology”. I expect papers in such a journal report on experiments. An experiment should have some goal — e.g., to see if it is feasible to construct something with a given strength; or to measure properties like drag, weight, etc. The paper does not seem to describe an experiment (what is the hypothesis or goal), nor report result (what did we measure or observe).

  The paper has details about construction which may be interesting to readers here.

• The paper’s abstract makes vague and unlikley claims about the benefits of centerless/hubless wheels:

  • In old bicycle the effort required to ride the bicycle is more. In the hubless cycle the gear transmission is provided which reduces the rider’s effort to pedal and ride the bicycle.
  • [In a] hubless wheel, the handle bar is attached to the wheel edge which [improves] steering efficiency.

  It could be “old bicycle” means hub-type wheels, in which case they seem to be claiming a gear transmission is more efficient. They do not explain or measure. Or, it could be that “old bicycle” means the donor bike they used. In which case they could have gotten reduced effort by using different sprockets. The paper does not explain.

  Also not discussed: how to build their bike, but keep the original gear ratios — which the bike probably had for a reason.

  They do not explain what is “steering efficiency”, so it is hard to analyze the claim. “Steering efficiency” is not a standard term. I do not know of any problems with steering hub-type wheels which I can think of as “steering efficiency”. Nor do I see where steering the rim solves a problem with hub-type wheels, efficiency or otherwise.

• It uses a “support hoop” [Wheel structure types] to support the wheel.

• It uses “rim-on-rollers” bearings [Bearing layouts].

• It uses rim brakes.

• It uses a ring gear drive [Drivetrain types].

• There is no discussion of mouting fenders or racks.

• The wheels use a “rim-on-rollers” approach [Bearing layouts]. There are separate wheels and bearings for lateral and radial loads.

  To support weight, most wheels use smooth idler wheels running on a smooth track. This bike supports radial loads with spur gears which engage the ring gear. Only one gear transmits torque.

  Supporting loads with gears is handy: it avoids the need to create both a ring gear and rolling tracks for radial support wheels. However, supporting radial loads with gears is probably much less efficient, and wears the ring gear faster.

  The rear wheel drive gear is placed at the bottom of the wheel. This requires a longer chain. But there may be two benefits. First, the rider’s weight resists the gear pushing out of engagement, which may reduce the amount of structure needed. Second, the spur gear is larger than the other radial support gears. The highest radial load is at the bottom of the wheel, and using the largest gear there probably provides better support and maybe lower drag than the same force going through a small gear.

• The ring gear is exposed and so probably wears faster than a housed gear. For an academic project, this is not surprising.

• It is not stated whether the wheel bearings use seals. One photo appears to show a shielded (not sealed) bearing.

  Bearing lifetime is not directly an issue for an academic project. But if ridability is hurt by seal drag, it would be good to include this in the design and analysis.

• There is no evaluation of the bicycle, such as rolling drag, drivetrain drag, or how it compares to a similar bicycle using conventional wheels.

Links:

• [https://www.ijraset.com/research-paper/design-analysis-and-fabrication-of-hubless-bicycle as of 2025-03-06]

Hotard-Campbell-Collier

           
From [https://www.core77.com/posts/25754/Reinventing-the-Hubless-Wheel-Transport-Is-a-Trunk-for-Your-Bike-by-David-Hotard-Matthew-Campbell-n-Edwin-Collier] as of 2025-03

Notes:

• The purpose of the centerless/hubless wheel is better cargo storage.

• It uses a “support hoop’ [Wheel structure types] to support the wheel.

• It uses “rim-on-rollers” bearings [Bearing layouts].

• No brake is shown; it appears a rim brake woulld work.

• It is a front wheel, so there is no need to drive it.

• The front wheel is a rack. There is no discussion of mouting fenders.

• “The final prototype is not rideable since it is made out of wood.” So all claims (drag, cost, weight durability) are proposals, and might be much different than something you can ride.

• There is no discussion of drag/friction.

• There is no discussion of cost, but it is a prototype. Nothing appears expensive.

• “The wheel alone is very lightweight since the storage compartment is made out of PET-G—it’s comparable to the weight of an OEM road bike wheel.” However, it is not at all clear how strength, durability, etc. of a product woulld compare. There have been many wheels over the years which are light, but which fall apart in use.

• Bearing durability is likely to be poor for the same reasons as discussed above: some of the bearings are close to the road, so are likely exposed to lots of grit; and are likely to be immersed in dirty water — e.g., rolling through a deep puddle. If the bearings are not sealed, it is likely gunk will get in to the bearings and damage them. If the bearings are sealed, the extra exposure compared to a hub bearing means durability is probably worse; at the same time, seal drag is likely to be much higher than hub bearings, maybe enough drag that you notice and care.

  Since this is a “cargo” application, riders probably care a lot about durability. On the other hand, they are probably okay with a lot more drag than riders with a “race” application.

• Durability of the rollers and track are unknown. Online comments on the Black Hole suggested that rollers wore out quickly, but it was unclear from comments if the problem was the bearings or the roller surface where it meets the track on the rim.

• A goal is “commuters [can] travel with any bag, not just a rack-able bag” One common alternative is a rear basket. It is somewhat easier for somebody to steal a bag from a rear basket while you are riding. Another common alternative is a front basket. The bike is somewhat tippier when stopped with a high-up load; and a load may interfere with lights mounted on the handlebars.

       
  From [https://m.media-amazon.com/images/S/aplus-media-library-service-media/e2e28df2-b632-4f02-b5c2-b32b4168036f.__CR0,0,970,600_PT0_SX970_V1___.jpg] as of 2025-03, [https://www.aliexpress.us/item/3256805684504089.html] as of 2025-03.

Links:

• [https://www.core77.com/posts/25754/Reinventing-the-Hubless-Wheel-Transport-Is-a-Trunk-for-Your-Bike-by-David-Hotard-Matthew-Campbell-n-Edwin-Collier as of 2025-03]

  Reinventing the (Hubless) Wheel: ‘Transport” Is a Trunk for Your Bike, by David Hotard, Matthew Campbell & Edwin Collier
  Core Jr
  2013-10-17

  Although it may be more expensive to produce and currently less structurally sound than a traditional wheel, we know that in many cases people are willing to pay more for a design that satisfies their needs. We did a lot of research looking at concept bikes, current products and observing users. The hubless wheel storage system brought those three areas of research together by giving commuters the ability to travel with any bag, not just a rack-able bag, and also showed a practical purpose for a hubless wheel concept.

Karthi et al.

     

• Karthi et al. is an academic bicycle using centerless/hubless wheels. It is not intended as a product or path-to-product.

• Why does it usse centerless/hubless wheels?

  • Looks.
  • Better balancing.
  • Lower weight.
  • Smaller size.
  • The reduction in chain length reduces the fatigue of the rider.
  • The steering has been enhanced.
  • Use of rollers makes the motion of the bicycle more facilitated.

  Note these are taken from both the paper’s abstract and conclusions; you need to read both, as the two have distinct lists (except “looks”, which appears in both).

• It uses a “support hoop” wheel structure [Wheel structure types].

• It uses a rim-on-rollers bearing layout [Bearing layouts]. The rollers are flanged, with radial loads are taken directly (without sliding) but side loads cause sliding [Roller layouts].

• The approach to bearing seals is not discussed [Bearing seals].

• The centerless/hubless wheel uses a rim brake [Brake types].

• It uses a ring gear to drive the rear wheel and has one ratio — aka single-speed [Drivetrain types].

• Fenders and racks are not discussed [How do you mount fenders and racks?].

• Karthi et al. say one reason for centerless/hubless wheels is better balancing, but there is no more discussion. There is a vague claim about “steering efficiency”, but it is not clear if this is the same or different; nor what is “steering efficiency” or how it affects the bike or riding the bike. This claim seems unlikely.

• Karthi et al. say one reason for centerless/hubless wheels is less weight than conventional spoked wheels. It seems likely their example wheel is much heavier than an ordinary wheel. Weight for their design is not discussed otherwise. This claim seems unlikely.

• Karthi et al. say one reason for centerless/hubless wheels is reduced size. Or, more precisely, they say “high ... size” is a problem with conventional bicycles. They do not discuss how their centerless/hubless wheel changes that. This claim seems unlikely.

• It is widely understood that most friction in a chain drive is where the links articulate, which is where they wrap around the sprockets. Karthi et al. claim “The reduction in chain length enhanced the fact of less loss of energy. It will also reduce the fatigue encountered by the driver.”. There is no discussion why chain length would affect efficiency or rider fatigue. They do not evaluate the efficiency of their drive. This claim seems unlikely.

• Karthi et al. claim the steering is enhanced; and that steering the rim directly improves steering efficiency. “Steering efficiency” is not a standard term. I do not think of any other term or concept which makes sense here as “steering efficiency”. Conversely, I am not aware of any problem with steering which is improved with steering the rim directly. They do not explain, and they do not evaluate the bike’s handling nor report on steering efficiency. This claim seems unlikely.

• Karthi et al. “Use of rollers makes the motion of the bicycle more facilitated.” I think this is a claim that using rollers reduces rolling drag, presumably compared to a conventional hub-type wheel. It seems almost certain that their wheel has much higher drag than conventional wheel, for reasons discussed above. They do not evaluate the efficiency of their drivetrain. If I understand the claim (maybe I do not), this claim seems unlikely.

• The load limit of the wheel is not discussed. A rim without spokes is normally much weaker. Karthi et al. weld together the rim and ring gear. Depending on how they are welded together, the ring gear may improve the wheel strength.

• Since the ring gear is welded to the rim, it is effectively integral. Replacing the ring gear thus also means replacing the rim. Conversely, if the brake wears out the rim, then replacing the rim requires replacing the ring gear.

  The ring gear is mild steel and is exposed to mud/dirt/etc. It is likely the ring gear will wear far faster than something similar but run clean and well-lubricated.

  Durability, service, and costs are not discussed — for the ring gear, or more generally.

• Hub-type bearings can last a decade or more of daily service with little or no maintenance. The rollers for Karthi et al.’s wheel are exposed to mud, grit, etc., so seem likely to wear quickly if used in a dirty environment. Durability and maintenance costs for rollers are also not discussed.

• Karthi et al. write: “The outer rim is fitted with a tubeless tyre. The tyre is radial ply tyre. Tubeless tyre offers long life as compared to tube tyres because the tubeless tyre gets easily damaged.“ I do not understand this comment: tube-type and tubeless tires are often of similar construction and durability. It is not clear if the comment was meant specifically for centerless/hubless wheels, or is just a general comment.

  It seems plausible that a tubeless can be used to simplify placement of the valve stem. The paper does not discuss how the tire is inflated.

• There is no information about the wheel bearings, seals, durability, or long-term costs.

• The drive ratio is not specified and is not compared to the original bicycle’s drive ratio.

  Riders often want a multi-speed drivetrain, there is no discussion of how to do that — multi-speed is probably outside the scope of the project, but is still an important topic, at least mentioning it would have been useful.

• Karthi et al. is an academic project, so you would expect it to ignore at least some issues that are important for a product. However the Abstract and Conclusions mention several attributes (balancing, weight, size, etc.) but then never discusses or measures them.

• [https://mail.irjet.net/archives/V6/i3/IRJET-V6I31131.pdf as of 2025-04]

  • The outer rim is fitted with a tubeless tyre. The tyre is radial ply tyre. Tubeless tyre offers long life as compared to tube tyres because the tubeless tyre gets easily damaged.

  • The reduction in chain length enhanced the fact of less loss of energy. It will also reduce the fatigue encountered by the driver.

    The reduction in chain length reduces the fatigue of the rider.

  • In normal cycle, the steering (i.e) the handle bar is attached to the hub of wheel. But in hubless wheel, the handle bar is attached to the wheel rim which increases steering efficiency.

    The steering has been enhanced.

  • input power is given to the rear gear meshed with the internal teeth rack welded with the outer rim.

    Rack is the member which engages with the gear to transfer the drive. It is bent as per shape of the outer rim. Rack is made of mild steel material which can be easily transformed to necessary shapes for providing power transfer to the wheel. The individual parts like rack, outer rim, inner rim nylon rollers and bearings are assembly.

KOSMOS

              
From [https://www.kickstarter.com/projects/451828310/the-worlds-first-hubless-foldable-e-bike] as of 2025-03.

• The KOSMOS is a folding e-bike. It appears to use ISO 406 wheels, and a front centerless/hubless wheel fitted with an airless tire.

• Why use centerless/hubless front wheel? KOSMOS says the goal is to improve over a hub-type wheel with

  • “drastically” lower bike weight;
  • less air drag, for “vastly” improved speed;
  • a lower center of gravity.

• It appears to use a “support hoop” wheel structure [Wheel structure types].

• There are no details on the bearing arrangement [Bearing layouts] or bearing seals [Bearing seals].

• The centerless/hubless wheel uses a rim brake [Brake types].

• Since it is only a front wheel, there are no drive issues.

• Some photos show a truncated fender, which is unlikely to work as well as a regular fender. A KOSMOS accessory fender has stays and looks to bolt to the structural hoop, so a good fender could be used instead. There is no obvious provision for a front rack. More: [How do you mount fenders and racks?].

• A centerless/hubless wheel associates a new maker with a long history of suspect and bad claims from prior makers; anybody looking at a new centerless/hubless wheel may assume the worst. Thus, a new maker can probably help themselves and their customers by providing relevant data even before they are asked.

  Unfortunately, KOSMOS is asking for money, but does not give measured, computed, or even forecast data on their wheels (as of 2025-03). Little information plus asking for money probably reminds people of [Cyclotron]. People were suspicious of Reevo for similar reasons. By asking for money and not providing wheel data, KOSMOS probably scares away many people who would otherwise be interested.

• Ordinary hub-type wheels are often light. KOSMOS claims their centerless/hubless wheel makes the bike “drastically” lighter, but provides no further information.

  • An ordinary ISO 406 hub-type wheel could be under 500 grams without using exotic parts. (E.g., Velocity A23 290 grams, Bitex front hub 80 grams, Sapim Laser 28x at 4 grams/ea., Sapim Polyax Aluminum 28x at 0.35 grams/ea.)

  • For a centerless/hubless wheel, a common 406 rim might be 300 grams, and a support hoop might be another 300 grams. That is already 100 grams heavier, even without bearings. So it seems likely a KOSMOS front wheel is heavier than an ordinary hub-type wheel. Especially as the KOSMOS wheel is likely to be expensive, so there is an opportunity to up-spend on a lighter hub-type wheel [A fair comparison].

  • “Drastic” is vague — they could claim 10 grams saved from the bike is “drastically” lighter. Most people would probably disagree that is “drastic”, but then what is drastic?

    Suppose “drastic” is 10%. The bike’s claimed weight is about 22,000 grams, so a 10% drop in weight is 2,200 grams. KOSMOS seems to claim their hubless/centerless wheel weighs less than a 500-gram hub-type wheel, so the most it could save is 500 grams. Thus it could not save 2,200 grams.

    Broadly, the claim seems unlikely, and KOSMOS does not provide data, such as wheel weight, to support their “drastically lighter” claim.

• KOSMOS claims their wheel has less air drag, but gives no test data or other support for the claim. Aerodynamics is tricky, and things which look more-aerodynamic are sometimes worse. KOSMOS also claims better aerodynamics makes the bike “vastly” faster. But again, no supporting data.

  On this bike, even a truly aerodynamic wheel is likely to have only a minor speed benefit, so little useful speed benefit. Factors include:

  • It uses airless tires, which increase rolling drag and make the bike slower [Airless tires]. Going slower reduces the benefit of aerodynamic improvements.

  • Even if a KOSMOS wheel has better aerodynamics, prior centerless/hubless wheels such as [Reevo], have had high bearing drag. KOSMOS does not address rolling drag, so a reasonable assumption is that the KOSMOS wheel has high bearing drag, like the Reevo. High drag could cancel any aerodynamic benefits and make you slower overall.

  • Consider rider+bike air drag. Even with a rider in an “aero tuck” position, rider air drag is much bigger than bike air drag. And wheel drag is only one part of a bike’s air drag. The KOSMOS rider position is upright, which significantly increases the rider’s air drag. In turn, a wheel’s air drag is an even smaller part of total drag.

  • The bike is aimed at urban riders, who will mostly wear “street” clothes, with worse aerodynamics than bike racing clothes. Similarly, riders will often have luggage, which adds air drag. More rider and luggage drag further reduces a wheel’s effect on drag.

  • This is an e-bike, so average speeds are probably higher than a pedal-only version of the bike. That increases power going to wheel air drag, but also power going to the rider air drag. Wheel drag as a fraction of all aerodynamic drag stays small, even at higher speeds.

  KOSMOS says the wheel makes the bike “vastly” faster. The word “vastly” is vague

  • KOSMOS could claim saving one second from an hour ride is “vastly” faster. However, I think many people would say 50% faster (150% the speed) is vast; while 10% (110% the speed) is faster, but not “vastly” faster.

  • Here is a simplified example of potential benefits from a wheel with less air drag. Suppose a hub-type wheel is 10% of all bike+rider air drag. Suppose the KOSMOS wheel is perfect and has zero drag, so using it reduces total bike+rider drag from 100% to 90%. Air drag power is cubic in speed — going 2x as fast, power goes up by 2 × 2 × 2 = 8x. Suppose all drag is air drag (ignore rolling drag): if air drag goes down to 0.90, speed goes up by (1/0.90)^1/3 ≈ 1.036 or about 3.6%, which would save you 2 minutes in an hour.

  • Again, “vastly” is vague. My guess is most people would say “3.6% faster” is much less than “vastly” faster. 3.6% is based on several assumptions, and is likely wrong. But one assumption to get to 3.6% is the centerless/hubless wheel has no drag at all — since a real wheel has drag, it could be the actual speed improvement is much less than 3.6%.

  This example shows the value of specific claims. It also suggests people should be careful about how seriously they take all claims: consider what actual evidence is presented to support each claim.

• If you reduce weight near the road, then the average height of the remaining weight — the center of gravity — moves up. KOSMOS claims their wheel lowers the bike’s center of gravity. That seems at odds with their claim that the wheel is lighter.

  A similar comment applies to wheel+fork or wheel+unfork.

  A high center of gravity tends to help stability: a penny-farthing or other tall bike can be easier to balance at low speed, while a recumbent is often harder. KOSMOS claims a lower center of gravity helps riding stability. A low center of gravity tends to help stability when you are stopped; and tends to help maneuverability; but at the expense of riding stability.

  Usually, a lower center of gravity makes it harder to keep your balance while riding. KOSMOS claims “And thanks to the reduced center of gravity, KOSMOS Hubless is easy to keep your balance, even for beginners.” This seems contradictory.

  KOSMOS claims: “The reduced center of gravity has an inverse effect of increasing traction, making turns much easier.” I do not see how the center of gravity affects traction. Also, traction is not normally an issue for making turns. Also, claiming it makes turns easier seems to contradict their previous claim that it helps stability.

  The bike+rider are much heavier than the difference in center of gravity between hub-type and KOSMOS centerless/hubless wheels. Thus, both wheels probably have good handling — I see no reason a KOSMOS wheel will hurt handling. But contradictory claims here leads me to wonder what else may be wrong? Either wrong claims, or things done wrong on the bike.

• What has KOSMOS done to demonstrate the safety and load limit of the wheel? Since their centerless/hubless wheel is new, potential buyers will wonder: is it inferior to conventional hub-type wheels? Some questions may include:

  • If you ride through puddles, etc., can grit suddenly jam the front wheel? More broadly, a new wheel design can have new ways to fail. Things which slow you down are annoying, but things which could throw you are dangerous.

    Conventional hub-type wheels can jam due to stick-in-spoke kinds of events. The KOSMOS wheel gets rid of that risk. Are there new ways to suddenly jam? And if so, what riding makes that most likely — so when should a rider be most careful?

  • What is the weight limit? An answer probably involves testing, in which case potential buyers probably want to know how it was tested, what were the results, and how that compares to an ordinary spoked wheel.

    And since the fork and fork/wheel mount are new for the KOSMOS wheel, they should be included.

• There is no information about the front wheel bearings or seals.

  In all-weather urban riding, it is likely the wheel will occasionally roll through a deep puddle, so the lowest part of the bearing will be submerged in dirty water and exposed to other junk. In such use, if the bearing is not sealed, it will probably wear out quickly. If sealed, there may be a lot of drag — maybe enough to significantly reduce range and/or top speed compared to a similar e-bike with hub-type front wheel.

  Hub-type wheel bearings are often quick and fairly cheap to replace. Some centerless/hubless wheels look easy [Black Hole], but others look labor-intensive and thus expensive to replace [Reevo]. Some use bearings which are probably expensive. KOSMOS does not address wheel bearing durability or replacement cost.

  KOSMOS also notes the warranty does not cover “conventional” wear items. It is possible that fast-wearing wheel bearings are not covered by the warranty.

  Since the front wheel bearing is not a standard bike part, it is not well-tested nor well-understood like a hub-type wheel. Since it is not standard, if you have problems, you cannot simply buy another maker’s wheel or parts. People considering the bike should probably be concerned about durability and cost.

• Centerless/hubless wheels can have much more rolling drag than hub-type wheels. KOSMOS does not address this. It may be KOSMOS wheels are similar to other high-drag centerless/hubless wheels.

  KOSMOS is an e-bike, which can compensate for wheel drag by using more motor power. However, if wheel drag is large enough, it can hurt acceleration, top speed, and/or range.

  In turn, people considering the bike should probably be concerned about overall wheel drag.

• KOSMOS claims that their centerless/hubless wheel “helps with sudden stops and faster, better riding.” They do not explain how or why.

  This highights a general issue: potential customers should probably be concerned about all products offered by sellers making unsupported claims [Unsupported claims].

• The centerless/hubless front wheel uses a rim brake, which means the rim is a wear item. On the [Reevo], the rim is separate from the rest of the wheel, and so can be replaced without needing to replace other expensive parts.

  KOSMOS does not explain what is needed when the rim wears out. Potential customers should probably be concerned that the wear surface is integral, so a worn rim requires expensive custom parts; and if you cannot get them from KOSMOS, there may be no other source, making the bike unridable.

• Both rim brakes and disk brakes can be good choices. A summary of advantages and disadvantages of each is at [Brake types].

  KOSMOS uses a rim brake on the centerless/hubless front wheel. One reason may be standard disk brakes cannot be fitted, so mounting a disk brake would need more development [Brake types]. However, KOSMOS instead says a rim brake “makes the brakes much more reactive on dry pavement.” The term “reactive” is not standard; they do not explain what is or why you want it. They also do not explain what you lose on wet pavement.

  KOSMOS has a disk brake on the rear. When the bike is folded, it is inside the fold, so is protected from getting bent in handling the folded bike. There is still a risk of damage when e.g., parked at a bike rack, but it is the same as other bikes with disk brakes.

  KOSMOS writes the mixed brake types “gives you the best of both worlds”, although it seems reasonable to argue it also gives the worst of both worlds.

  They also write “On the heavier back wheel, a more powerful brake is used”. That choice seems odd: under heavy braking, the front wheel carries more weight, does more of the braking, and so needs to dissipate more power. Many bikes with front and rear disk brakes use a larger front rotor — giving more leverage, and more surface area so it can dissipate more power.

  Using mixed brakes is reasonable. It is somewhat common for riders to use a front disk brake and rear rim brake — sometimes because the frame lacks a disk brake mount; but even with a rear disk mount, people sometimes use a rim brake to save weight or avoid warped-rotor maintenance hassles.

  A front rim brake is a reasonable choice, but some riders may prefer a more-predictable front disk or drum brake. That would be easy with a hub-type wheel. I see no way to mount a disk or drum brake on the KOSMOS centerless/hubless front wheel. That contradicts the KOSMOS claim “All of the Pros of a Hubless Bike With NONE of the Cons.”

Links:

• https://www.kickstarter.com/projects/451828310/the-worlds-first-hubless-foldable-e-bike as of 2025-04]

  • All of the Pros of a Hubless Bike With NONE of the Cons

    Peddle [sic] assist

    Hubless bikes are known for their revolutionary design and futuristic technology. But the high costs, complexity, and massive bulk of these bikes make it hard to offer these bikes at an affordable price.

    The KOSMOS Hubless series solves all of these issues and more by being both lightweight AND affordable.

    Why KOSMOS Hubless?

    Hubless Front Wheel

    The front wheel of KOSMOS Hubless is made with NO spokes and NO hubs. This structure reduces the bike’s weight and air resistance for faster, smoother riding. And with a lower center of gravity, you’ll feel much more secure while you ride.

    The reduced center of gravity has an inverse effect of increasing traction, making turns much easier.

    This design feature also helps with sudden stops and faster, better riding.

    And thanks to the reduced center of gravity, KOSMOS Hubless is easy to keep your balance, even for beginners.

    This futuristic take on the hubless design is the first of its kind. Not only is the minimalist design visually striking, but the high-tech aesthetic makes this bike perfect for city riding.

  • By taking out the hubs and spokes, we managed to drastically reduce the bike’s weight.

  • Efficient Brake System

    The combination of a V break [sic] for the front wheel and a disc brake for the back wheel gives you the best of both worlds.

  • Front Wheel V Brake

    This design feature not only contributes to making the bike feel lighter, but it also makes the brakes much more reactive on dry pavement.

    Back Wheel Disk Brake

    On the heavier back wheel, a more powerful brake is used for stable braking power even at high speeds. Whether you’re in rain, snow or mud, these breaks [sic] work just as well in all conditions. And the even distribution of heat keeps these brakes strong even after the passage of time.

  • Finally, we’ve perfected the KOSMOS Hubless series: a peerless collection of bikes that takes all of the pros of hubless bikes and combines them with state-of-the-art safety, convenience, and aesthetic features.

  • The hubless wheel makes bike riding smoother than ever, [...]

  • *Damage to tires, brake pads and other parts that typically suffer from conventional wear and tear are not covered by the warranty.

• [https://kosmosbike.com as of 2025-04-02]

  • Revolutionary Hubless Front Wheel

    The hubless wheel’s reduced weight and air resistance vastly improve speed and the overall riding experience. And with a reduced center of gravity, this bike is exceptionally well-balanced.

  • Efficient Brake System

    The combination of a V break [sic] for the front wheel and a disc brake for the back wheel gives you the best of both worlds.

Nulla

        

Notes:

• This is as a design concept. The purpose of the centerless/hubless wheel is looks/sytling.

• Tuvie writes:

  However, it is not sure whether you will have a safe riding with this [...]..

  They were actually writing about the seat, but I think the comment applies to the bicycle as a whole.

• This is “only” a design concept, but I do not see the point of a design concept for something which will not work. The wheels are not supported against all loads. For example, if you put weight on the handlebars, the fork crown will rub on the front tire, and the front tire will jam against rotation.

  It can make sense to have a design where “We don’t know if this can work”, but it seems pointless to have a design using things which cannot work.

• There are other doubtful parts of the design. For example, the rims are structural, but have about the same cross-section as a conventional lightweight rim. A conventional lightweight rim will collapse if you put weight on it without tight spokes to support the loads. Convential rims are aluminum, but even high-strength materials are unlikely to be strong enough to fix it.

• Braking does not seem to be considered.

• Things like this can serve a useful purpose, such as a study in what is bad (could work, but not well) or impossible (riding with a jammed front wheel). Try to fix problems with this (put a roller on top of the front wheel so it won’t jam; use deeper-section rims), re-evaluate for new problems (the roller has high friction, the wheel will collapse if there are side loads) and repeat. Once you solve enough problems, compare to a conventional wheel. A process like this can help explain much more specifically what are the problems with centerless/hubless wheels, and why conventional wheels are conventional.

  But... they do not seem to have done that.

• I included this because it seems unworkable, and undesirable even if you could — see Tuvie’s comment about the seat. Yet it seems to have gotten at least some attention.

  I think it helps demonstrate a bigger point about why centerless/hubless wheels keep attracting interest. And why unworkable [Cyclotron] and workable but bad [Reevo] designs are able to raise money, apparently without backers spending much time to understand what is hard, and whether the design might have actually solved the problem, has just papered over it, or it is a scam and there is no design.

  Second, I think it shows the value of constructive criticism. That is, why something is bad, or hard. Many comments online — about everything, not just centerless/hubless wheels — are along the lines of “That’s dumb”, “It’s a bad idea”, or “It’ll never work”. Rather than “That’s dumb because ...” and then explain why, often a combination of (a) it does not solve the problem; (b) it introduces serious new problems; and/or (c) it claims to solve a problem which is hard-to-solve, and they have not demonstrated they really solved it.

  I hope that the Nulla — a design concept getting treated as an engineering proposal — is obvious enough to help introduce readers to analysis of more complicated “seems real” cases.

• [https://www.tetongravity.com/forums/showthread.php/131530-a-bike-without-spokes-nor-a-chain as of 2025-03]

• [https://www.tuvie.com/nulla-minimalist-and-stylish-bike-concept as of 2025-03]

• [https://www.arch2o.com/nulla-bike-a-minimalist-bike-concept-bradford-waugh as of 2025-03]

• [https://mmminimal.com/nulla-bike as of 2025-03]

• [https://www.core77.com/posts/4346/nulla-chainless-hubless-bicycle-concept-4346 as of 2025-03]

• [https://web.archive.org/web/20090302100814/http://bradfordwaughdesign.com/index.php?/project/nulla-bike as of 2025-03]

  Despite the attention attracted, the Nulla bicycle does not have a solid backer.

  Snarky comment: dumpster fires often get a lot of attention, but with a few notable exceptions, not a lot of solid backers. Social media promotes a lot of dumpster fires, so I guess it is easy to confuse “attention”, “interest”, and “value”.

• [https://web.archive.org/web/20100322220429/http://bradfordwaughdesign.com/index.php?/project/nulla-bike as of 2025-03]

  Bradford is working on possible mechanical and design updates to this design.

  Snarky comment: better than working on impossible udpates!

Reevo

        
from [https://uncrate.com/reevo-hubless-e-bike as of 2025-03] [https://mikeshouts.com/reevo-hubless-electric-bicycle-indiegogo as of 2025-03] [https://www.designboom.com/technology/reevo-hubless-electric-bike-09-23-2020 as of 2025-03]

              
screen captures from [https://www.youtube.com/watch?v=MgPUpccQ_mw as of 2025-03]

Notes:

• The Beno Reevo is an e-bike with centerless/hubless wheels. It was sold by Beno from around 2022 to around 2024. They claimed over 1,800 delivered; others have estimated 100-200 delivered.

  Online reviews say Reevos were unreliable. Also loud, slow, heavy, and uncomfortable. But poor reliabiliaty makes the other problems much less important: if it won’t go, it does not matter how loud or slow it was when it did go.

  Reliability problems included electronics and other things not related to the centerless/hubless wheels.

• The goal of using a centerless/hubless wheel is looks.

• The wheel structure is “support hoop” [Wheel structure types].

• It uses “rollers-in-rim” bearings [Bearing layouts].

  It appears from photos that Reevo’s rollers are

  • Cartridge ball bearings surrounded by rubber. That is similar construction to wheels used in inline skates, saketeboards, and so on. It appears from photos the wheels are about 30-40 mm diameter and maybe 10 mm wide. That is much smaller than standard skate/etc. wheels.

  • Built using shielded rather than sealed bearings [Bearing seals]. Shields are non-contact, which eliminates seal drag. But: it also eliminates some bearing protection against dust, mud, and dirty water, and other things which can damage the bearings.

• It uses rim brakes. Various online comments reamark they are significantly worse — longer stopping distance — than ordinary rim brakes.

• It uses a ring gear to drive the rear wheel [Drivetrain types].

• There do not appear to be mounts for fenders or racks.

• Some videos show Reevo wheels with a lot of side-to-side play. It is unclear what is the source of the play.

  • It could be design — all Reevos have it, because the design requires play, or because it specifies looser tolerances, which have play.
  • It could be manufacturing — some Reevos came with play, others did not.
  • It might be wear — new wheels are “tight”, but after use develop play.

• The rims have 17 rollers, which run in a track on the support hoops. The track forms a channel. Since it is on the hoop, it faces outward. Thus, any dirt which gets on to the track will tend to fall off [Bearing layouts]

  The rollers look like a urethane rubber tire around a standard cartridge bearing. The cartridge bearing appears to be shielded, not sealed [Bearing seals]. Using shields instead of seals means the bearings would likely be contaminated with grit and water after, say, rolling through a deep puddle, which would mean short bearing life. However, it appears Reevos had other reliability problems which limited how much they could be ridden, thus limiting the number of bearing failures.

  Online reviews say the wheels were loud and had high drag. Seals instead of shields would probably increase drag significantly.

• Reevo uses rim brakes.

  Online reviews say the brakes were weaker than other common rim brakes, and that the Reevo takes a long distance to stop, even when the brakes are dry.

  Some videos online show the wheels with a lot of side-to-side play. That would lead to pad rub with high-leverage brakes. I wonder if Beno fitted the Revo with low-leverage brakes, to increase pad travel and avoid pad rub when the brakes are released? I have no evidence to support this, but it would explain why the brakes are weak, even though higher-leverage brakes are available.

  The Reevo uses an off-the-shelf rim, screwed to the rotating wheel parts. The Reevo uses rim brakes, so the rim is a wear item. A separate rim allows it to be replaced without replacing other parts.

• The Reevo rear wheel uses a ring gear cast in to the rim, and a spur gear attached to the support ring to drive it. [Drivetrain types]

  Online reviews say the drivetrain is much louder than other e-bikes, loud enough to make Reevos annoying to ride.

  The ring gear runs close to the ground, so will likely be contaminated after running through, say, a deep puddle. Contaminants could cause faster gear wear, especially since the ring gear is aluminum. First, aluminum is soft, so more kinds of grit can wear it quickly. Second, aluminum wear particles turn in to aluminum oxide, which is abrasive: a lot of sandpaper uses aluminum oxide as the abrasive grit. So even without grit intrusion and even with good lubrication, there might be fast gear wear, and grit intrusion would make more wear debris, thus make more aluminum oxide, and thus make wear even faster. [Ring gear wear when dirty]

• The Reevo uses a single-speed drivetrain.

  Online reviews say it was too high of a ratio to climb normal hills without electric assist; and too low to provide useful pedal power when riding on flats with electric assist, despite the Reevo being slower than many other e-bikes.

• Reevo’s advertised weight was 25 kgs (55 pounds). Online reviews claim the actual weight was 30-35 kgs (65-75 pounds), although I have not seen where anybody weighed one.

  It is not clear how much of the bike’s weight is from the centerless/hubless wheels. They are clearly heavier than standard hub-type wheels; but other parts of the Reevo are also heavy. And it is an e-bike, so has battery and motor weight. On the other hand, it has single-speed gearing, lightwieght rim brakes, and no suspension.

• Although Reevo was a product, I do not find any suggestion “beno” (and variants) or “reevo” (and variants) were trademarked or otherwise protected.

  That seems odd, at least if they expected the product to be successful.

• How many bikes were delivered to paying customers? Several places online claim between 100 and 200 of were made and delivered, but I have not seen sources or an explanation for this — it might be a bad estimate which gets repeated, or maybe it is right. The Reevo Delivery Update #19 claims many more:

  [https://www.indiegogo.com/projects/reevo-the-hubless-e-bike#/updates/all as of 2025-03]

  We want to share that we have delivered a total of 1,888 bikes so far.

• Reevo’s Indegogo page says their wheel is tested as safe to 265 lbs (120 kgf).

  Since the centerless/hubless wheel is a new thing, anybody buying it should be concerned about wheel safety. So it is great they tested. However, the simple claim “265 lbs” should leave you with at least two questions:

  • What is the test protocol, and what were the results?

    “How you test” has a big effect on the results.

  • How does that compare to a standard spoked wheel in the same test?

    Since we know spoked wheels, that is a good starting point for comparison.

• Reevo has high rolling drag. It is an e-bike, which can cover up drag to make a bike feel good anyway. However, drag can hurt e-bike range, acceleration, and top speed.

When the wheels are unloaded (hanging free), they have a lot more rolling drag than an ordinary bicycle wheel. The spin-down demo in Berm Peak’s “Teardown” video suggests to me it could be 100x or more. First the Reevo wheel stops much faster. Second, the Reevo’s airless tire probably weighs more than the bicycle tire, so as a flywheel has more energy — meaning the Reevo’s bearings are even worse.

With a load, the drag difference between ordinary and Reevo wheels may be much larger or much smaller. The Reevo’s rollers use cartridge ball bearings, which should have similar load-versus-drag to a bicycle’s hub bearings. However, the Reevo’s rollers look to be covered in urethane rubber. Rubber can be quieter and more forgiving of dirt than metal-on-metal. However, rubber has higher rolling drag than clean metal-on-metal. Drag under load for the rollers is thus probably much higher than just the bearing drag.

Even a small amount more drag can be important. As a rough estimate, hub bearings may dissipate about 1 Watt in normal riding (both hubs). If a big version of a hub bearing has 20x more drag [Drag/Friction, above], that is 20 Watts. If rubberizing the rollers doubles that, it is 40 Watts. If a Reevo’s average power is 200 Watts, then the same speed with hub-type wheels would only draw 161 Watts. In turn, the Reevo’s range could be about 20% less than it would be with hub-type wheels.

Those are all estimates, but should at least give a sense of the problems being demonstrated by the Reevo.

• Reevo uses airless tires. Possibly for no-flats reliability. Possibly because it is harder to fit pneumatic tires on a centerless wheel — where do you put the valve stem?

  Airless tires have a lot of drag [Airless tires], so may contribute to complaints about Reevo’s top speed and range. In other words, only some of the blame is with centerless/hubless wheels.

  Pneumatic tires can be fitted to centerless/hubless wheels, so problems with airless tires are not an inherent flaw of centerless/hubless wheels. However, it appears there is no way to mount pneumatic tires on the Reevo’s wheels.

• The rear rim has a built-in gear. The teeth face inward, called a “ring gear”. A small spinning “spur” gear on the frame drives the ring gear [Drivetrain types].

  Revo drives the spur gear through many stages of other gears, which probably hurts efficiency and adds noise. However, high losses in the airless tires and centerless/hubless wheels are probably large enough to hide the gearing losses.

  I did not see claims about durability, but my guess is it wears out as fast or faster than a conventinal chain or belt [Ring gear wear when dirty]. Chains and belts are often forgiving of misalignment and dirt; gears tend to be more seneitive. Part of the ring gear sits close to the road, meaning it is likely to get dirty — for example, the first time it is rolled through a deep puddle. The Reevo”s ring gear appears to be aluminum, which is softer faster-wearing than hardened steels normally used in gears and chains. Also, aluminum wear particles turn in to aluminum oxide, a common sandpaper abrasive. Wear might be reduced by occasionally removing dirty lubricant and replacing it with fresh lubricant, but the Reevo has no obvious way to do that.

• Wear is a problem, partly due to the cost of dealing with wear.

  Gears are more complicated and expensive to make than sprockets.

  Also, the gear is integrated with the rim, so when the gear wears out, you have to pay for both gear and rim, even if the rim is otherwise in good shape. With a bolt-on sprocket, you just replace the sprocket.

  So even if the gear rim was is mass production, it would be expensive to replace, compared to an ordinary sprocket.

  A worn chain drive needs to replace both sprockets and chains. Bicycle derailleur chains are often expensive, but a centerless/hubless wheel can use a single-speed chain. A new chain is an added cost, but it seems likely that it is still much less than the cost of a gear rim.

  A Reevo is much more complicated to disassemble than a conventional bike. Labor costs to replace a gear rim would be more than to replace sprockets, unless sprocket replacement also requires disassembly. Because it is complicated, many fewer riders would have the tools and skills needed to change it themselves and avoid the labor clost.

• [https://www.reddit.com/r/shittykickstarters/comments/ivk95y/comment/g76z9ms as of 2025-03]

  “[Reevo Hubless E-Bike] - Anyone know anything about this company? It looks suspiciously like the old Cyclotron scam with a few new bits stuck on.”
  jjreinem
  2020-09-18

  Reevo_bikes
  2020-09-30

  Our goal has always been to develop an e-bike that is sexy yet functional. Early on in the design process, we decided to build Reevo around hubless wheels to create that “wow” factor. That was not an easy feat. We have to purpose built entire electric drivetrain to negate the downsides of this design choice.

  I read that as:

  • Centerless/hubless wheels were used for looks, not functional benefits.
  • Centerless/hubless wheels have problems (disadvantages).
  • Beno/Revo thought — or at least claimed — they had compensated for the problems.

• My summary: Beno/Reevo used centerless/hubless wheels for looks. They acknowledged centerless/hubless wheels have problems, but claimed they had a design that would adequately fix those problems. The bike they shipped had several problems, including that centerless/hubless wheels made it slower and noisier. But problems with the Reevo being unreliable may have meant “It’s broken, so I cannot ride it” was even more important than “It is slow, loud, and uncomfortable, so I don’t want to ride it.”

• [https://web.archive.org/web/20201003103626/https://www.cyclingaddicts.com/reevo-hubless-e-bike-review as of 2025-03] “Reevo Hubless E-Bike Review — Scam? We Won’t Be Rushing To Purchase One” CyclingAddicts 2020-09-21

• [https://www.reddit.com/r/shittykickstarters/comments/ivk95y/reevo_hubless_ebike_anyone_know_anything_about as of 2025-03] [https://www.reddit.com/r/shittykickstarters/comments/ivk95y/comment/g76z9ms as of 2025-03]

  “[Reevo Hubless E-Bike] - Anyone know anything about this company? It looks suspiciously like the old Cyclotron scam with a few new bits stuck on.”
  jjreinem
  2020-09-18

  Reevo_bikes
  2020-09-30

  Our goal has always been to develop an e-bike that is sexy yet functional. Early on in the design process, we decided to build Reevo around hubless wheels to create that “wow” factor. That was not an easy feat. We have to purpose built entire electric drivetrain to negate the downsides of this design choice.

• [https://newatlas.com/bicycles/reevo-hubless-ebike as of 2025-03]

  “Hubless Reevo ebike pushes the limits of engineering ... and credulity”
  Loz Blain
  2020-10-06

  Hubless wheels, on the other hand, have basically one purpose in this application: to look cool.

• [https://www.indiegogo.com/projects/reevo-the-hubless-e-bike as of 2025-03]

  • $1,382,695 USD by 738 backers on Nov 5, 2020

  • Reevo’s hubless wheels are meticulously impact tested in our lab to be safe up to 265lbs. The wheels are triple-sealed from the elements for superb long term reliability.

• [https://jimmymacontwowheels.com/reevo-hubless-bike-reveals-over-blown-expectations-of-investors as of 2025-03]

  “Reevo Hubless Bike Reveals Over-blown Expectations Of Investors”, jimmymac August 19, 2022

• [https://www.indiegogo.com/individuals/24648046 as of 2025-03] Also seems to have done business as “beno+ TECHNOLOGIES”. Although “beno technologies” and “BENO INC.”, and “Beno incorporated” may be good alternatiave search terms.

• [https://web.archive.org/web/20240305110354/https://www.reevobikes.com/reevo as of 2025-03], Reevo bikes web site [http://www.reevobikes.com/reevo] archived 2024-03-05.

  Spokes Are So Last Year

  Turn heads...and corners with the sleekest e-bike on the road. Meticulously engineered for performance

  [...]

  We’ve reinvented the wheel

  And the future comes rolling in. No spokes, no transparent centre, no gimmicks — just pure engineering excellence. Our patented hubless wheel design is the first of its kind.

  The LED headlamps [...]

  Snarky comments:

  • No gimmicks? I think they mean: within the gimmick of the centerless/hubless wheel, there are no additional gimmicks.
  • “Pure engineering excellence.” I am reminded of an old joke: “He’s so smart, he can make anything work, no matter how bad an idea it is.”
  • “First of its kind” But, unfortunately, not the last.
  • "Engineered for performance": it might be argued that “poor performance” is, indeed, one kind of performance; and by engineering away good-peforming hub-type wheels, they have indeed “engineered for (poor) performance.”

  Serious comments:

  • They seem to give no reason for why use centerless/hubless wheels. That is odd, as it seems like one of the first questions people would have. When they do not answer “obvious” questions, it suggests either they are trying to hide something, or that they are confused.

  • Is there any mainstream lighting which is not LEDs? Why do they feel a need to point out “LED”? Because centerless/hubless wheels are so backwards, they are grasping at straws to convince you it is not entirely backwards? Or because the use of centerless/hubless lights might lead people to wonder if it also uses carbide lamps?

    This is a (very) minor point, but beware that, yes, folks lie; but also do things to distract you — so they can seem to say something, without actually having to lie and say it.

• [https://jimmymacontwowheels.com/reevo-hubless-e-bike-appears-to-be-a-massive-failure as of 2025-03] jimmymac January 20, 2025

• [https://www.youtube.com/watch?v=khw8kpt0SGA as of 2025-03] HOW LONG WILL THIS RARE HUBLESS BIKE LAST? - PROPER DURABILITY TEST! Sam Pilgrim Feb 16, 2025

• https://www.reddit.com/r/bicycling/comments/zah4yv/reevo_bike_finally_arrived/

• https://mrgadget.com.au/reevo-the-hubless-e-bike-review

• [https://www.youtube.com/watch?v=AB7pBrudFbg as of 2025-03]

  “This Rare Futuristic eBike is a Total Nightmare”

  Seth Alvo
  Berm Peak
  2025-01-19

  “From what I can find, of the 2,700 people who ordered a Reevo Hubless Bicycle, only about 150 actually got them delivered [...]”

• [https://www.youtube.com/watch?v=z6dXmQPlvUs as of 2025-03]

  “I bought the Cheapest Reevo Hubless eBike in the USA” Citizen Cycle [https://www.youtube.com/@CitizenCycle925 as of 2025-03], 2025-01-24.

• [https://www.youtube.com/watch?v=MgPUpccQ_mw as of 2025-03]

  “What’s Inside The World’s Worst eBike? Reevo Teardown”

  Seth Alvo,
  Berm Peak Express
  2025-02-09

• [https://www.youtube.com/watch?v=IbaJ97-_dOM as of 2025-04]

  This Trash Raised Over $6 Million
  penguinz0
  2025-01-22

  @Daddyoh94
  2 months ago [approximately 2025-02]

  They DEinvented the wheel

Ross

           

Notes:

• The goals of centerless/hubless wheels are (a) a more durable wheel, by getting rid of loose spokes; and (b) a safer wheel, by preventing clothes and body parts from getting tangled in spokes.

• The wheel structure is “structural rod” [Wheel structure types].

• It is a “large bearing” design [Bearing layouts], with two large bearings per wheel.

• No brake is shown; it appears a rim brake would work.

• It uses a face gear driven by a spur gear. The idea is broadly similar to a ring gear, although it differs in details and trade-offs [Drivetrain types].

• There is no discussion of fenders or racks.

• Large bearings are heavier and more expensive than hub-size bearings. It is not clear that two bearings per wheel has advantages over one bearing per wheel; see, for example [Top Secret].

  The bearings are relatively close to the rim beads. That might make them more sensitive to rim damage than a wheel which places the bearing(s) further away.

  The bearings run close to ground level, so may be contaminated by e.g., rolling through a deep puddle. Once contaminated, they are likely to wear quickly, and are likely very expensive to replace. Bearing seals likely protect against deep puddles, but also likely cause enough seal drag that riding speed is significantly reduced.

• The Ross drivetrain takes power around two 90° corners. Bevel gears and other angle drives can be much less efficient than normal in-plane gears. When bevel gears are loaded (that is: driving, not coasting) they also try to pry themselves apart. That requires a stiff-enough support to keep them engaged. The support can add weight. Bevel gears have more complicated shapes, so are also often more expensive than ordinary in-plane gears.

  For these reasons, bevel gear drives in bicycles have often been heavier, more expensive, and more fragile than chain drives.

  Several other centerless/hubless wheels use a ring gear. Ring gears also have ejecting forces.

  • However, a ring/spur engagement has both gears curve the same way. That makes the engagement more tangent, which reduces both sliding drag and ejecting forces. The bevel/face gears in Ross do not have the same curvature, which increases sliding drag and ejecting forces.

  • Ring gear drives can optionally use a larger spur gear, which helps make engagement yet more tangent. In Ross, width is limited, which limits the size of bevel/face gears — and thus the ability to solve problems by using larger gears.

• One of Ross’s goals for centerless/hubless wheels is better wheel durability, by getting rid of the spokes. The rationalle is spokes get loose and weaken the wheel, so a spoke-less wheel can be more durable.

  An alternative is a spoked wheel built so that the spokes stay tight. Spokes often loosen when the wheel is run close to overload. Spokes more often stay tight in a stronger wheel built with many spokes, which are light (stretchy), laced to a stiff rim, and tightened enough. A wheel built this way can be ridden hard for a decade or more with little or no maintenance. For a wheel built this way, service life is more often limited by external damage (hits something or gets hit) or by other wear-out (rim brake track wear, rim spoke bed cracking). More: [Ordinary wheels].

  Stronger and more durable spoked wheels can have slightly more air drag, add some cost, and add some weight — perhaps 300 grams per wheel. However, heavier hub-type wheels can compare favorably in many ways to centerless/hubless wheels.

• One of Ross’s goals for centerless/hubless wheels is better wheel safety: getting rid of the spokes makes it impossible for clothes and body parts to get tangled in spokes.

  One alternative is to use more spokes. Closer spoke spacing makes it harder for things to get tangled in the spokes. More spokes can also improve wheel strength and durability. More spokes adds some air drag; but a Ross wheel likely has significant bearing drag. For ordinary riding, an ordinary wheel with many spokes may have less total drag. One spoke and nipple is often around 5-6 grams, so going from a 32-spoke wheel to a 48-spoke wheel might add 90 grams per wheel. This is likely small compared to the added weight of a Ross wheel.

  A second alternative is some kind of wheel/spoke guard. Bicycles since the late 1800s have commonly used some kind of frame-mounted string, fabric, or sheet plastic guard to keep clothes and children’s feet out of the spokes. Guards are usually light, and if designed for it, can also reduce spoke air drag. It appears a Ross wheel is likely much heavier. Historical wheel/spoke guards are also cheap and do not have moving parts to wear out.

  A third alternative is a wheel-mounted guard, in effect turning a spoked wheel in to a disk wheel. Front disk wheels often lead to poor handling in crosswinds. A safety guard could be made air-permeable; that would reduce the aerodynamic advantage, but also reduce the sensitivity to crosswinds. Wheel-mounted guards are also light and can be cheap — although race-optimized guards are often expensive, trading cost for better aerodynamics and lower weight.

• I included Ross because it has goals beyond “looks good”. So it is a good example for how you might compare a centerless/hubless design to alternative designs which also address those goals.

Links

• [https://patents.google.com/patent/US6224080B1/en as of 2025-03] [https://patentimages.storage.googleapis.com/72/92/2a/182631c1ddfc47/US6224080.pdf as of 2025-03]

  Spokeless bicycle system
  United States Patent US6224080B1
  Bennett Ross
  Granted 2001-05-01

Sada

           
From [http://www.sadabike.it/en as of 2025-03]

Notes:

• The Sada is a folding bike. It uses centerless/hubless wheels with the goal of folding smaller/faster than with conventional wheels.

• The Sada front wheel is "structural rim"; the rear wheel is "structural rod", with a cantilevered rod [Wheel structure types].

  The Sada front wheel and Terpstra both use a “structural rim” approach. But there are significant differences; highlighting them may highlight the Sada’s specific design features.

  The Sada uses a different roller arrangement: the rim is flanged, with rollers on the tread side which reach around the tread to roll on the flanges. A third roller, inside, rolls on the inside of the flanges.

  The Sada roller and rim flange layout mean that radial loads are in-line with the rollers. Lateral loads are also mostly transformed to radial roller loads. The rollers have consistent diameter where they run on the rim flanges, so do not slide, and so do not have sliding drag. In contrast, the Terpstra’s roller-in-groove layout means radial loads press on the side of the roller, which has varying radius where the roller and groove meet, and so the Terpstra has scrubbing drag for radial loads.

  The Sada rear wheel uses three internal rollers. It puts a drive roller and the supporting structure near the ground contact point. Although the structural rod is cantilevered, the rim has very low leverage on the rollers, which should dramatically improve wheel strength, and also reduce roller/wheel drag.

• It uses “rim-on-rollers” bearings [Bearing layouts].

• It appears to use a fixed gear (no freewheel), and otherwise looks to have no brakes. It uses a friction drive, which is somewhat likely to slip in the wet, leaving the rider with no brakes.

• It uses a friction drive [Drivetrain types].

  A friction drive is somewhat likely to slip when wet.

  It appears to use a Schlumpf High Speed Drive (now made by Haberstock Mobility [https://www.drive-mobility.com/en/drives as of 2025-03]) to drive the sprocket faster than the cranks. That partly compensates for the rim drive’s low ratio. It still looks to have a lower-than-normal ratio (fast pedaling, low road speed).

• There is no discussion of mouting fenders or racks.

• The Sada proposal is a folding bike, and the wheels are thinner than ordinary spoked wheels, as there is no axle, no hub spacers, etc.. “Thinner” is a benefit for folding bikes, as the wheels stack smaller, so it helps the bike to fold more compactly. The Sada rims are flanged for roller support, so are quite a bit thicker than the tire — e.g., it appears a tire which is 25 mm wide needs a rim which is about 50 mm wide.

  Ordinary hub-type wheels have substanial thickness: wide-spaced flanges for wheel strength, and even wider at the axle to attach to the frame. Front wheels are often over 80 mm at the flanges and over 125 mm at the axle. Rear wheels with a single sprocket are often over 60 mm at the flanges and over 145  mm at the axle.

  Alternatives include:

  • Regular wheels, but without spacers. The rear sprockets can stay with the frame, as in RAS and later Cinelli Bivalent and Reyedel hubs, or, more recently, as in FUBi prototypes. The FUBi wheels are listed as 50 mm wide, which is only slightly wider than many ordinary tires, so they lay about as flat.

    The Sada rims are flanged, and appear 50 mm or wider. So may be even thicker than the Fubi’s. The Fubi appears to fold to similar dimensions as the Sada, although the Fubi may take longer to fold to this size.

    The center of the Sada wheels is open, which may be an advantage to stack other stuff when stowing the bike. The FUBi wheels are flat, but the centers are filled with spokes.

               
    From [https://laughingsquid.com/the-sada-bike-a-bicycle-that-folds-down-to-the-size-of-a-backpack] as of 2025-03 [https://www.indiegogo.com/projects/fubi-world-s-most-compact-bike as of 2014-05], [http://www.fubi.com as of 2014-05, 2018-05]

  • X-spoked wheels, such as Shimano’s “Sweet 16”. They require special rims. They are normally used for a wider bracing angle to reduce spoke count and/or improve wheel strength. But x-spoking could also be used to narrow the flange spacing without losing wheel strength. X-spoke wheels require special parts, but not as special as for the Sada wheels. X-spoked wheels may be slightly heavier than standard wheels, depending on the rims you would have used otherwise.

  • Cast or other one-piece “tri-spoke” wheels, often using a few spokes and made of cast magnesium or carbon fiber. With suitable bearings, the hub width can be reduced to the width of a normal narrow tire (e.g., 30 mm) and so the wheels can stow flat. These are heavier than spoked wheels, but usually well under 1 kg per wheel for carbon fiber versions and often under 2 kg per wheel for magnesium versions.

  It appears these alternative “stow flat” wheel choices would allow the Sada to fold almost as flat, but without needing centerless/hubless wheels.

  • The Sada wheels require width for flanges that engage rollers which rearch around the tire. That means a 50 mm wide rim cannot use a tire wider than about 25 mm. In contrast, the other wheel choices can use as wide a tire as the total width — e.g., a 50 mm tire for 50 mm total width. A wider tire can improve ride comfort.

  • The Sada wheels have flanges which are near the ground, and not “hidden” in the profile of the tire. Riding close to rocks, curbs, etc., may thus damage the flanges. Conventional wheels do not have that problem.

• [http://www.sadabike.it/en as of 2025-03]

  • Packed in a backpack

    This hubless bike features a frame that folds away for maximum portability. It can even be packed in a regular-sized backpack, with others objects!

  • No spoke!

    The design focuses on the bike’s space-saving advantages. Unlike most bikes, which usually get their strength from the tension in the spokes, the strength of the Sada Bicycle is put into the rim. Sadabike has spoke-less wheels!

  • Foldable

    Sadabike is a hubless bike which can fold up to the size of an umbrella with one movements, making it perfect for commuters who need to make lots of short trips.

Terpstra

        
From [https://www.startupselfie.net/2023/12/17/penn-e-farthing-hubless-wheel-penny-farthing-ebike as of 2025-03]

Notes:

• Electric non-pedaling penny-farthing built by Christopher Terpstra.

• The goal of using a centerless/hubless wheel is art (looks).

• The wheel structure is “structural rim” [Wheel structure types].

• It uses “rim-on-rollers” bearings [Bearing layouts].

• Only the front wheel is centerless/hubless; the front wheel has no brake.

  The rear wheel has a band brake, although it is not clear from pictures if it is connected.

  This is an art bike, so “no front brake” is not a problem.

• The front wheel has no drive.

The rear wheel is a conventional hub-type wheel and is driven by an electric motor.

• Although it is ridable, the front wheel is probably dramatically weaker than a spoked hub-type wheel. And, the rollers which hold the rim are at the wrong end of a very long lever — the rim. So either the wheel or the roller supports are likely to fail under a side load which is held easily by a spoked wheel.

  A weak wheel is fine for an art project, even though it is a problem for “daily driver” bicycles.

• There do not appear to be mounts for fenders or racks.

  This is an art bike, so not a problem.

• Centerless/hubless wheels are great when done for art.

• [https://www.startupselfie.net/2023/12/17/penn-e-farthing-hubless-wheel-penny-farthing-ebike as of 2025-03]

  Hubless-wheel electric penny farthing is a unique fusion of past and future, 2023-12-17.

The Q

        
From [https://odditymall.com/diy-hubless-fat-tire-bicycle as of 2025-03]

Notes:

• Bicycle using centerless/hubless wheels, built by “The Q”.

  It is an art/demo project, not meant as a prototype or product.

  Ridable, but not meant to be as fast, cheap, light, or durable as a regular bicycle.

• The goal of using a centerless/hubless wheel is art (looks).

• The wheel structure is “support hoop” [Wheel structure types].

• It uses “rim-on-rollers” bearings [Bearing layouts]. It appears the bearings are shielded, not sealed; and also some are welded in place, making service more difficult. Since this is an art bike, durability and easy/cheap service are not concerns.

• Only the rear wheel has a brake, it is a drivetrain brake [Brake types]. Since this is an art bike, one brake is enough.

• It uses chain drive to a large sprocket on the rear wheel, with an intermediate hub to step up gear ratios [Drivetrain types]. It appears to be single-speed. Since it is an art bike, one speed is enough.

• There do not appear to be mounts for fenders or racks. Since this is an art bike, that is not a problem.

• Centerless/hubless wheels are great when done for art.

• [https://www.youtube.com/watch?v=RMUrJhet-1M as of 2025-03-08] “Insane Hubless Bicycle” The Q 2021-06-22

• [https://odditymall.com/diy-hubless-fat-tire-bicycle as of 2025-03-08] “This DIY Hubless Fat Tire Bicycle Looks Incredible” Ryan 2021-06-23.

Thorpe

     
From [https://patents.google.com/patent/US436844A/en] as of 2025-03

[https://patents.google.com/patent/US436844A/en as of 2025-03]

Bicycle
Inventor Thomas J. Thorp
US Patent US436844A
Granted 1890-09-23

Notes:

• The goal is a treadle drive bicycle, with friction drive to the inside of the rear wheel.

  You might expect “why” is a goal which directly benefits a rider or owner — faster, more comfortable, more cargo, etc. However, patents often omit larger goals, and instead focus on things like “If your goal is to build a friction drive to the inside of the wheel, then one way is: ...”

  It is not explicitly stated why use centerless/hubless wheels. I suspect (a) so the friction drive can drive the inside face of the rim, without spoke interference; and (b) so the two treadles can be connected directly — otherwise, a connection would need to pass around the wheel to avoid going through the spokes.

• The wheel structure is “structural rim”. Optionally, the front uses a “support rod” support [Wheel structure types].

• It uses “rim-on-rollers” [Bearing layouts]. Some run on the inside of the rim, others run on the tread. The tires are presumably airless, as pneumatic tires were uncommon in 1890.

  The rollers are in-line with radial forces. They are flanged to resist side-loads.

• Only the rear wheel has a brake, it is a spoon brake [Brake types]. It was common at the time to use just one spoon brake.

• It drives using a friction roller on the inside of the rim [Drivetrain types]. It is single-speed, as were almost all bikes at the time.

• There do not appear to be mounts for racks.

  Each wheel has a “mudguard”, which is a scraper running on the tire tread. It is positioned before the rim reaches any rollers, so presumably reduces the amount of junk reaching them.

  The rollers are up high, not down by the road. However, the tire can pick up grit/mud/water/etc. from the road, and carry it up to the rollers. The mudguards are positioned to scrape the tire before it reaches any rollers.

  Aside: the front wheel mudguard is high enough that I suspect stuff scraped off the front wheel winds up on the rider’s feet. I note modern versions of this idea put the front wheel’s scraper/brush much closer to the ground. For example, the 2014 demo/prototype Teage × Sizemore “Denny” [https://www.thespoken.cc/oregon-manifest-2014-teague-x-sizemore as of 2025-03-15].

• I suspect the friction drive alone would make this bicycle inefficient, even by the standards of 1890s bicycles, which often used solid tires; bushings rather than ball bearings; and block chain rather than roller chain. However, friction drives tend to be inefficient. Modern rubber compounds (e.g., urethane) can have less rolling drag, but are still less efficient than many chains, toothed belts, gears, etc.

  Rollers on the tire tread probably also add a lot of drag. Good penumatic tires on smooth pavement can cost 10 Watts or more rolling drag for each wheel. Solid rubber tires tend to have much more drag. A roller on the tire is somewhat like adding another road contact. In other words, roller-on-tire probably doubles the drag.

Top Secret

        

From [https://www.tuningblog.eu/en/E-Scooter-ATV-Quad/hubless-e-bike-top-secret-497637 as of 2025-03] [https://www.topsecretebike.com/#projects2 as of 2025-03]

Notes:

• Prototype for an e-bike product.

• The stated reason for using centerless/hubless wheels is looks, not for functional reasons.

• The wheel structure is “support hoop” [Wheel structure types].

• Each wheel turns on a large bearing [Bearing layouts]. Some pre-production pictures suggest it is shielded, but not sealed [Bearing seals].

• The front wheel uses a large disk brake, with a special rotor but standard hydraulic caliper and levers. The rear wheel uses a standard disk brake as a drivetrain brake [Brake types].

• The rear wheel is driven via ring gear [Drivetrain types].

• There do not appear to be mounts for racks.

  Some prototypes show small fenders which may help protect the bike. They are too small to protect the rider. The fender mounts do not obviously extend to support a normal fender.

• Patent drawings show either a standard cartridge bearing, or a custom bearing; the bearing is almost the size of the rim. Videos online appear to show a custom bearing being made and spun.

  Videos online suggest that even without bearing seals, it has much higher free-running drag than a hub bearing. The video is not detailed, but may show a full-complement bearing; the patent also shows one with a bearing cage, which would reduce the number of balls, but might reduce rolling drag.

  Consider a large bearing for a centerless/hubless wheel, and a small-diameter bearing for a hub-type wheel. Suppose both are of similar design and construction. A large bearing may have less drag at the bearing than the small one. However, drag at the tire might still be higher, since a tire has low leverage over a large bearing, but 20:1 or similar leverage on a hub bearing.

• The lowest part of the bearing is much closer to the ground than a hub bearing. Thus, it is likely to be exposed to more dust/mud/dirty water/etc. than a hub bearing. Rolling the wheel through a deep puddle could submerge the bearing. Dirty water can carry in grit, as well as being corrosive. During the rainy season, riders may ride through deep puddles several times a day.

  One patent has a design to block debris from entering a bearing; it appears to be a simple labyrinth seal. Non-contact shields and labyrinth seals can keep out most dry grit. However, non-contact seals are much worse at resisting liquid flow compared to contact seals.

  Contact seals can provide much better protection. But seal drag can be a lot more — enough to reduce top speed and/or range; and enought to make the bike harder to ride when just pedaling, with no motor assist.

  More-frequent bearing service and/or more-frequent bearing replacement are also options. But these are extra work and cost compared to a standard hub-type wheel and bearings. Large bearings are likely to cost 100x or more the cost of standard hub bearings, so frequent replacement may add a lot to the cost of owning the bike.

  More on this at [Bearing seals].

• The patent shows a gear ring which is separate from the rest of the rim and wheel assembly. That means each part can be made of a material suitable for a specific job. Compare to the [Reevo], where the gear is integral, and so the various parts need to compromise on the material choice, which can lead to higher weight, faster wear, etc.

  The separate ring gear also means when it wears out, it can be replaced without replacing the rest of the rotating wheel parts. Compare again to the [Reevo], where the gear is integral; when it wears out, the other integral parts must be replaced.

• The rim is integral with other rotating parts of the wheel. Compare again to the [Reevo], which uses an off-the-shelf rim bolted to the rotating wheel parts. The Reevo uses rim brakes, so the rim is a wear item; a separate rim allows it to be replaced without replacing other parts. Since the Top Secret uses disk brakes, the rim is not a wear item.

• At least one online video appears to have sound from the bike being ridden, probably under electric power. It seems loud, but it is unclear if that is wheel noise, drivetrain noise, or something else. Note also that loud noises can be produced at low power. A loud wheel might mean a lot of lost energy, but it could be loud without using much energy.

  A loud wheel is a disadvantage for ride comfort, even if the power cost is low.

• The patent text says it intends to improve on prior centerless/hubless wheels; but does not suggest the result is as good as a hub-type wheel.

• The front brake uses a very large brake disk and a conventional hydraulic caliper and levers. The front wheel is held by a large bearing, not rollers. Thus, disk “wander” is probably less than for a conventional bicycle disk rotor.

• The rear brake is a conventional bike disk brake which brakes the rear wheel via the drivetrain. A possible disadvantage is that a drivetrain failure can also mean a brake failure.

This is an e-bike, so should be less sensitive to added drag than a regular bike. Also, drag is something you can notice in the first ride: either it is okay for your use, or it is not. Drag, noise, and so on are likely to be addressed in reviews.

• That said, I would be concerned about bearing durability. Durability is hard to evaluate in one ride or in a short-term review.

• The bearing is likely to be very expensive — 100x or more the cost of a hub bearing. So regular replacement may add a lot to the cost of operating the bike. Or, if replacement bearings are not available, it can leave you with an expensive but unridable e-bike.

• Further, your particular riding may have a big effect on bearing durability. Someboady who rides only when it is dry might have good bearing life, which might encourage you to get one. But if you ride all-season, your bearing life could be much less.

• It could be that all-weather bearing durability is good. However, there is a risk that it is poor, and that this makes the bike much more expensive or unridable. So this should be a concern for anybody buying it.

• Bearing durability is a general concern for all centerless/hubless wheels. I mention it here because (a) the Top Secret wheel bearings may be a lot more expensive than bearings for e.g., the Black Hole or Reevo; and (b) Top Secret seems likely to offer bikes for sale, whereas most other centerless/hubless wheels are either historical or have no clear path to market — in other words, it is unlikely you have a chance to buy the others, and thus less likely that durability matters to you.

• [https://www.tuningblog.eu/en/E-Scooter-ATV-Quad/hubless-e-bike-top-secret-497637 as of 2025-03]

• [http://jnbmobility.com as of 2025-03] [http://jnbmobility.com/bbs/board.php?bo_table=tl_gallery as of 2025-03] Address : 251 IMPERIAL HWY STE 411 FULLERTON CA 92835-1058 Contact : topsecret@jnbmobility.com

• [https://www.topsecretebike.com as of 2025-03-15]

• [https://www.youtube.com/@jnbmobilityInc./videos as of 2025-03-15]

• [https://patents.google.com/patent/KR102625329B1/en as of 2025-03]

  A hubless wheel for bicycles having a stationary wheel devided into two parts
  South Korean patent KR102625329B1, 2024-01-15
  Inventor 이재철

  English translation:

  [T]he present invention can provide the hubless wheel for a bicycle that conforms to low weight and low noise bicycles and has improved assembly efficiency.

  Hubless wheels with this structure have the advantage of a clean and simple appearance because they do not have a hub or spokes.

  However, most hubless wheels have the disadvantages of being heavier, louder, and less easy to assemble than traditional wheels.

  The present invention is an invention derived to solve the problems of the prior art described above, and seeks to provide a hubless wheel for a bicycle with reduced weight and noise and improved assembly.

• [https://patents.google.com/patent/KR102684730B1/en as of 2025-03]

  A hubless wheel having a improved structure
  South Korean patent KR102684730B1, 2024-07-12
  Inventor 이재철

  English translation:

  Depending on the structural characteristics of hubless wheels, the bearings placed between the fixed wheel and the rotating wheel are manufactured in relatively large sizes. Therefore, in hubless wheels, bearings are an important factor in determining the weight, assemblyability, and noise characteristics of the wheel.

  The present invention seeks to provide a hubless wheel with improved assembling and noise characteristics that reduces weight through improvements in bearings and their mounting structures.

• [https://patents.google.com/patent/KR102747704B1/en as of 2025-03]

  A hubless wheel providing high assembly strength and foreign object blocking structure
  South Korean patent KR102747704B1, 2024-12-27
  Inventor 이재철

  English translation:

  [I]ts main purpose is to provide a hubless wheel having improved assembly strength and a function of preventing foreign matter from entering.

Twist

              
From [https://inhabitat.com/jose-hurtadoss-extraordinary-hubless-twist-bike-can-be-turned-into-a-tandem] as of 2025-03

Notes:

• This is a design exercise. It appears the main theme is converting two single bikes to a tandem, and centerless/hubless wheels are for looks, not function.

  There does not seem to be an explanation how the tandem rolls over bumps without lifting wheels — and, likely, breaking things. It may be the design handles bumps; but it is not clear that it does. If the design is superficial on the main goal, it is probably also superficial about centerless/hubless wheels.

  So: we need to be be careful what we “learn” here.

• I included it because the front wheel structure appears to be a hybrid of “structural rod” and “support hoop” [Wheel structure types]. Notably, the front wheel’s structural rod reaches only somewhat near to the normal tire contact patch, so some wheel loads are still taken by the hoop.

  The more-direct “structural rod” load path should make this lighter than a pure “support hoop”, which supports all loads via a less-direct load path.

  However, the wheel’s center is no longer open. If an open center is key to (say) carrying cargo or even looks, the hybrid design may not be suitable.

• The quick-release fork could also be done with a hub-type wheel, and something like it has been done for prototype folding bikes using hub-type wheels. So that is not specific to centerless/hubless wheels.

• [https://inhabitat.com/jose-hurtadoss-extraordinary-hubless-twist-bike-can-be-turned-into-a-tandem as of 2025-03]

Ujet

           
From [https://www.ujet.com/case-study/in-wheel-motor as of 2025-03]

The Ujet scooter includes the “Ujet One" and “Ujet Founders Edition". Both are scooters, not e-bikes, so there are no pedals and no pedal drivetrain. However, the Ujet uses centerless/hubless wheels, so may be interesting/instructive.

Notes:

• The use of centerless/hubless wheels appears to be driven by the motor design and a goal of light weight, even if it means the price is higher.

• The wheel structure is “structural rim” [Wheel structure types].

• It uses a “large bearing” design [Bearing layouts].

• Both wheels use disk brakes [Brake types].

• It is a scooter, not a bike or e-bike, so has no pedal drivetrain [Drivetrain types].

• The body forms a cantilevered rack. The front fender is attached to the front wheel’s support hoop. The front fender is probably too short: it looks like some dirty road water thrown up by the front wheel will get on the rider. That said, a longer fender or mudflap would be easy to add using the current approach. The rear fender is integrated with the body/rack.

• It uses a motor in the rear wheel which has a large-diameter ring structure for the motor coils.

  The tires are roughly the diameter of BMX tires, so smaller than ordinary MTB and “road bike” wheels. In turn the motor’s “large-diameter ring structure” is still smaller than a typical bicycle wheel.

  They could use this motor design but instead have a wheel/rim structure which reaches inboard to a hub.

  • For example, spider attached to the rim close to where the disk brake rotor attaches. That would add weight for the spider, hub, and a longer stay from the scooter frame to the hub. It would eliminate the large-diameter wheel bearings, but they are probably much lighter than the the parts which would replace them (spider, hub, and stay).

  • Notably, the large bearings rest motor/wheel parts which are needed for both wheel types (centerless/hubless and hub-type), so getting rid of the large-diameter bearing does not lead to other savings.

• The front wheel is also centerless/hubless. It appears to be a styling decision, and may be heavier and more expensive than a similar hub-type wheel.

  The front wheel is small. The support hoop and wheel bearings are larger than for a hub, but they are much smaller than a similar design for a large bicycle wheel. In turn, the Ujet may have only a small weight penalty for using a centerless/hubless front wheel rather than a conventional hub-type wheel.

  With a front hub, a conventional hub-mounted disk brake could be used. However, given the (small) size of the Ujet wheels, and given the higher brake power needs of a scooter vs. a bicycle, you probably do not want to use a smaller disk brake rotor. So there is probably no weight savings from using hub-mounted brakes.

• The bearings appear to be large-diameter industrial cartridge bearings. Here, “large diameter” is still much smaller than conventional bike wheels. And since the scooter has higher loads than a bicycle wheel, a hub-type bearing would be larger than the 6001 bearing discussed above.

  • In turn, the amount of leverage that the wheel “gives up” is less than the 20:1 figure discussed above for bicycles. Thus, the bearing drag ratio between centerless and hub-type wheels is less than 20:1.

  • For similar reasons, the added seal area is less than for bicycle wheels. In turn, the added seal drag of a large bearing is less than the 400:1 figure discussed above. For example, if the bearing seal is 8x as large as a hub bearing seal, the seal drag would be about 8 × 8 = 64x more.

  Rolling through a deep puddle can submerge the bottom part of the bearing in dirty water. It would take something much deeper to submerge hub bearings. The Ujet bearing seals are not discussed. It seems likely to me that it uses normal contact seals. Contact seals should be adequate for ordinary submersion. See the next point for seal drag power.

• The Ujet’s motor is 3.5 kW, compared to often 0.75 kW or less for an e-bike. The Ujet’s battery is also larger than typical e-bikes.

  Since added bearing and seal drag is (probably) less than with larger bicycle wheels; and since the battery and motor are bigger, the added bearing and seal drag power is a smaller fraction of the bike’s motor power and battery capacity, compared to a typical e-bike.

• There are no pedals, so no drivetrain other than the electric motor, and thus no complications from that.

• It uses rim-mounted disk brakes. A scooter will have larger brakes than an e-bike anyway. The Ujet wheels are smaller than most e-bike wheels. So the Ujet adds little disk rotor size or weight compared to a hub-mounted rotor.

  The use of large bearings (rather than rollers) means the Ujet’s wheel should have little lateral play, and thus is adequate for a disk brake.

• The status is unclear: web pages suggest a shipping product, but I did not find reviews, sites selling them, parts or service/etc.

Links:

• [https://www.ujet.com/case-study/in-wheel-motor as of 2025-03]

• [https://www.moto2s.cn/news/24402.html as of 2025-03]

• [https://www-moto2s-cn.translate.goog/news/24402.html?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en-US&_x_tr_pto=wapp as of 2025-03]

• [https://www.youtube.com/watch?v=DswFPt5JRRg as of 2025-03]

• [https://www.youtube.com/watch?v=OPPyRjpqr2A as of 2025-03]

• [https://www.ujet.com/support as of 2025-03] Ujet Electric Scooter User Manual [https://assets.website-files.com/5eabe9b381573a3335dfacda/5ee8706e07de1d2bc3ac5ec5_5e4699429fcf751ca37abf45_USER_EN_05.12.2018-min.pdf as of 2025-03] tires 80/80x14 MC26 Capri Ujet Electric Scooter Specifications [https://assets.website-files.com/5eabe9b381573a3335dfacda/5f6de2d33c1f8c1ce9fc2705_5e56776b6725fd6ead31f53d_Ujet--specifications.pdf as of 2025-03]

• [https://www.youtube.com/watch?v=yulLLuuqXUs as of 2025-03]

• Ujet’s most-recent youtube video (as of 2025-03) is from 2019-08-23, which suggests it is not an active product/product. [https://www.youtube.com/watch?v=OwtKxzb4eAs as of 2025-03]

Yale demo

           
From [https://www.core77.com/posts/15986/yale-mecheng-students-build-bike-with-hubless-wheel-15986] as of 2025-03

Notes:

• This is an experimental bike. It is not meant as a path to a product.

  An experiment should have some goal — e.g., to see if something is feasible; or to measure drag, weight, etc.; or something else. It is not clear what was the goal for this project. It seems likely there is a project write-up, but I did not find a copy.

• Why a centerless/hubless wheel: in an online posting, one author wrote:

  [W]hy we went for the spokeless bike. First, it looks cool. Second, we only had a semester so we wanted to pick something that was both feasible and challenging. Also, you can do a lot of things with the space that opens up [...]. Finally, the fact that we couldn’t find pictures of a real spokeless bicycle online [...].

  This seems to be saying (a) for looks; (b) to see if building one shows it is useful for something, but we did not think in advance that it was useful; (c) to learn why centerless/hubless wheels are not common.

  I assume there is a project write-up, which might say more, but I did not find a copy.

• The wheel structure is “structural rim” [Wheel structure types].

  The rollers are spread almost half-way around the rim. That should make it stronger and help reduce roller drag compared to narrower spacing, such as the [Terpstra] or [Sada].

• It uses “rim-on-rollers” bearings [Bearing layouts]. The rollers run in grooves, see also the [Terpstra].

  Running in grooves is good for lateral loads. But a radial load — the rider’s weight — presses on the side of the wheel. The side of the wheel slides as it meets the groove, and slides more as it moves away, causing roller friction/drag and wear.

  I suspect that the roller drag makes it feel like the brake is always dragging. That may be fine for a prototype or e-bike, but not for ordinary riding. Drag also means wear. Ordinary dirt/mud/etc. can act as an abrasive to speed up wear. And worn-off pieces of aluminum can turn to aluminum oxide, which is a grit used in some sandpaper — so even without dirt/mud/etc., there will be abrasive wear. Thus, the rim and rollers may wear quickly in some kinds of ordinary riding.

• The centerless/hubless wheel does not have a brake. It is not discussed in web postings, but there is probably a class report, and it might discuss brakes.

• It uses a ring gear to drive the rear wheel [Drivetrain types]. It uses a 2-stage drive so the gear ratio is similar to a bike with hub-type wheels. It is a single-speed drivetrain.

  The ring gear is a toothed rubber belt attached to the rim. Toothed rubber belts for bicycles can have drag comparable to a chain drive. It is unclear how efficiency changes when a toothed belt is attached to the rim — it could get worse, better, or no change.

  Bicycle belt drives also have good durability — roughly similar to bicycle chain drives. It is unclear if attaching to a solid backing changes that.

  The belt drive runs closer to the ground than in typical bicycle use. The Yale bike puts the belt in the center of the rim, which should protect it more from dirt than a belt drive used like a chain drive such as [The Q] However, it will still get dirty if the bike (say) rolls through a deep puddle. It is not clear how that affects durability.

• There do not appear to be mounts for fenders or racks; they are not discussed in web postings. Again, there is probably a class report, it might discuss fenders and racks.

Other comments:

• Snarky comment: the headline “Yale students build spokeless bicycle in one semester, now looking for jobs.” made me think “Yeah, I wouldn’t hire them, either.”

  They may be great engineers, but the social media version of their project is unflattering.

• Of course it is fine to build “whatever”. For art projects, it desirable to do things which make people think. Practicality is not the issue, and impracticality can be a benefit.

  For engineering, it is often useful to quantify things — drag, cost, weight, durability, noise, etc. So you might build a bad one, to show how much you lose compared to the good one. Measurements of the bad one explain why it is worth spending extra effort/cost/etc. to make the good one.

  That said, many engineering school projects seem more like art: they may provoke you think, but provide few answers. For example, the Gafoor bike [Gafoor et al.] has no drag, weight, or other measurements; and almost no discussion of why you might want it; or even subjective evaluation of how the wheels performed. That’s too bad, because they had a working bike and could have answered lots of those questions.

  For the Yale bike, the builder’s online comment is disappointing, since it suggests no specific goals other than to build one. I hope they were interested in things like drag, cost, weight, and durability. Even if they thought centerless/hubless might be worse, it would be useful to have measurements: worse by 2x, 10x, or 100x? In which case, a reasonable answer to “Why?” is “To measure it and put numbers to how it compares”.

  For the Yale bike, there is almost certainly a write-up for the class. I did not find it. I did not look very hard, but that also reflects a problem: social media sites usually show as much eye candy as possible, then gloss over the most basic information — like where to get more information!

• [https://www.engadget.com/2010-02-17-yale-students-build-spokeless-bicycle-in-one-semester-now-looki.html as of 2025-03] Yale students build spokeless bicycle in one semester, now looking for jobs Richard Lai Senior Reporter Feb 17, 2010

• [https://www.core77.com/posts/15986/yale-mecheng-students-build-bike-with-hubless-wheel-15986 as of 2025-03]

• [http://cozybeehive.blogspot.com/2010/02/spokeless-bicycle-wheel-design-from.html as of 2025-03]

  [...] students form groups of 3-4 (or more) and under the guidance of an advisor will seek to realize a design idea. That may either be for an innovative new product or something which may offer genuine improvements to an existing product.

  That said, what would still be interesting is to learn the rationale behind choosing a spokeless wheel to work on for a project. Just for being cool or something else?

  [...] the more you try and fiddle with what exists, the more you sacrifice what’s already there in terms of design trade-offs. I suppose that’s the thing with design. It’s tricky. Gain some, lose some. Gain-gain is rare. What’s better in terms of taking loads, transmitting motion, and reducing friction - spokes, ball bearings and an axle or a two extra cranks, chains, and a timing-belt kind of setup?

  What I would have done if I were them would be to visit the nearest patent office to see what was done by individuals in the years past in spokeless wheel design. Were they successful? Did the idea SELL? Did it actually make someone’s life better?

• [https://www.reddit.com/r/technology/comments/b2g91/spokeless_bicycle as of 2025-03]

                         

  zhaolander 2010-02-15

  There were nine seniors total (all mechanical engineering majors). Our professor was actually a working full time senior mechanical engineer at Sikorsky.

  As you’ve probably noticed, only the back wheel is spokeless. This was done for several reasons: Manufacturing the rear wheel/rim was very expensive, so we only had one machined to see if it would work. We knew that if we could get it working for the rear wheel, it would definitely work for the front. But in the end, we ran out of time (class was only one semester long).

  Edit: Hey guys thanks so much for all the comments! I just wanted to elaborate on why we went for the spokeless bike. First, it looks cool. Second, we only had a semester so we wanted to pick something that was both feasible and challenging. Also, you can do a lot of things with the space that opens up where the spokes use to be. You can stick an electric motor in there. You can install some sort of gyro balanced storage basket. Finally, the fact that we couldn’t find pictures of a real spokeless bicycle online really sealed the deal.

  2nd edit: Also, this is the first prototype! We weren’t even thinking about suspension and all that stuff. If you notice, we only have one brake installed. All we wanted to do was prove the concept of a human powered spokeless bike.

• “You can stick an electric motor in there.” Many e-bikes already have motors in the wheel, how does centerless/hubless change that? A better motivation would be if centerless/hubless somehow improves on what is possible using a hub-type wheel. But they do not say that.

• “Gyro balanced storage basket” — Okay, you could, but why does anybody want that? How does that make it easier to carry anthing people actually want to carry? Again, how does it improve on a hub-type wheel?

• “We couldn’t find pictures” — in fairness, in 2010 there were a lot fewer pictures online.

  And while a patent search might have found Lew’s 1993 “Black Hole” patent, it is assigned to “Wear and Tear” which by 2010 was no longer operating.

Background

• Bearings

  • Bearing layouts
  • Bearing seals
  • Roller and track drag
  • “Big Bearing” Alternatives
  • Derailleur jockey pulleys
  • Roller layouts
  • Roller/rim wear
  • Safety of roller-supported rims

• Brakes

  • Brake types
  • Disk brake attached to a rim

• Drivetrain

  • Drivetrain types

• Structure

  • What is a centerless/hubless wheel, anyway?
  • Wheel structure types
  • Fork weight
  • How do you mount fenders and racks?

• Comparisons

  • Wheel aerodynamics are still incomplete
  • Airless tires
  • Speculation: why is hubless appealing?
  • Ordinary wheels
  • A fair comparison
  • Unsupported claims

Bearing layouts

Centerless/hubless wheels use at least three bearing layouts:

• Large bearing
• Rollers-in-rim
• Rim-on-rollers

• “Large bearing”

  The wheel turns on one or two large-diameter bearings. In effect, this is a hub bearing scaled up to about the same size as the rim.

  Often, only one bearing is needed: ordinary deep-groove bearings support both radial and side loads.

  Examples: [Ross], [Top Secret], [Ujet]

• Rollers-in-rim

  The rim as several or many rollers or “little wheels”. The little wheels roll on a track on the non-rotating part of the wheel.

  Example: [Reevo]

• Rim-on-rollers

  The non-rotating part of the wheel has several rollers. The rim has a track which runs on those rollers.

  Examples: [Black Hole], [Terpstra], [The Q], [Gafoor et al.], [Yale demo]

Is “hubless” just a big hub?

Digression: At the top, I wrote

One view is that there is no such thing as a “hubless” wheel: somebody enlarged a hub to almost the size of the wheel, then used a very large hollow axle. Put this way, the question is whether a very large hub has advantages over a standard-size hub.

One view is that roller-based systems did not simply “enlarge a hub”, and so they are truly hubless. That is reasonable view. But it also reasonable to say that most hubs use ordinary bearings, but you could also build a hub which runs on rollers.

For me, the interesting thing is “what can you build, and how well or poorly does it work?” Names (“hub”, “hubless”) are useful ways to describe things and help organize our thinking. But even when you and I use different names for a thing, the thing’s attributes (drag, cost, weight, durability) are the same.

Does dirt/mud/water fall out, or get trapped?

The [Reevo] rims have 17 small wheels, which run in a track on the support hoops. The track is a U-shaped channel. Since it is on the hoop, it faces outward. Thus, any dirt/mud/water which gets on to the track will tend to fall off.

However, the rollers in the rim sit in “pockets”. If you dip the front wheel in a deep puddle, the lowest few pockets will fill with water... but then gravity holds it there, and it does not drain out. That means more time for water to work in to the bearings and damage them.

Or, at least, that is how it looks in photos. Maybe there are drain holes, but (a) I did not see them, and (b) small drain holes tend to plug up over time.

Similarly, dirt and mud that gets in the pocket will tend to be held in by centrifugal force while the wheel is rotating. When it stops, it can fall out of the top and side pockets, but will tend to stay stuck in the low pockets.

A “rim-on-rollers” design does not tend to trap stuff around the rollers. The track on the rim could be U-shaped, an inverted-U, or something else. A U-shaped track will tend to trap dirt/mud/water, as above. A dirty track will wear faster than a clean one, and also wear the rollers faster.

In comparison, stuff that gets on an inverted-U track will tend to be thrown off by centrifugal force, and also tends to run off/fall off under gravity.

Why do hubs use two bearings?

Digression: you might wonder: since centerless/hubless wheels sometimes use a single bearing, why do ordinary hubs have two bearings? Answer: leverage.

• Small wheels can use a single bearing -- many derailleur jockey pullies use just one bearing. The side loads are relatively small, and the pulley diameter is small, so lateral loads are small enough that bearing drag and durability are good with just a single bearing assembly.

• I have ridden a prototype folding bike with a small front wheel that is supported by only one 6001 bearing. I doubt it would be durable in regular service. But it was a working bicycle wheel using a hub supported by one bearing.

• As a wheel grows in size compared to the bearing, lateral loads get more leverage on the bearing. In other words, if the tire diameter grows but the bearing does not, then a given side load at the tire causes more load at the bearing. You can address this by using stronger bearings, but that adds weight. Using two bearings gives the bearings more leverage against the load, so can give high strength with less weight.

• Using two bearings is also stiffer. That is, with similar total bearing weights, the one-bearing approach allows more hub and rim side-to-side motion. Using two bearings leads to less side-to-side motion both at the hub (good for disk brakes) and rim (good for rim brakes).

• Spoked wheels need the hub flanges fairly far apart in order to give good wheel strength against side loads. Rear wheel flange spacing is usually 50 mm or more, and front spacing often 80 mm. It is thus convenient to use two bearings, one under each flange. A single bearing in the middle would likely require more structure in the hub, and so would hurt cost and weight.

Bearing seals

           
From [https://pibsales.com/bearings/sealed-open-and-shielded-bearings-comparison] as of 2025-03, [https://www.machinerylubrication.com/Read/29452/using-labyrinth-seals] as of 2025-03

Standard cartridge bearings are often described as one of:

• Open — you can see the internals of the bearing.

  An open bearing has no protection from contaminants. But if used in an oil bath, oil can flow freely through the bearing.

• Shielded — a flat metal “washer” mostly covers the distance between outer and inner races. The shield is attached to the outer race and does not quite reach to the inner race. It is non-contact, so has no drag. Like an open bearing, it lets in debris (grit, water, etc.). Since the gap is small, it lets in less, and also keeps out grit bigger than the gap. Small things and liquids can still get past the gap between shield and inner race, and then damage the bearing.

  Shielded bearings are used commonly for indoors equipment where the bearings are not exposed to junk except for airborne dust. The shield reduces the amount of debris which gets in to the bearing, compared to an open bearing. A lot of airborne dust is soft enough that it does not damage the bearing. Some airborne dust is hard enough to damage the bearing, and too much dust can displace lubricant. Both hurt bearing service life, but bearings in this service often last decades.

  Shielded bearings are also often used in wheels for inline skates, because the have less seal drag than sealed bearings. However, when used in dirty conditions (e.g., rainy streets), bearings may fail after a few days use, due to grit and water in the bearings. As it fails, bearing drag goes up, the bearing gets rough and loose, and it may sometimes jam.

• Sealed — A rubber “washer” is attached to the outer race and has a rubbing contact on the inner race. The rubbing contact excludes most grit and water, although high-enough pressures (e.g., a pressure washer) can overcome the seal. Where the seal contacts the inner race, grit can get between the seal and the race and rubs on the rubber seal, wearing it away; over time, wear is enough the seals leak. Since the seal rubs on the inner race, it has seal drag. For a bearing carrying full load, the bearing’s inherent drag is larger than seal drag. But for a lightly-loaded bearing, seal drag may be much more than the bearing drag.

  Sealed bearings are used commonly on bicycles hubs, bottom brackets, suspesion pivots, derailluer pullies, and so on. One cause of sealed bearing problems in bicycles is that seals wear, grit and water gets inside, and the bearing is damaged by grit and/or corrosion.

In addition, many bearings use

• Labyrinth Seals — one part is attached to the outer race, another to the inner race. They overlap, but without touching. Usually, one or both are “shaped” so there is a zig-zag path to get from outside the bearing to inside the bearing. That is, there is no way to shine a light on it and have it shine directly inside the bearing.

  Two basic ideas of labyrinth seals are:

  • There is no straight path in to the bearing, so mud, grit, and dirt are likely to get stuck to the sides of the labyrinth, so do not get inside.

    Traditional bicycle hubs are often oil-lubricated. Oil is added to the middle of the hub, and a little leaks out through the labyrinth seals. The oil makes the labyrinth sticky, so even dry airborne dust is likely to stick without getting in to the bearing. Oil also helps to flush out trapped gunk, so it cannot come loose and work its way inside. A down-side is the oil makes some amount of sticky mess on the wheel, so you are more likely to get dirty if you handle it. Oil can also get on the brakes.

  • The two parts are close-together towards the outside, then further-apart towards the inside. If water lands on the labyrinth, surface tension will tend to hold it where parts are close together, which keeps it from getting inside. If water is pushed in, surface tension tends to pull the water back out to where things are close-together.

    Similarly, oil inside the hub can run out. The oil tends to get trapped at the outer part of the labyrinth (small gap), but surface tension pulls it out of the inner part (large gap). Outside junk can land and get trapped in the oil at the outer part, but it tends to get stuck there, since there is no easy way for the dirt to get worked in to the inner part.

  Labyrinth seals are often used along with cartridge bearings, but the seal is separate: cartridge bearings do not come with labyrinth seals. You can order a shielded or rubber-seal bearing cartridge, but not a labyrinth-seal cartridge.

  Labyrinth seals are used commonly in bicycle cup-and-cone bearings. They are sometimes are used as additional protection for cartridge bearings with shields or rubber seals. The outer labyrinth helps keep the inner shield or seal stay cleaner.

  Simple labyrinth seals are easy and cheap to build, but also easy to mis-design and get wrong. There are many examples of bicycle bearings where it appears the designer wanted to build a labyrinth seal, but did not understand how. In the worst case, mis-design leads to something which tends to help dirt and water get in to the bearing.

  Labyrinth seals can be very effective. However, labyrinth seals provide very little protectection against high-pressure water jets or immersion/submersion.

  • For lots of riding, it is rare that hubs are submerged, so labyrinth seals can work well.

    Older Sturmey-Archer hubs use them, and there are examples of hubs from the 1940s/etc. which have been used daily for decades in wet weather with little or no service other than occasionally adding oil, and the bearings and gears inside are in good condition.

  • If the bearing is immersed, a labyrinth can slow down the flow of water and can limit the maximum size of grit which enters the bearing. However, even very fine grit can lead to bearing damage. Fine grit can also interfere with free rolling of the bearing, even before there is bearing damage [http://pardo.net/bike/pic/fail-036/index.html as of 2025-03].

  • A labyrinth seal can protect a contact seal. When immersed, there may be air trapped inside which reduces the amount of grit and water which reach the contact seal. The labyrinth also acts like a sieve and limits the kinds of grit which can get through and reach the contact seal. The contact seal stays cleaner, so wears more slowly. Seal wear makes it more likely bearing immersion will get grit and water in the bearing, so by keeping the contact seal cleaner, a labyrinth seal can help improve the protection provided by a contact seal.

  • A labyrinth can also slightly delay water intrusion, as it simply takes a moment for water to “run the maze”. If the immersion is brief, this can reduce the amount of grit and water which reach the seal or bearing.

  • A labyrinth seal can also slightly reduce the pressure of water on a seal. If there a splash or brief immersion, there is friction for water to “run the maze”, so the first water to reach the contact seal is at lower pressure. Lower pressure makes it easier for the contact seal to resist intrusion. If external pressure is sustained, eventually the pessure inside will reach the outside pressure.

  For a centerless wheel, the bearing is much closer to the ground. Many bikes will occasionally go through water deep enough to submerge the lowest part of the wheel, and thus the lowest roller or the lowest part of a large bearing. Even though labyrinth seals are very effective for hub bearings, it is likely they are much less effective for bearings in centerless/hubless wheels.

  Once water and grit are inside, they can add drag and start to damage the bearing.

  Labyrinth seals add little weight or cost and no drag, so are probably a good idea for most bicycle bearings, including those in centerless/hubless wheels. Although labyrinth seals provide useful protection, they can only do so much.

  Thus, it is unlikely that even good labyrinth seals on a centerless/hubless wheel can give bearing life anywhere near as good as ordinary hub bearings.

Bearing seal drag for large bearings

Centerless/hubless wheels run the bearings close to the ground. A deep puddle can submerge the lowest part of the bearing in dirty water. If the bearing is shielded but not sealed, dirty water can get in the bearing and start damaging it. The bearing may then fail fast — e.g., in a few months. And, centerless/hubless wheel bearings cost more than hub bearings. So your cost goes up twice: first, you have to replace non-shielded bearings more often; second, it costs more when you do it — more parts cost, and more labor cost.

Switching to sealed bearings could keep the bearing clean inside, so it lasts a long time. But seal drag can be big enough to slow you down. I wish I could tell you how many Watts go to seal drag, but I do not have that information. Instead, here is an explanation why seal drag can be small in hub bearings yet large for centerless/hubless wheel bearing, even when using the same kind of seal. And, in turn, why you should be cautious about seal drag for bicycles (and maybe e-bikes) using sealed bearings for centerless/hubless wheels.

• Suppose seal drag in hub bearings costs you 0.01 Watt at 20 kph. For a rider putting out 100 Watts, that is 0.01/100 = 1/10,000 or 0.0001x of their power.

  Over a 10-hour ride, that might make you (10 hours) × (3600 seconds/hour) × 0.0001 = (3.6 seconds) slower. That is enough to lose a race. But for ordinary riding, it is too small to even notice.

• On a 6001 bearing, the seal is about 14 mm diameter. If the tire is 620 mm diameter, it has 620/14 ≈44:1 leverage over the seal drag.

  In other words, if 6001 seal drag is D, then the seal’s drag at the tire is D/44. D/44 is the drag that you (or gravity, or inertia) has to overcome to keep the seal turning.

• Suppose we switch to a bearing which is twice the size and has the seal running at twice the diameter. A 6005 bearing has a seal running at about 27 mm, which is close. Let‘s assume it is exactly twice the seal diameter. There are two effects:

  • The tire has half the leverage, compared to a 6001 seal.

  • The seal is twice as large. That means there is twice as much seal area rubbing on the bearing’s inner race. Every place the seal touches, it needs to push just as hard as the 6001 seal to get a good seal. But there is twice as much seal rubbing area, so the total seal drag is twice.

  So (a) seal drag doubled, and (b) leverage is halved.

  The 6001 drag was D, so now we have (D × 2) / (44/2) = (2×2) × D/44 = D/11.

  In other words, doubling the seal diameter led to four times the drag at the tire.

  In the same way, a bearing which is 3x as big has 3×3=9x seal drag. And a bearing 4x as big has 4×4=16x seal drag. And so on.

• If the hubless/centerless wheel bearing seal is 540 mm diameter, then it is 540/14 = 38.6 times the diameter of the 6001 bearing seal.

  As above, that means 38.6×38.6=1490x the seal drag.

  If the 6001 bearing was 0.01 Watts at 20 kph, then the hubless/centerless wheel bearing seal is 1490 × 0.01 Watts = 14.9 Watts. If you ride at 100 Watts, that is about 15% of your total power.

  That is enough to turn a 10-hour ride in to an 11-hour ride. Definitely enough slower to notice!

• I do not know what is the actual seal drag or power for a 6001 hub bearing. So the above “Watts” and riding time estimates are (probably) wrong. So why give them? My goal is to show seal drag might be unacceptly high, so folks who care will be sure to get good numbers — and avoid unhappy surprises.

  Also, I suspect seal power is higher than 0.01 Watt at 20 kph. In which case the scenario is even worse — more than 15% of rider power.

• Note also this is only counting seal drag. A large bearing has the same “leverage” problem as the seals. If hub bearings take 1 Watt at 20 kph, then a 38.6x larger bearing may take 38.6 Watts, plus the 14.9 Watts for seal drag.

To summarize, non-sealed bearings are likely to fail fast in ordinary riding, since the bearing is close to the road, making it easy for junk to get in the bearing. Seals can probably fix that, but seals are likely to slow you down a lot.

Seal drag is less of an issue on an e-bike, but might still be enough that you notice — shorter range, slower acceleration, and lower top speed.

Maybe somebody will get seal drag small enough. “If it works, it works”. But beware of vague claims, such as “we reduced seal drag to accpetable levels”. It might be acceptable to them, but that does not mean it is acceptable to you or anybody else!

Look for something numeric, like “Watts at 20 kph”. Riders often put out around 100 Watts, so if seal drag is 1 Watt or less, you probably won’t mind, and maybe won’t notice. If it is 10 Watts or more, you probably care a lot.

Bearing seal drag for “many rollers”

The discussion above assumes a wheel bearing almost as big as the rim and tire. [Reevo] rims instead have 17 rollers, which run in a track on the support hoops. Does using many little rollers avoid the “big seal” problem described above?

It reduces the problem, but does not entirely avoid it. Here is why.

• Simplified explanation:

  The Reevo has 17 rollers in the rim. So, basically, you are replacing two hub bearing seals with 17×2 = 34x seals. So seal drag for “many little rollers” is 17×38.6 = 656x the hub bearing seal drag. That is better than 1590x ... but not much.

• More complete explanation:

  Each roller has a bearing in it, so there is some leverage from the outside of the roller (which runs on the track) to the bearing seal. For example, a 6001 bearing has about a 14 mm diameter seal and 28 mm outside diameter. So if you run a 6001 bearing directly on the track, it has about 2:1 leverage. If you put the 6001 in a 42 mm roller, that gives about 3:1 leverage. In turn, 17×(38.6/3) = 218x the drag.

218x is definitely better than 1590x, but still a lot worse than 1x.

In practice, we care about Watts of drag, and not whether it is 10x or 100x or 1000x. Either the drag power is big or small. How big is it at 218x?

• I do not know, but consider this estimate: if hub seal drag is 0.01 Watts, then 17 rollers with 3:1 leverage is 218x more seal drag, or about 2 Watts.

• Will you notice 2 watts, or will you care? It depends on the rider and riding. But this estimate seal drag here is large enough to be concerning, so this is an area where you want actual numbers.

Note that switching from a big bearing to many rollers also means adding a track where the rollers roll. There is roller/track drag, in addition to bearing drag and bearing seal drag. Although roller/track drag is not seal drag, if you switch to rollers to reduce seal drag, then roller/track drag is in-effect part of the seal drag. See below for more.

To summarize, “many rollers” has the same basic “lots of seal drag” problem that happens with a big bearing. “Many rollers” may make things a little better, but it may still be too big to be acceptable. E-bikes may tolerate more drag. But the numbers above are guesses, and I may be guessing too low.

A centerless/hubless wheel maker should have actual numbers. If they do not, you should be concerned: if they lack wheel drag information, what other critical information are they also missing? Or, if they know and will not say, they may be hiding a problem.

Bearing seal drag for “a few rollers”

The [Black Hole] uses just four rollers. Only three are close to the ground — so maybe the top one could run without a seal. That gets rid of the 17x problem, above. Does that solve the seal drag problem?

• One problem is centerless/hubless wheel bearings can be more expensive to replace than hub bearings.

  How much more? It varies a lot.

  • Standard bearings the size of a rim can cost well over US$1,000. A custom bearing (shown in some videos) may be just as expensive.

  • [Reevo] uses rollers with off-the-shelf cartridge bearings. Each bearing might be about the cost of a hub bearing, so with 17 rollers, only about 17/2 = 8.5x as much to buy the bearings. However, removing and disassembling the Reevo’s wheels is far more time-consuming than a bearing swap for many hubs. And since it is complicated, many riders will lack the tools, skills, and/or time to do the work, so cannot “do it yourself” to save labor costs.

  • The [Black Hole] uses a roller which looks to use a standard cartridge bearing, and might be replaced more easily than many hub bearings, meaning lower labor costs. There are four bearings, but labor may be less, so the total cost might still be less than the cost of replacing some hub bearings.

  Bearing purchase and labor costs vary a lot with the design. For some, costs are sure to be much higher than for hub bearings. For others, the cost may be less than the cost of hub bearings.

• A second problem is bearing life may be dramatically shorter than for hub bearings. Even if the basic service cost is okay, multiplying by more-frequent service can make it grow.

  People running inline skates with shielded bearings sometimes find that after skating in the rain, the bearings are rough after a few days. Bearings in centerless wheels may be better protected, but are still easy to submerge. It seems likely that people who ride regularly in the rain may need to replace unshielded bearings several times per year, maybe more. Versus hub beaerings may last a decade or more of all-season use without service.

  Some wheel designs combined with some riding styles may give good durability. But buyers should be beware that some riders may get durable service, while other riders need frequent replacements.

  Also, a bike may not be ridable while it is waiting for service. And, time spent getting it to and from service is time not doing something else. How much this “costs” depends a lot on the rider. But it is a disadvantage of service-intensive approaches.

• A third problem is that bearing drag is already much higher, so you might not be able to delay service.

  Many people ride for years on hubs with slightly damaged bearings. Since hub bearing drag is small, the higher drag of a damaged bearing may be unimportant. For example, if hub power increases from 1 W at 20 kph to 2 W at 20 kph, the rider might not care, and might not even notice.

  • Some bearing damage can also allow brakes to rub, causing drag, which increases the power. So total power from worn bearings may be bigger than just the bearing power.
  • A damaged bearing can fracture and lead to serious hub damage, and also make the bike unridable, stranding the rider. Even if a damaged bearing has low drag power, there may be reasons to replace it sooner rather than later.

  With a centerless/hubless wheel where bearing drag is already (say) 5 Watts at 20 kph. Doubling that due to bearing damage could mean 10 Watts at 20 kph. For a rider putting out 100 Watts, that is a big increase in drag.

  In other words, you might be willing to ride hub-type wheels with minor bearing damage, but might be much less willing to ride centerless/hubless wheels with a similar level of damage.

Roller and track drag

A centerless/hubless wheel can run on a large-diameter version of a standard bearing; or can run on rollers. Things affecting drag are similar, but not the same. Here is some further discussion of drag for rollers-in-rim and rim-on-roller wheels [Bearing layouts].

• Bearing drag power:

  As noted above, a large bearing may have a lot more drag than a hub bearing. Do rollers reduce that? Short answer: no. Longer answer: yes, but only by the ratio of roller diameter to bearing diameter.

  If a 6001 bearing normally has 44:1 tire-to-bearing leverage, then putting the same bearing in a 3:1 roller means 44/3=14.7x the drag. A tire-sized big bearing has 44x the drag. I used 1:1 for simplicity. A real bearing is smaller than the tire, but at 90% the diameter still has about 40x the drag.

  If hub bearings take 1 Watt at 20 kph, then a 3:1 roller takes about 15 Watts at 20 kph. A 1:1 big bearing takes about 44 Watts.

  You can have any number of rollers and bearing drag is about the same. Why? Because bearing drag depends on the load. If you spread some load over N bearings, each bearing will have about 1/N drag. So for one bearing or for many, the bearing drag is about the same. (“1/N” is an approximation, but is close enough for the current discussion.).

  Note that seal drag happens when a bearing turns, whether or not it is loaded — so N bearings will have N times the seal drag.

  Suppose a rider puts out 100 Watts, and suppose hub bearings usually cost 1 Watt. Going to 14.7 Watts of bearing drag could add about an hour to 10-hour ride.

• Roller-on-track drag:

  A roller can somewhat reduce bearing drag, but the roller needs to roll and so adds roller-on-track rolling drag. A big bearing has bearing drag, but no roller-on-track drag. So a fair comparison needs total drag for rollers that includes both bearing drag and roller-on-track drag.

  Rolling a roller on a track has some drag where the roller meets the track, in addition to the roller’s bearing and seal drag. Also, roller-on-track drag goes up when the roller and track are dirty. Even fine dust can somewhat increase rolling drag. Low drag in the workshop might still mean high drag when you are outside riding.

  How much is roller-on-track drag? I do not know. So here is an example using a made-up cost: suppose 3 Watts for roller-on-track drag. Then you have 14.7 Watts of bearing drag plus 3 Watts roller-on-track drag for 17.7 Watts total (plus seal drag if the bearings have seals).

  That is still much better than 44 Watts for the big bearing, but much worse than 1 Watt for the hub bearings.

  I do not know if 3 Watts is reasonable, but it is easy enough to measure. And is something you want to make sure is included: “We reduced seal drag” is not so useful if they did it by adding lots of other drag!

  Note that roller-on-track drag depends a lot on details: metal-on-metal can be efficient when clean, but dirt can hurt the efficiency a lot. The Reevo appears to use rubber-on-metal, which may be a lot less efficient when it is clean; but might be more efficient when the track is dirty. If a maker gives you only one number, it is probably for clean running — and may be small compared to real-world drag for dirty running.

To summarize: using rollers instead of big bearings may reduce drag in some situations, but can add roller-on-track rolling drag. When the track is dirty, efficiency may suffer. I do not know if rollers are more or less efficient than a big bearing, but either one looks to be a lot more drag power than a conventional hub bearing.

“Big Bearing” Alternatives

A centerless/hubless wheel can rotate on rollers, or can rotate on a wheel-size bearing. Here, I refer to a wheel-size bearing as a “big bearing” The [Top Secret] e-bike uses “big bearing” wheels.

Unfortunately, big bearings can be expensive. A hub-size cartridge bearing, such as the 6001, can be bought retail in small volume for around US$1 each for a cheap one, maybe US$10 each for a good-quality one. In contrast, an industrial cartridge bearing the size of a bicycle rim can cost over US$1,000. Top Secret makes their own bearing, but it seems likely it will still be muchssssssss more expensive than hub-size bearings.

Another issue is the lowest part of the bearing runs close to the ground. For example, rolling the wheel through a deep puddle can submerge part of the bearing. If the bearing is not sealed, water and grit can get in, leading to rapid bearing wear. A seal can help keep out water and grit. But, unfortunately, a good seal on a large-diameter bearing can have much more drag than a similar seal on a hub-size bearing [Bearing seals].

High price together with short service life can make it expensive to own and ride a big-bearing bike. Is there a way to bring down the cost? Maybe, here are some possible alternatives.

• Rollers

  The [Black Hole] and [Reevo] and other bikes use rollers with hub-size cartridge bearings. The Black hole has 4 bearings versus 2 in a standard bicycle hub, so the cost only slightly more. The Reevo is similar with 17 bearings per wheel.

  For both of these, seal drag is very likely high enough to be a problem (see above), but non-sealed bearings are likely to wear out quickly in ordinary riding, such as riding in the rain. Bearings might wear out in months, whereas hub bearings often last years of all-weather riding without service. Although the bearings are cheap, frequent replacement can multiply the cost. For the Reevo, bearing replacement also looks to have a high labor cost.

• Cheaper "big bearing"

  Cartridge bearings used in hubs are more precise than needed. But at a dollar each and fairly durable, there is not much to be saved by switching to custom low-precision bearings.

  Traditional cup-and-cone bearings are slightly less precise, but also adequate, and adequate even if somewhat mis-adjusted. You can even run a wheel with pitted bearings and races; until it disintigrates, it can work well enough for some uses.

  A centerless/hubless wheel may be able to use a bearing which is less precise than a hub bearing. Some hub bearing “slop” is amplified by the length of the spokes — a small motion at the hub can lead to a bigger wobble at the rim. In contrast, a big bearing is already at the rim, so there is no amplification.

  What are some options for a cheaper big bearing?

  Restaurant turntables often run on bearings made of cheap stamped steel. They are close to the sizes needed for a centerless/hubless wheel, yet cost only tens of dollars, perhaps US$100 in wheel sizes. They give long service holding loads similar to the weight of a bike and rider, at least when used inside a restaurant and hidden under a turntable from splashes and spills.

  Turntable bearings are axial bearings, whereas most bicycle loads are radial. However, the same stamped-bearing ideas could probably be used to make a wheel-size radial bearing at a size and load capacity needed for bicycle wheels. But at a cost perhaps well under US$100.

  However, a disadvantage of cheap stamped steel compared to bearing steel is that cheap steel is more easily damaged by grit than a high-quality steel. Turntable bearings are used indoors and hidden inside the turntable where gravity keeps away almost everything except airborne dust. In contrast, a bicycle wheel bearing could easily be submerged in mud, dirty water, and lots of other abrasive grit.

  A good-enough seal can improve durability. But it is hard to make a seal which both does a good job and has low drag. Even harder to do it at low cost.

  Thus, a stamped-steel big bearing might be much cheaper. But it may have a shorter service life than a better bearing, so more-frequent replacement plus more labor costs could eat away at the savings. As with big bearings, seals can help, but for some uses, seals are likely to have too much drag.

  To sum up: cheaper big bearings may be workable, but it does not look promising.

• Plain bearing material innovations

  Sliding-contact “plain” bearings are by far the most common type of bearing. But they are not used in bicycle wheels, probably in part because:

  • Plain bearings need to run slightly loose, in order to run free. Howevever, slightly loose can lead to brake rub — for both rim and disk brakes. Plain bearings can be made with higher accuracy, to run tighter, but that costs more.

    In comparison, ball bearings can be tightened until there is no "slop" and even can run a little bit tight ("preload").

  • Drag of plain bearings is often more than ball bearings. For many riders and riding, hub drag is small enough to ignore. But make it (much) bigger, and people will notice.

  Many pedals use a plain bearing near the crank, and drag is tolerable. A hub using plain bearings would have bearing drag similar to a pedal bearing with the rider’s weight on the pedals (not the seat). Details:

  • Pedals use a plain bearing near the crank, but usually use a ball bearing away from the crank. That may cut friction almost in half compared to using all plain bearings.

  • Many riders spend most time pedaling seated. Load on the seat does not go through the pedal bearings. Thus, typical drag (sitting + standing) is much less than standing drag. In contrast, a wheel bearing carries the rider’s full weight at all times.

  • Drag power goes up as the speed goes up. Hubs often turn several times faster than the crank — a 48T chainring sprocket driving a 24T sprocket is 2:1; driving a 16T sprocket is 3:1. Even if drag friction is independent of speed, drag power is basically (drag friction) × (speed), so power goes up with speed. So even under equal load, the slower-turning pedal bearing saps less of your energy than a faster-turning hub plain bearing.

  • Pedals often use bushings as a way to reduce pedal thickness. Bushings are thus a compromise to save space. Hubs typically have room for ball bearings, so do not need to compromise this way.

  Various materials are used to make low-friction plain bearings. Many materials are somewhat soft, and many need to be lubricated by a small amount of oil (which may be included in the material). Oil will tend to hold grit. And oiled or not, grit that gets in will tend to chew up the (softer) bearing material.

  That means most plain bearing materials are not suited to a centerless/hubless wheel: either they have high drag, or they need good seals, and the good seals have high drag.

  However, some experimental materials such as boron carbine may allow low-friction plain bearings which can run without lubrication, and which are hard enough they grind up debris before the debris damages the bearing. That might allow a bearing to run without seals, so that it has low drag; and yet is still durable.

  Unfortunately, many such materials are (a) still experimental, and (b) likely to be very expensive for the forseeable future.

• Magnetic levitation

  With no rubbing/sliding contacat, magnetic levitations bearings can have lower drag than ball bearings. Since there is no contact between parts, they do not wear, even when dirty.

  Some magnetic bearings are “active” and require external power. The power has to come from somewhere, so it is effectively drag.

  Other magnetic bearings are “passive” and can run without external power.

  However, it is hard to make a magnetic bearing assembly where all loads are carried by passive bearings. Thus, systems use either

  • A mix of magnetic and non-magnetic bearings. Major loads are carried by passive bearings. At least one load direction is carried by some other kind of bearing.

    Using multiple bearings adds cost, and other bearing types add drag.

  • A mix of passive and active magentic bearings. All loads are carried by magnets, but at least one load direction is carried by by an active bearing. Active bearing power can come from an external source (e.g., battery) or can be generated by the rotating parts (a built-in generator).

    Multiple bearings and control electronics add cost. Active power is in effect bearing drag, even though some of the power is dissipated outside of the bearing. In other words, if you need 1 Watt to run the control electronics, that is power that had to come from somewhere, so in effect is bearing drag, even though it was the electronics that got warmed by the power, and not the rotating bearing parts.

  Mixed magnetic/non-magnetic bearings in centerless/hubless wheels are likely to suffer grit/wear problems described above, for the non-magnetic part of the bearing assembly. Total bearing friction may start out less due to the loads carried by magnetic bearings, and because they can run without seals. However, once the non-magnetic bearings wear, they need service. It seems likely the service intervals and cost will be similar to non-magnetic bearings described above.

  Active bearing components take power, but the total power can be similar to ball bearings. Even with power for control electronics, the bearing drag might be on the same order as hub-type ball bearings. I should emphasize “might” — I am not aware of guides/etc. for energy cost vs. bearing design vs. load. There probably are such guides, and I just do not know about them.

  Even if the rolling drag (including active power input) is higher, aerodynamic savings from reduced windage may mean total drag is lower. This is unlikely to be significant for ordinary cyclists, but might be useful for racing.

  A wheel-size magnetic bearing, power electronics, etc. is likely to be heavier than spokes and a hub. If it is only a little bit heavier, that may be fine for general use. If it is a lot heavier, there may still be specific races where it is acceptable — some hour record bicycles have used weighted wheels — but especially heavy wheels might not be acceptable for ordinary riding.

  A wheel-size magnetic bearing is likely to be much more expensive than ordinary hub-size ball bearings. It might be cheap enough to use in expensive bicycles. Notably, bicycles are relatively tolerant of bearing “slop” compared to turbines and many other uses. So lower-cost designs and manufacturing might be acceptable for bicycles.

  I am not aware of any issues which make magnetic bearings inherently a bad match to centerless/hubless wheels. However, it seems likely that they would be limited to “botique” uses due to cost and weight compared to a spoked wheel.

  • [https://www.magneticbearings.org/passive-bearings as of 2025-03]
  • [https://mechanical-engineering.com/magnetic-bearing as of 2025-03]
  • [https://ntrs.nasa.gov/api/citations/20020038851/downloads/20020038851.pdf as of 2025-03]
  • [https://www.magneticbearings.org/app/uploads/publications/ismb6/ismb6_submission_12.pdf as of 2025-03]

Derailleur jockey pulleys

As an aside, note that derailleurs sometimes use large-diameter jockey pullies and other "tricks" to reduce chain drag caused by pulley bearings. It may seem odd to spend effort on jockey pulleys: they carry only a light load from the derailleur’s chain tension spring. Which is tiny compared to the load on hubs.

However, simple plain-bearing pullies can use over 1.2 Watts for a pair of pullies. Low-drag pullies can disspiate under 0.1 Watt — over 1 Watt less [https://ceramicspeed.com/pages/11-tooth-derailleur-pulley-wheel-efficiency-test as of 2025-03]. For a rider putting out 100 Watts, that is 1% of their total power.

Pulley power consumption depends both on the bearing type and also how fast the pullies turn. If you are on an 22-tooth rear sprocket, an 11-tooth derailleur pulley spins at twice the speed of the sprocket and wheel. If you could double the pulley diameter, pulley power would drop in half.

For racers, 1% can be the difference between winning and losing. For an ordinary rider, a plain-bearing pulley does not slow you down much. And once you have a “good” bearing, fine-tuning probably does not matter.

But it seems thought-provoking: pulleys carry only a light spring tension load, yet bearing drag matters “enough” for some riding that it gets attention. I mention it here as another way of thinking about why bearing drag for a centerless/hubless wheel is important — the diference in drag power between a hub-type wheel and centerless/hubless wheel is much more than the 1 Watt difference between kinds of derailleur pulleys.

Roller layouts

There are various designs for how rollers support the rim. How the rollers and rim engage can have a big effect on drag and maybe on how fast the rollers and track wear.

Some options include:

• Flanged roller on track, or roller on flanged track. Some loads bear direclty on the roller face. Perpendicular loads bear on the flange. The [Reevo] uses rollers on a flanged track.

• Rollers for radial loads; separate rollers for lateral loads. [The Q] uses this approach.

• V-shaped rollers on a V-shaped track. The [Black Hole] does this.

• Pairs of angled rollers on a V-shaped track. This is suggested by the [Black Hole] patent.

• Consider the [Terpstra] and [Yale demo], which use a groove on each side of the rim. There are rollers on each side of the rim which run in the grooves.

  When you put weight on the bike, rollers near the top of the wheel support most of the weight. The rim’s groove pushes on the side of the roller. As it rolls, there is sliding between rim and roller, because contact points on the side of the roller are at differing radii: when the roller turns, points near the edge of the roller travel twice as far as points on the roller at half the radius. Since the roller and groove are both rigid, the differing distances lead to sliding. Sliding means drag.

• Consider instead the [Sada] front wheel. It uses flanged rims which are wider than the tire. There are rollers above and below the flange. The rollers contact the flanges using the face of round rollers, so the whole contact is the same radius. Therefore, it can turn without sliding.

Side loads/etc. make things more complicated, but one observation is the rollers support the rider’s weight all the time, whereas side loads tend to be both smaller and intermittent.

Some details:

• Sliding adds drag. Sliding also adds wear, especially when the rim and rollers are dirty: dragging grit across the rim and roller under load is like running sandpaper on them while you ride.

• A rolling contact can also be damaged by grit. First, even a “rolling” contact has some sliding. Second, grit can be pushed in to the rim or roller, causing divots and bumps which are slightly rough. Over time, irregularities can flake off.

• Rims can tolerate some sliding wear: rims wear slowly enough that rim brakes are practical, even though the pads slide on the rim when applied, and even though the rim and pads are often dirty. But brakes are used only intermittently. Consider riding with a brake which is always dragging, and especially in a gritty environment. It will wear faster.

  Also, in some riding, rim brakes wear out rims quickly. Here is a rim that was used with rim brakes and wore out in six months of road riding in the Pacific Northwest: [http://pardo.net/bike/pic/fail-026/RIM-003.html as of 2025-03]. Part of the problem here is wear rates are unpredictable, and probably also for roller-supported rims. Even if rims usually last several years, there may be conditions where a rim wears out in just a few months. So you may need to check e.g., monthly, just in case it is wearing out fast. Hub bearings do not need monthly wear inspection, so regular checking of centerless/hubless wheels is an added time cost.

• Roller wear is also an issue. Rollers should be cheap to replace, but both parts and labor are added costs compared to hub-type wheels.

• Roller-supported rims may be prone to some jamming, which leads to wheel, bike, and/or rider damage. Even if jamming is rare, it could be a big extra cost when it happens. See [Safety of roller-supported rims] for details.

Roller layout can have a big effect on drag. And where there is drag, there is wear. The parts, labor, and checking are also added costs compared to a hub-type wheel.

Safety of roller-supported rims

For roller-supported rims, objects can get caught between roller and rim. Unfortunately, once things are trapped between roller and rim, the rolling motion of the bike tends to pull them in more. Some objects may pass though, but others will tend to jam the roller and rim and may damage the bike and/or lock the wheel and throw the rider.

• Small objects — grit, leaves, etc. — can roll through. Larger objects might pass through if they if they can push apart the rollers and rim. If the object is not able to push things apart, the object may be crushed. These things cause momentary extra drag, but as long as extra drag is rare, it should be fine.

• Objects might jam without passing through. High jamming forces may spread apart the rim and roller and damage the rim, roller, and/or the framework supporting the roller.

• Or, jamming forces may lock or eject the wheel. This can damage many parts. It can also throw the rider from the bike, causing injuries.

Of course regular bikes also fail. And, a spokeless wheel never has failures or accidents having to do with spokes — so it may be reasonable to suffer more of some kinds of failures, in order to get less of others.

One question, then, is: how likely are these sorts of failures for centerless/hubless wheels? I do not know, but below are some observations.

• Ordinary bikes already have limited clearance between the rim and brake pads. On one hand, it is rare for things to seriously foul the brake pads. On the other hand, there is often 1-2 mm between pad and rim; brakes often let the pads move side-to-side; and spoked wheels are somewhat flexible side-to-side. Perhaps most important, the pads meet the rim at about a right angle, which tends scrape off things thicker than 1-2 mm, so it is hard for large debris to get in the gap.

  Rollers supporting a rim are different. They need to run in contact, not 1-2 mm away. Rollers need to be stiff, so the rim does not wander excessively, pry itself out, or jam. A roller against the rim is a “funnel”, which tends to trap things, rather than scrape them off. Where the roller and rim come together, it resembles feed mechanisms in presses, shredders, and other equipment where the goal is to force things between the rollers/cutters/etc. So once junk gets to the rim/roller area, it is likely to be drawn in forcibly, increasing the risk of jamming the wheel.

• The Reevo’s rollers are hidden behind shields. That makes it hard for large debris to ever reach the roller/track. In turn, that makes it more likely whatever does get there is small enough to pass between rim and rollers.

  In contrasat, the Terpstra and Yale bike have the rollers relatively out in the open. Things could get trapped because they stick to the rim, but also things that simply get too close, rub, and get pulled in — branches, clothes, body parts, and so on.

  The Reevo suggests a hybrid approach, namely a scraper and shield to protect each roller — something to scrape off large debris before it reaches the roller, and something to protect against things falling in to the rim/roller area. In essence, a shield for rollers, instead of the Reevo’s shield for the whole rim.

It seems likely that with suitable shields, safety can be good. However, even with a small gap, there is some chance things can build up inside. For example, riding through grass, each blade may be able to slip through a gap in the shield, then a bunch of blades may get tangled together and jam the wheel. Similarly, mud might slip through the gap but build up and jam the rim/rollers instead of passing through.

Brake types

Some brakes used on centerless/hubless wheels include:

• None
• Tire tread “spoon” brake
• Rim brake
• Disk brake
• Drivetrain-mounted brake
• “Fixie” -- a rider-resists-the-pedals drivetrain brake
• Electric field — a wheel motor used as a brake.

For many bikes, rim and disk brakes are the most common choices. Here is a quick comparison; more information about them and other brake types is below.

• Rim brakes are lighter, and are harder to knock out of true when handling or parking in a shared rack; but they are less predictable in wet/mud, and they wear out the rim so can be expensive due to the labor cost of replacing the rim.

• Disk brakes are more predictable and the wear parts can be replaced with little labor; but are heavier, and disks are easy to knock out of true in handling/parking, leading to annoying brake rub.

• Both rim and disk brakes can fail in serious ways that injure the rider. Unfortunately, failures depend a lot on details of the brake system, making it hard to make general statements — one brake might dissipate high power without any problems, while a similar-looking brake might fail suddenly and completely at half that power. Whether a brake is suitable sometimes depends more on how the brake was designed, and less on whether it is a rim or disk brake.

• A small-diameter rim means a rim brake has less cooling area and so can be easier to overheat. A large rotor can be fitted to a small wheel, so it can still dissipate high power.

In more detail

• None

  Cheap, light, and easy. But bikes need brakes on both wheels to stop reliably in all uses. “No brakes” is fine for an art bike. But for ordinary riding, if you see a wheel without a brake, one of your first questions should be “How would you add a brake?”

• Tire tread “spoon” brake

  A metal plate presses on the tire tread. The plate is somewhat spoon-shaped, hence the name “spoon” brake. Spoon brakes have been common on bicycles since the 1800s, and are still sometimes used, mainly on cheap bikes with solid (airless) tires. They tend to wear the tire quickly, and are affected by water/mud/dirt/etc. that sticks to the tire tread.

• Rim brake

  Rim brakes are used widely and can be a good choice. Rim brakes have various problems. Including:

  • Rim brakes can be unpredictable in wet weather. Riders can learn to work around the problem — ride in a way that allows much longer braking distiances in wet/muddy conditions. But a brake that simply works reliably is an advantage.

  • The brake wears out the rim. Worse, wear is unpredictable — in clean conditions, a rim may last a decade. In especially dirty conditions, the same rim type can wear out in a week. A replacement rim is often expensive to buy and may also be expensive labor to install.

  • When the rim gets hot, the tire rubber gets slick and the tire can move on the rim. Two not-rare hot-brake failures are:

    • The tire rotates relative to the rim, and tears the valve stem out of the inner tube. The tire goes flat suddenly, and a flat tire has much less lateral traction.
    • The tire shifts eccentrically on the rim and then blows off. The tire goes flat and comes off the rim suddenly, leading to a complete loss of lateral traction.

    Both of these failures are likely to cause a crash, as well as a complete loss of braking. Whereas an overheated hub brake may cause a sudden loss of braking, but does not hurt tire traction. Although crashes are common after hub brake failures, there is at least more chance to choose what to crash in to.

  Rim brake cooling can be improved by using a deep-section “aero” rim with more surface area, at the cost of somewhat more weight. A deep-section rim can be much stiffer than a shallower rim and can significantly improve both wheel strength (one-time loads) and durability (long-term use). In other words, added weight for a safer brake can also have other benefits. Whereas a larger disk brake rotor improves braking but not other things.

  Although rim brakes are used widely, hub brakes — disks, drums, band brakes, coaster brakes — are often more predictable, longer-wearing, and cheaper to maintain. For many riders and riding, these factors make hub brakes a better choice.

  Aside: rim brakes can be made more predictable and longer-wearing. For example, Wilson’s “Positech” [http://pardo.net/bike/pic/fail-015/000.html as of 2025-03-15]. But currently (2025) no brakes like this are available commercially.

• Disk brake

  A disk brake rotor is attached to the rim

  Disk brakes are often:

  • More predictable than rim brakes. This is partly because of how the brake works, and partly because the disk rotor is usually up high above the road, so is less often covered with mud/water/etc.

  • Cheaper and easier to replace worn parts. Disk rotors often last longer in dirty service than a rim; both because of material choices, and because the rotor is less often covered with stuff. Disk brake rotors are often cheaper than rims; and replacing the rotor is mainly a matter of a few bolts and a few minutes work. Replacing a rim is much more complicated, increasing labor cost. Disk brake pads are often more expensive than rim brake pads, but often last longer in dirty service.

  Disk brakes fail when overheated. Failures are varied and unpredictable.

  • Failures can be annoying, such as bad smells, somewhat reduced brake strength, and a rotor that rubs and makes noises. Failures can also be dangerous, such as sudden total loss of braking. There are also other kinds of failures. The kind of failure varies a lot depending on brake design.

  • Failures happen when a lot of power is going in to the brake. The maximum safe brake power varies probably 3:1 or more between brakes. And some brakes can fail catistrophically at low power, while other brakes may tolerate much higher power and also have much safer failure — e.g., gradual loss of brake power versus sudden and total loss.

  • Brake failure is easy enough to measure, but is not widely reported (as of 2025-03). That makes it harder for buyers to choose brakes.

  For a centerless/hubless wheel, disk brake problems include:

  • There are not standard disk brakes which fit to a rim. If centerless/hubless wheels were in volume production, there would be standard brakes to fit. But now and for the forseeable future, no standard brakes will fit.

  • A large disk rotor is ... larger, which adds cost and weight. A rim-size rotor operates at lower force, so it may be possible to use a different design or materials to offset the weight. However, that is additional development effort, and not sure to work. For example, the disk runs close to the ground, and the extra dirt may cause light materials to wear too fast.

  • Conventional disk brakes are more sensitive to lateral wobble or “play” than rim brakes. Some hubless/centerless wheels have a lot of play. There are low-play wheel choices, but they can introduce their own problems (friction, cost, etc.). Thus, a conventional disk brake may work poorly for some wheels, even if rim brakes work well.

  • Putting the disk near the rim means it is also closer to the road, so more likely to get splashed with mud, dirty water, etc. That may hurt its predictability. That may also cause it to wear faster, increasing the effective cost.

  Note that a rim brake is a disk brake. We call them different things because it is useful and convenient. But if you see a rim-size disk, it may be worth asking how that makes it more like a rim brake than a disk brake.

• Drivetrain-mounted brake

  A disk, drum, or other hub-type brake is connected to the wheel via a chain, belt, or other “drivetrain” mechanism.

  This is somewhat a natural fit for the rear wheel, which already driven. However, for a non-driven wheel, the “brake” drag, cost, weight, and durability are all hurt by needing to add a drivetrain just for the brake.

  A drivetrain brake has the disadvantage that a drivetrain failure, such as a dropped chain, also means a brake failure.

• “Fixie”

  A drivertrain brake in which the drivetrain cannot frewheel, so the pedals turn any time the bike is moving. The rider can resist the motion of the pedals, which has the effect of braking the wheel.

  A fixie brake can be light and cheap. However, it tends to have worse braking and requires more rider skill to use — fixies are known for causing untrained riders to crash. Also, the bike cannot coast, so it can limit how the bike is used. Also, there are few off-the-shelf choices for multi-speed fixie gearing — that either limits use, or requires drivetrain development.

  For most riders and most riding, being able to coast is important, and an explicit brake is probably a better choice.

• “Electric field”

  An electric motor can be used as a brake, in two ways:

  • Regenerative braking, also called “regen”. The wheel turning the motor generates electricity, which is fed back to the battery and can (partly) recharge it.

    As the bike slows down, power generation falls off, so braking is reduced. As the speed approaches zero, brake force approaches zero. Thus, regen is useful to recover lost energy, but it is not a full-function brake.

  • Driven braking

    The motor is used as a motor, but motor power is used to slow the bike, rather than speed it up or keep it moving.

    Driven braking takes power, so it runs the battery down. Also, if the battery goes flat, a driven brake stops working.

    I am not aware of any e-bikes using driven braking. Instead, they use a separate brake. That suggests for a bicycle, the costs are bigger than the benefits.

  Some limitations and complications of electric field brakes:

  • Both regen and driven braking need the wheel to drive the motor. But many electric drives use a freewheel so cannot use either regen or driven braking.

  • Both regen and driven braking are limited by motor strength. Basically, a motor cannot slow a bike any more briskly than it can speed up the bike.

  • It may also be hard to make an electric brake as reliable as a mechanical or hydraulic brake.

  • Adding braking also complicates the motor controler. First, the controller needs new braking features. Second, the controller may need to handle higher power than for a motor only used for acceleration. These may make the motor controller a lot more expensive, even in volume production.

  Regen and driven braking may be good as a brake, but might not be suited as the brake for a wheel. It might be okay to mostly use the motor as a brake; then for heavy braking or a flat bettery, use a “backup” rim brake. That avoids most rim wear, and also avoids problems of an electric-only brake. But using both adds weight and complexity for the backup brake.

All of these brakes work with hub-type wheels. Some conventional brakes, such as drum brakes, work well with hub-type wheels but are hard to use with centerless/hubless wheels, except as a drivetrian brake.

For most uses, there seems to be no advantage to a rim-mounted brake. A centerless/hubless wheel uses a rim-mounted disk because it has to, not because it is better. Intuitively, if you can put a brake on the rim of a centerless/hubless wheel, you can put the same brake on the rim of a hub-type wheel. If a rim-mounted brake had advantages, we would probably do that already, but we do not.

The main case where we do use a rim-mounted brake is a design which integrates the rim and the brake — since one thing works as both a rim and a brake, it can save weight. This integrated “rim plus brake” is often called a “rim brake” :-)

Disk brake attached to a rim

Here is an example of a disk brake attached to the rim of a spoked hub-type wheel.

        
From: Sardinia Hot Bikes Facebook 2025-03, [https://classic-cycle.com/community-bilder/community-bild-von-rudecycles as of 2025-03]

[https://www.gadgetreview.com/21-game-changing-bikes-paving-the-way-for-a-greener-future as of 2025-03]

The Rude Cycles “LA120” is a custom chopper-style bike that features pneumatic elements for adjustable ground clearance. This bike is designed for performance and style, making it a standout in the cycling community. The unique design allows for flexible riding experiences, whether on the road or performing tricks.

This bike is perfect for those who want to make a statement while riding. The Rude Cycles “LA120” combines aesthetics with functionality, making it a popular choice among custom bike enthusiasts.

Rim brakes will rub if the rim “wanders” side-to-side. Disk brakes are usually even more sensitive to wander. The brake disk is attached to the rim and depends on the rim/wheel to position the disk. So if the rim wanders, pad/rotor rub can be an issue.

This is an art bike rather than daily transportation, so more rub may be acceptable than for a daily transit bike.

The rim is also very wide, so should be laterally stiff, and thus relatively unaffected by minor spoke tension changes or by side-to-side forces when riding. A rim which is much less stiff may have more wander and thus brake rub.

A centerless/hubless wheel may have a large-diameter wheel bearing, which can reduce rim wander compared to hub-type wheel. That could mean a centerless/hubless wheel does not have brake rub, but a hub-type wheel using the same rim-mounted brake has rub. A fair comparison should match weight and/or price — if the hub-type wheel (and fork/frame, etc.) is much lighter, then add rim stiffness until the weights/costs are similar. A stiffer spoked wheel may wander less and so work without rubbing.

Drivetrain types

Centerless/hubless wheel drivetrain ideas include:

• Chain or belt drive to a large sprocket attached to the rim. Example: [The Q].
• A ring gear facing towrds the center; a spur gear to engeage it. Example: [Reevo].
• A friction roller which runs on the inside of the rim. Example: [Sada].
• A tire made of two rings, with an external belt running between them [Arthur].
• Electric, where pedals drive a generator to make electricity, which powers a electric motor to drive the wheel.

A chain or belt needs to clear the tire, so the sprocket on the rim needs to be to the left or right of the tire. The sprocket is nearly as large as the tire, so runs close to the ground. In ordinary riding, you may ride close to rocks, curbs, etc., which can hit and damage the sprocket or chain/belt. Making things stronger can resist damage, but adds weight and cost. Ordinary chain/belt drives also run outboard of the tire, but are much higher, so less likely to hit things.

A chain or belt that drives a rim-size sprocket needs to be roughly 3x as long as one driving a hub-size sprocket. The chain or belt for a rim-size sprocket is under much less tension, so maybe it can be of lighter construction to offset 3x length. However, making it lighter also makes it more fragile.

A ring gear can sit inside the rim, and so when you ride close to stuff, it hits the tire and not the rim — just like a conventional hub-type wheel. Gears are often more sensitive to alignment that chains and belts; that may add cost/weight/etc. Gears often need to be made more precisely than sprockets, which can add cost.

With centerless/hubless wheels, both chain/belt and ring gear drives have parts which run close to the ground, so are more likely than conventinal drives to be bathed in dust and mud or immersed in dirty water. That can lead to higher drag and faster wear than a similar chin/belt/etc. used to drive a hub.

Gear drives are often durable and have low maintenance. Many Sturmey-Archer hub gears from the 1940s have been used daily for decades with little maintenance other than periodically adding a little oil. However, gears often wear much faster when dirty. Sealing a centerless/hubless ring gear against contamination is hard because the seal needs to be large — roughtly the size of the rim. Seals thus tend to either high drag or inferior sealing.

[Yale demo] uses a ring gear made from a toothed rubber/fabric belt. Although it is exposed, toothed belts often give adquate service, even in dirty riding. It has the potential for good service, it is not clear if it actually does: attaching the belt rigidly to the rim may change how it wears. It runs at e.g., 10x larger diameter than a normal belt drive — which means it runs at 10x lower force; but also 10x higher rate of tooth engagements, which in dirty service might cause rapid wear. Or not — tooth belts in clean service often run high power at high speed and give good durability. It is unclear what is the durability of a design like the Yale bike.

Friction drives are prone to slip when wet. Friction drives often wear quickly when they slip, and especially when they slip and are dirty. Some friction drives are durable even under dirty slip: railroad wheels are friction drives, and often run very high forces in dirty conditions with 1% or more slip. However, “heavy steel-on-steel railroad wheels wear well” does not tell you much about light rubber-on-aluminum bicycle friction drives.

Friction drives often have poor efficiency. That would appear as increased rolling drag. Although they often have poor efficiency, friction drives can have good efficiency: railroad wheels (when slip is low) can be more efficient than gears. However, railroad wheels may not tell you much about bicycle drives.

An electric drivetrain uses pedals to drive a generator to produce electricity; the electricity then runs an electric motor to drive the wheel. When combined with a battery to make an e-bike, it is often called a series hybrid drivetrain.

I am not aware of any example centerless/hubless wheel proposals using an electric drivetrain, but included it in this disucssion because

• Some people online say they read the [Cyclotron] proposal as using an electric drivetrain. That is not my reading, but I might be reading it wrong. (There are suggestions online the Cyclotron was a scam; in which case there might not be any correct reading.)

• Many centerless/hubless wheel proposals are for e-bikes, which already have a motor. All the other centerless/hubless wheel drivetrains are problematic, so it seems worthwhile to consider an electric drive — and it seems only a matter of time until it comes up.

Electric generator/motor drives are used widely and can be very efficient and durable with low maintanence; but typically with much heavier construction than in bicycles. Some notes:

• There are prototype bicycles and e-bikes with generator/motor drive, but they have (or seem to have) much lower efficiency than a chain drive. I do not know of much test data, but I have seen suggestions of around 60% efficiency for motor-generator, 80+% efficiency for chain+hub gears, 90+% for clean derailleur gearing, and 95+% for clean single-speed chain or belt.

• Lower efficiency means slower riding. For some riders and riding, that will be unacceptable. For other uses, it may be okay — 60% efficiency might still let you ride at 75% of the speed; and an electric drive can offer a wide range of “gear ratios” which can be adjusted automatically and continuously, making the bike much more rideable for casual riders — and possibly faster than a bike with conventional gears, as casual riders often just leave the bike in one gear.

• Why do I claim poor efficiency is okay for an electric drive, but similar-size friction losses are a deal-breaker for centerless/hubless wheels? Because an electric drive has advantages which can help it to “pay” for worse efficiency. Whereas the higher friction of a centerless/hubless wheel is just just a drag, without offering you much (except looks).

  A similar argument applies to [Airless tires]: they ride slower and are less comfortable. But for some riders and riding, never-flat is more important than fast or comfortable.

• An electric drive might also be sealed effectively from the environment, leading to lower maintenance and higher availability than other choices. Notably, a direct rim drive motor like the [Ujet] looks to avoid all the dirt/water sealing problems associated with both chain/belt and ring gear drives.

• More info: Andreas Fuchs, Series Hybrid Drive-System: Advantages for Velomobiles, [https://hupi.org/HPeJ/0015/SeriesDriveHybridVelomobiles.pdf as of 2025-03] Also [https://en.wikipedia.org/wiki/Hybrid_vehicle#Vehicle_types as of 2025-03].

Electric drives are likely both heavier and more expensive than other drive types. For an e-bike, there is already an electric motor, although it may need to be up-sized to handle rider+battery power instead of just battery power. An e-bike has heavy batteries, which means a smaller percentage weight gain from adding a generator. The cost and weight of a pedal-driven generator are likely similar to cost and weight of the electric motor.

Aside: many places have laws which require on-road bicycles and e-bikes have a direct mechanical connection (such as a chain or belt) from the pedals to the wheel. Even if there is no battery, an electric-only connection may be categorized as a motorcycle, or might not be permitted as a bike. Some regulations may allow a bicycle (with no battery or other energy storage) but not an e-bike. Regulations vary a lot by region; can change; and anyway this is an aside, not a technical concern. However, regulations might discourage use/adoption of electric drives.

Aside: many online articles refer to electric drive as a “digital drive” or sometimes a “chainless digital drive”. This name seems to be a failed attempt to sound smart or trendy. But is bad term because it is misleading in several ways. Including:

• A belt drive is “chainless”. A penny-farthing is “chainless”. A lever/treadle drive is “chainless”. A geared shaft drive is “chainless”. The Ceramic Speed shaft/roller drive is “chainless”. Various cable, string, hydraulic, cam, and other drives are “chainless”.

• Everything is digital. There are digital derailleurs — for, you know, chains. There are digital hub gears. Combine digital hub gears with a belt or shaft drive, that is also a “chainless digital drive” You could build a penny farthing which automatically adjusts the crank length according to your cadence — behold, the chainless digitial drive penny farthing!

Everybody is, of course, free to call it whatever they want. Just be aware if you use the term “digital drive”, people may assume your understanding is limited to whatever is in a recent shiny press release — possibly a press release from somebody who themselves has a limited understanding.

What is a centerless/hubless wheel, anyway?

One way to view a “hubless” wheel is that it simply has a large hub with a large hollow axle.

Ordinary bicycle wheels have a hollow axle so the quick release or thru-axle can hold the hub in the frame. If you make the bearings bigger, you can make the hollow axle bigger.

A centerless/hubless wheel is taking that to an extreme where the hole in the middle is large enough that it is at least half the wheel’s diameter. (I made up “half”, there is no specific size requirement.)

Another view is that a rim-on-rollers design like the [Reevo] is different, because it does not use a central bearing. However, you could build a hub which runs on rollers instead of bearings. Probalby there are patents proposing just that, but I did not check.

A “structural rim” design [Wheel structure types] has rollers that support only part of the rim, and the rest of the rim is cantilevered. For example, the [Terpstra] and [Sada]’s front wheel. You could also build a hub that way, although it seems less likely anybody would.

So maybe “structural rim” is a truly “hubless” design, and not just a big hub. However, many things which apply to a big hub also apply to a structural rim design: bearing and seal drag, and so on. Going to a “hubless” wheel does not suddenly change what is needed, nor does it suddenly ease drag/cost/weight/durability issues compared to a big hub design.

Put another way, thinking of a “hubless” wheel as a large hub with a large hollow axle might be a little bit misleading, but at the same time can be useful to help to point out many issues. And, “hubless” still has problems similar to a large hub with a hollow axle.

Wheel structure types

There are several types of centerless/hubless wheels. There are no standard names, so here are some I made up, based on what structures the wheel has for carrying loads:

• Support hoop
• Structural rim
• Support rod

        
From [https://odditymall.com/diy-hubless-fat-tire-bicycle as of 2025-03] [https://www.startupselfie.net/2023/12/17/penn-e-farthing-hubless-wheel-penny-farthing-ebike as of 2025-03] [http://www.sadabike.it/en as of 2025-03]

In more detail:

• Support hoop

  There is a non-rotating hoop inside the rim. Loads are transferred fairly directly through bearings or rollers from the rim to the support hoop.

  The rim carries some load, so is not really a non-structural rim. But you can use a relatively light rim.

  Examples: [Gafoor et al.], [Reevo], [The Q], [Top Secret].

• Structural rim

  There are large spans where the rim is free-standing, without any bearings or rollers to support it. The support rollers are away from the tire contact patch, so the rim needs to be strong enough to carry all tire/wheel loads across a long span.

  Note that a standard bicycle rim is only strong when when laced in to a wheel. If you take a freestanding rim, place it vertical, then sit on the top, the rim will collapse. Take an identical rim and lace it in to a spoked wheel, and the rim plus spokes and hub support your weight.

  In a “structural rim” centerless/hubless wheel, the rim needs to be strong enough carry loads from the ground to the support rollers, but without using spokes. Thus, it needs to be stronger than a conventional rim. It also needs to be stronger than a rim for a support hoop centerless/hubless wheel.

  Getting rid of the support hoop saves the weight of the hoop, but adds weight to the rim. A structural rim wheel may still be lighter overall, but the weight difference is more complicated than just subtracting the weight of the support hoop.

  A structural rim is supported by rollers, which are only on one part of the wheel. The rim and rollers thus need to engage in a way that supports loads in all directions. For example, the Terpstra has grooves on either side of the rim, and guide rollers run in the grooves. The grooves mean rollers on the sides of the rim can also transfer radial loads.

  If the rim is wider than the tire, then rollers can run on the tread side of the tire to support radial loads. However, a wide rim is more likely to get damaged when rolling close to curbs, rocks, and so on.

  Rollers can instead run directly on the tire, but this is likely to have much more drag. Also, the tire is flexible, so rollers on the tire may allow the wheel to wander excessively.

  Structural rim examples include [Terpstra], [Nulla], and the front wheel of the [Sada].

  A structural rim may put much higher loads on the rollers than other approaches.

  • For example, suppose the [Sada] front rim roller tracks are is 50 mm wide, and the tire contact with the ground is 500 mm away from the rollers. A side load of 10 kgs at the ground turns in to a 100 kg force on the rollers.

  • Modest side loads could therefore stress the rollers and roller supports more than the rider’s weight. The rim, rollers, and supports can be strengthened to support these loads, but it adds weight. And, modest side loads may lead to much more roller drag, because a 10 kg side load is as-if 100 kgs extra weight is being supported by the rollers.

  • The higher forces can also cause faster wear of rim and rollers.

• Support Rod

  An intermediate approach is to use a support which reaches across the wheel to proide support near the tire contact patch. Thus most loads are carried directly from tire tread to rim to roller to rod to frame, rather than indirectly around a circle, as in support hoop or structural rim.

  • The wheel may have a “center”, but it is non-rotating. So it is different than a conventional wheel.

  • A support rod holds rollers near the bottom of the wheel, where they support the rim directly. This allows rollers placed similar to a support hoop, and so allows use of a less-structural rim than is needed for the “structural rim” approach.

  • Alternatively, a support rod can hold one part of a large bearing. It acts somewhat similar to a support hoop, but transfers loads directly across the center of the wheel, rather than indirectly around a circle.

  • A wheel also needs support against side loads and fore/aft motion. These loads are much less, so the structure to hold them can be lighter than for the main load, the rider’s weight.

  The structural rod load path is more direct than with a support hoop, so a support rod may be lighter. However, the more-direct support reduces the hole size. Many examples above use centerless/hubless wheels for looks; a structural rod changes the look significantly. If a centerless/hubless wheel bike uses the space — e.g., for cargo, or to hold a folded bicycle — a structural rod may have functional consequences.

  Some “support hoop” wheels use a single large bearing [Top Secret]. You can use a large bearing with a support rod, provided that the main load points are well-supported, so that the bearing does not need to carry loads where it is unsupported.

  The [Blood Falcons], [Sada] rear wheel, [Thorpe], and [Twist] front wheel, use a support rod.

  Arguably, the Black Hole is also a support-rod design, as the there are only 4 rollers, placed in a “peace sign” layout. The Black Hole’s center is filled with material for aerodynamic reasons, but only some of it needs to be structural. However, it is unclear if the main load path is across the center of the wheel (structural rod) or around the circumeference (support hoop).

There are not standard names for different kinds of centerless/hubless wheels, so I made up the above. They seem useful, but there may be better categories and names.

Fork weight

Centerless/hubless wheels need a special frame and fork — they do not directly replace existing hub-type wheels. Thus, the weight of centerless/hubless wheels is really the weight of the wheels and the frame/fork structures which support them. And, they should be compared with hub-type wheels plus their frame/fork.

Consider a “support hoop” wheel like the front wheel of the [Reevo].

• The Reevo’s support hoop looks like a heavy casting. A support hoop might instead be made of an extrusion, similar to how ordinary rims are made. However, an ordinary rim gets much of its strength from being laced in to a spoked wheel. A support hoop made like a rim has no spokes, so needs to be heavier than an ordinary rim.

• And, the support hoop weight is in addition to the rotating rim of the centerless/hubless wheel.

  You could think of the Reevo as using two rims for each wheel: a rotating rim, and a non-rotating rim. Whereas a hub-type wheel only needs a rim, spokes, and hub.

  Alternatively, you could think of the support hoop as being part of the fork. However, the total weight is the same either way.

• A support hoop connects to the frame via an “unfork”. The Reevo’s unfork is shorter than a conventional fork, but the part of the fork which is removed is the lower part of the fork legs. That is the lightest part of the fork.

The rear wheel is similar, but the connection to the frame is cantilevered. In a conventional frame for a hub-type wheel, the rear axle is supported by a tetrahedron, which is an efficient structure — light compared to the loads it carries. The cantilevered rear support of the Reevo is likely heavier than the rear frame for a hub-type wheel.

A “structural rim” wheel is broadly similar. The support hoops are replaced by a roller support structure, but this is also likley heavier than a conventional fork, etc.

A “support rod” wheel is probably lighter than support hoop or structural rim options. That said, roller supports and rollers add weight, and the rim needs to be strong enough without spokes to support normal loads. It seems likely to me that even structural rod is heavier than similar-cost hub-type wheels and fork/frame.

The rear also needs a drivetrain, so the frame and “rear untriangle” need to support it.

• For a conventional hub-type wheel, the added structure for the drivetrain is basically heavier chainstays and fancier dropouts. A diamond-frame hobby horse, with no drivetrain, could use lighter “chain” stays. And, a hobby horse does not to adjust chain tension, so no need for horizontal drop-outs or a mount for a chain tensioner/derailleur.

• A centerless/hubless wheel with chain or belt drive runs at lower tension than a hub-type chain or belt. Lower tension means the corresponding frame structures can be lighter.

• A centerless/hubless wheel with chain or belt drive runs at higher speed. There is not enough room for a much larger chainring, so some kind of step-up gearing is needed — e.g., a gearbox or 2-stage chain drive. The frame needs structure to support the step-up gears, which adds weight on top of the weight of the step-up gearing.

• A ring gear drive also needs to attach to the frame somehow. And a ring gear and spur gear tend to pry themselves apart under load. The structure needs to be stiff and strong enough to resist the pry-apart forces — although it may vary with the speicific design whether these forces are handled directly by the geartrain (adds geartrain weight but not frame weight) or by the frame (adds frame weight).

• For these reasons, it seems likely a drivetrain for a centerless/hubless wheel will be heavier than a similar drivetrain (same number and range of ratios) for a hub-type wheel. Chain or belt tension for a centerless/hubless is less than for a hub-type wheel, which may allow a lighter frame. However, the centerless/hubless drivetrain needs to do more — e.g., support step-up gears. So it seems likely a centerless/hubless wheel’s frame structure will be heavier than for a hub-type wheel.

E-bikes have less incentive to save weight: they are heavier anyway, and have more power, helping to mask the added weight. You can partly compensate for higher weight by adding battery capacity and making the motor bigger. These additions also add cost, so [a fair comparison] means comparing against a more-expensive hub-type e-bike.

The discussion above suggests that the fork and frame for centerless/hubless wheels are heavier and/or more expensive than for hub-type wheels. That is fine if you are paying extra for looks. If centerless/hubless wheels are being sold to you as being useful for something, what advantage do they offer which makes them worth higher weight and cost?

How do you mount fenders and racks?

For ordinary riding, you probably want fenders and racks. Or, at least, you want the option to use them. Even racers often want to use fenders during training rides. Fenders and racks usually attach to eyelets at the dropouts. Forks and frames for centerless/hubless wheels do not have dropouts, nor eyelets at the dropouts. So how do you mount fenders and racks?

Some centerless/hubless wheels, like [Reevo], use a non-rotating support hoop. You can attatch fenders and racks to that.

• In theory, you can attach fenders and racks to the support hoop. However, many examples above have non-structural shrounds around the support hoop structure, and the shrouds are probably too weak to hold a fender or rack.

• Even if you can get at the structural hoop, it might not be possible to attach things: is there already a mounting point? If not: where is it safe to drill and tap a hole, do you have the tools/etc. to do that, and does making a hole void the warranty?

• Fenders often attach with stays, which can be bent and shaped as needed. Racks carry a lot more weight, and are often much less adjustable. If you cannot find a suitable rack, you have to do without, even if there are mounting points. Does the bike vendor offer racks? Are they of a type suitable for your needs?

• There are retrofit “quick release” racks which mount to the seatpost. Most have a much lower weight limit than a conventional rack. They may be fine for carrying lunch and a jacket, but many are too weak to carry a normal load of groceries.

  Some are prone to shift side-to-side when loaded; it makes them more hassle to use, or added cost to keep buying/trying racks to find one that is stable.

  Many seatpost racks are much higher than a regular rack. Placing a load higher can make the bike to flop around more when you are stopped.

  Some people dislike the “afterthought” look of retrofit seatpost racks. If you are buying a centerless/hubless wheel for looks/style, then putting a goofy-looking rack on it may be unappealing. A regular bike with designed-to-fit racks and fenders is less exotic, but may be more appealing — and also work better.

       
  From [https://road.cc/content/review/6415-etc-seatpost-mounted-qr-pannier-rack as of 2025-03] [https://www.ilikeyourbike-shop.com/products/velo-orange-constructeur-bike-front-rack as of 2025-03]

Some centerless/hubless wheels use structural rims, which have no place to mount fenders and racks. You can, in theory, have cantilevered supports for fenders and racks, but I have not seen it discussed or proposed.

Do you care about fenders and racks? Many people do, many do not. Some talking points:

• Better fender mounts let you use better fenders, and good fenders protect you a lot more than poor fenders.

  Good fenders/mudflaps help you stay drier and cleaner with almost any level of rain gear. Even if you have a rain suit, you may get more wet without fenders.

  Also, with poor fenders (or no fenders), the suit gets covered with dirty/gritty water splashed up from the road. Stowing and cleaning a dirty rain suit is more of a chore than dealing with one which is merely wet.

• It can be hard to carry bulky and heavy items with only a backpack. A basic rack and pannier can allow you to carry 20 kgs of groceries/etc. But it needs a rigid mount. A floppy mount encourages the bike to wobble and shimmy. Many “easy on” racks are heavier than a regular rack, yet are only rated to carry light loads.

• Commuting and shopping in warm weather can be much less pleasant when wearing a backpack. A lot of bike riders use a basket or similar so they can carry a backpack to use off the bike at the ends of the trip.

Centerless/hubless wheels should work with fenders and rack. But a bike maker needs to offer sensible mounting points, and may need to offer racks to fit their bike, if nothing else fits. Are their racks good, or are they flimsy? If a rack is either week or flexible, it can make it hard to carry normal loads.

Any bike with centerless/hubless wheels can be designed for fenders and racks. There is nothing about centerless/hubless wheels which prohibits them. However, fenders and racks are a (very) practical issue, yet only some vendors address it. And “address it” might be fender and rack options which are poor at their job. If it looks like a fender but does not work like a fender, it is not really a fender.

Centerless/hubless wheels can use fenders and racks. But it is a “bad idea” when a bike is sold as “practical”, but you cannot use fenders and racks, or the options on offer are bad.

And if the seller cannot do fenders/racks right, what else did they get wrong? If they claim their centerless/hubless wheels are practical... well, maybe it meets their needs, but it is not very practical for you if it does not meet your needs.

Wheel aerodynamics are still incomplete

There are discussions above about wheels with “better” and “worse“ aerodynamics. Another complication is that we are still learning how to model and measure wheel aerodynamics. Thus something we currently think is better may turn out to be worse.

Here are two sources of uncertainty. ”Uncertain” means it might cause an error, or it might not.

• Many wind tunnel tests use direct headwinds. Measured this way, a specific wheel may show a benefit. But off-axis headwinds (headwind plus modest crosswind) can show the same wheel has more drag than some ”less” aero wheel.

  Off-axis headwinds are probably more common: only one wind direction gives you an on-axis headwind, but many wind directions give you an off-axis headwind. Some off-axis tailwinds can have an effect like an off-axis headwind.

  In turn, some ”better” aero wheels may in practice have more all-around air drag than some other ”worse” wheel.

• Many simulation models assume non-turbulent air. It is a reasonable assumption for high-speed/high-power vehicles like cars and airplanes. They go faster than most air, and also compress the air substantially and force certain air flows to dominate, even if the air starts out somewhat turbulent.

  Low-speed/low-power vehicles like bicycles have less effect on the air, and thus air flow may be more affected by the air’s initial turbulance than it is for high-speed/high-power vehicles.

  Further, bike riders often ride past bushes, buildings, and many other things which create turbulent air flows. An airplane may fly through non-turbulent air, but bikes are often riding through turbulent air.

  Thus, turbulent air may lead to large model errors for bikes. But ”how much” and ”when” (what kinds of turbulence) are not yet well-understood.

  Also, the amount of error in a situation (riding speed, air speed, kind of turbulence) likely depends on the wheel design.

  Again, some ”better” aero wheels may turn out to have more drag when you consider the many kinds of wind that you encounter, not just a few conditions which are easy to simulate/model.

The discussion above about centerless/hubless wheels assumes that we even know what is a ”good” wheel. But it may turn out that we are comparing the wrong wheels. This could either help or hurt ”no-spoke“ centerless/hubless wheels, we do not yet know.

Also, better designs may be possible once we understand more about what makes a good bike wheel. For example, Nullwinds proposes a spoke which is round for most of the length, but flattened (aero) near the rim [https://nullwinds.com/pages/vehicle-aero-aero-spoke as of 2025-03]. Their reasoning is round spokes have less all-around aero drag considering winds from many directions... except for high spoke speeds: if the spoke is moving fast enough, then an aero profile is faster even for off-axis winds. For bicycle wheels, they say only the part of a spoke near the rim is going fast enough so a non-round aero profile has lower all-around drag than round.

I have no way to verify (or refute) their claim, but it is an example of how a better understanding of air drag details can lead to better (maybe) designs.

At the same time, we should be somewhat careful about claims about the aerodynamics of centerless/hubless wheels. It seems obvious that getting rid of spokes gets rid of spoke drag; but “how much” is unclear. And in part because we are still learning about the aerodynamics of ordinary spoked wheels.

More:

• [https://www.hambini.com/modern-real-world-cycling-aerodynamics as of 2025-03]

Airless tires

One “bad idea” aspect of centerless/hubless wheels is they add significant drag.

Airless tires also add a lot of drag. Airless tires are, however, a good idea: the most common bicycle failure is flat tires. For some riders and riding, flats are common enough that the time spent dealing with them significantly hurts average riding speed. Also, flats are unpredicatble, so they make cycling a less-relible way to get places.

Why is “adds drag” okay for airless tires but not okay for centerless/hubless wheels? Because centerless/hubless tires add drag and other problems (cost, weight, poor durability); and in return give you: looks, and not much else. Airless tires add drag and weight and make ride comfort worse, but do give you something in exchange for the added drag.

Airless tires offer

• no time spent on flats;
• more-predictable trip times;
• no need for periodic tire inflation;
• maybe longer intervals between tire changes.

Unfortunately, airless tires add a lot of drag, and also hurt the ride quality.

• Ordinary pneumatic bike tires can take 20-70 Watts for both tires at 30 kph, 75 kg bike+rider, on a smooth road.

  Tire drag depends mainly on tire construction details: how the casing is made, does it have a flat-resistant belt, is the tread thin and smooth or thick and knobby, and so on. It also depends on the tire size and inflation pressure, where an over-inflated tire may have more drag than a properly-inflated tire.

  20-70 Watts is both a very wide range, and also a large fraction of an ordinary rider’s steady power output, which might be 100 Watts.

• Airless tires may add tens of watts to that. “May”: I do not know of any reliable measurements or comparisons.

  Some makers claim added drag of less than 10%. But 10% of what?

  • Do they mean (20 Watts + 10%) = 22 Watts? Do they mean (70 Watts + 10%) = 77 Watts?

  • Suppose a typical pneumatic tire is 45 Watts.

  • Then 22 Watts would be among the best tires made.

  • But switching from a 45 Watt typical tire to a 77 Watt airless tire would add 32 Watts. For a rider putting out 100 Watts, they would have to go up to 132 Watts, or 1/3 more effort, just to keep the same speed.

  Presumably, makers have reliable test data, which is what they use for making a “10%” claim.

  But they provide no data, so their claims may be meant to mislead you.

  Data should be something like [Watts] at [speed, load weight, road surface].

• Another page claimed an extra 70 Watts for a specific airless tire model [https://forum.slowtwitch.com/t/tannus-airless-tires/790249/4 as of 2025-03], compared to Continental GP4K tires.

  It is not clear if that number was repeatable, what were the test conditions (20 kph, ?load weight?, ?road surface?). It appears “power” was measured as rider power, which is somewhat harder to measure reliably than rolling power measured with a dynamometer; but a rider power measurement does include “jiggling flesh” power (see below).

• Airless tires also give a rougher ride.

  A rough ride is a comfort issue, but it is also a drag issue: a rough ride jiggles your flesh. That means rolling energy is used to jiggle your flesh, and some of that energy is lost as heat. The “jiggling flesh” loss is small on smooth pavement, but significant when things are rough. And airless tires are often proposed for urban/city use where streets often have rough pavement.

Airless tires show up often enough, but they seem especially common for centerless/hubless wheel designs.

• It may be the same folks inspired to make centerless/hubless wheels are also interested in reliable transit. If so, centerless/hubless wheels are a weak link in their thinking: centerless/hubless wheels are almost sure to be less reliable than hub-type wheels.

• Or, people proposing centerless/hubless wheels may be stuck trying to figure out where to put the valve of a pneumatic tire. Airless tires could simply be a workaround.

Pneumatic tires can be fitted to centerless/hubless wheels, so problems with airless tires are not an inherent flaw of centerless/hubless wheels.

• Ideally, each rider can consider their specific needs, and then choose whether to fit airless tires or pneumatic tires.

• However, pneumatic tires can only be used if they are part of the centerless/hubless wheel design. If you got a centerless/hubless wheel with no way to mount pneumatic tires, then you are stuck with airless tires — along with their high drag and rough ride.

• And adding tens of Watts of drag for centerless/hubless wheels and then more tens of watts for airless tires seems unfortunate. Even for an e-bike, the change may be significant. If some e-bike has a typical power of 250 Watts, and 50 Watts of of that is extra wheel and tire drag from centerless/hubless wheels and airless tires, that 50 Watts probably cuts the e-bike’s range by about 20%.

In other words, airless tires can be a very good idea for some riders and riding. But a centerless/hubless wheel design which prevents you from using a pneumatic tire is a bad idea.

Speculation: why is hubless appealing?

Maybe centerless/hubless is appealing in part because

• it’s different;

• many riders have had problems with spoked wheels: rim wobble, broken spokes, bad hub bearings, etc.;

• spoked wheels are a little bit mysterious: if a spoke breaks and you put in a new one, the wheel might still not run true; and the truing process is complicated.

Ordinary wheels

I note the second problem above — spoked wheel durability — is often an issue with makers trying to save a little weight and cost. Spoked wheels can be more durable.

Maybe (maybe!) centerless/hubless wheels would be less appealing if regular wheels had fewer problems.

A typical approach in the bike industry seems to be: build wheels with enough strength and durability that they give several years of service before problems start to show up. But: not more than is needed for “several years”.

• This approach leads to wheels that are somewhat lighter and cheaper. For example, a 32-spoke wheel is easier to assemble than a 48-spoke wheel. Most high-volume wheels are assembled by wheel-building robots, but it takes longer for a robot to assemble a 48-spoke wheel.

• This approach also leads to wheels that often develop problems after a few years of use.

Wheel durability can be improved a lot with:

• A stiff rim
• A thicker and more durable spoke bed
• More spokes
• Thinner spokes
• Better shielding of hub bearings

And for bikes using rim brakes:

• A thicker brake track

Doing all of these increases weight, cost, and air drag. But not lots. At the same time, they make a wheel which is both

• More durable: it lasts more years of ordinary use.
• Stronger: it can resist bigger loads without failure or loss of true.

My hand-wavy estimate is for typical bikes, doing all of the above adds about 300 grams per wheel (600 grams per bike) and a few tens of dollars (e.g, US$20).

$20 could be 5% higher cost for ordinary bikes (US$400). But: they could do only some of the above and still get some of the benefit. And for more expensive bikes, US$20 is less than 5%.

For those who are interested, a bit more detail.

First, how a wheel carries (some) loads is a bit odd. This “how a wheel works” is background to help explain “why make those changes“.

• The main load on wheels is radial: carrying the rider’s weight from the hub to the ground. When you put weight on the hub, spokes near the bottom of the wheel lose tension (get slacker). More load means more loss of tension.

• When you roll along, each spoke loses some tension and then rebounds once per wheel revolution. That is, as the wheel turns, each spoke briefly points at the ground; as it approaches, the spoke tension goes down some; as it moves away, it returns to full tension.

  This tension reduction is how the wheel carries load: if you take an unloaded wheel and put a load on the hub, the only thing carrying the load from hub to rim is the spokes. Some spokes pull the hub up, some pull down, some forward, some back. If you then put a load on the hub, the tension changes is some spokes. If you add up the change in tension for all spokes, it adds up to the load on the hub.

  Which spokes change tension, and how much? All spokes change tension a little. For most wheels, the main change is lower tension in whatever spokes are currently pointing at the ground.

  Your intuition may suggest tension should go up in the top spokes. But how a wheel carries load is odd. For most wheels, the only spokes with a significant change is lower tension in spokes pointing at the ground.

• As the wheel turns, some spokes stop pointing at the ground and regain tension; while other spokes start pointing at the ground and so start to lose tension.

  If the tire diameter is 637 mm, the circumference is 2 metres, so riding 1 km = 1000 m rotates the wheel 1000/2 = 500 times. If you ride 5.5 kms per day (and many riders ride a lot more than that), that is 2000 kms/year or 1,000,000 wheel revolutions in a year. 10 kms/day? More like 2 million wheel revolutions per year.

• That means each spoke has some tension loss then returns to tight a million times a year or more. Let’s call this “working” the spoke. All materials have microscopic defects, that includes spokes. Working often causes microscopic cracks to grow. With enough working over enough millions of cycles, cracks grow to where the spoke breaks.

  Rims have a similar problem around the spoke holes. Many rims fail due to cracks near the spoke holes. Those cracks come from millions of cycles of the spoke changing tension: when spoke tension changes, it works the rim. Microscopic cracks can grow until they are large enough the spoke and nipple pull loose from the rim.

• One name for this “crack growth” kind of failure is “fatigue”. You may read “fatigue failure” in some discussions. That refers to failures like this, which happen with modest loads over many cycles.

  There are also “strength” failures, where a much larger load is applied a few times — maybe just once. If you put a big enough load on a brand-new bike, the wheel will collapse, even if you don’t move the bike. That is a strength failure, not a fatigue failure.

  Strength failures are also an issue for bikes — carrying very heavy cargo or landing from a big jump. Also landing a small jump sideways, which puts a sideways (“lateral”) load on the wheel. Bike wheels are much weaker against sideways loads than radial loads.

• When you put a load on the hub, spokes near the ground go slack. As you put bigger and bigger loads on the hub, the spokes get looser and looser until some of them go completely loose.

  If you keep adding load, the spokes which are already loose cannot get any looser, so they cannot carry any more load.

  Some load can be carried along the rim to spokes which are close to the spokes pointing at the ground. This can slacken more spokes and carry more load. But only up to a point: eventually the rim buckles near the ground.

• Spokes pull towards the hub, but they also pull sideways. If you push sideways on the rim, spokes near where you are pushing will get tighter on one side and looser on the other. That is how the wheel resists a sideways force.

  However, if the spokes are already loose due to radial load, then they cannot get any looser. So they cannot carry more sideways load.

  Loose spokes on the “get tighter” side can regain some tension. That helps, but is only half the spokes.

  Also, the rim has to move sideways in order to re-tighten those loose spokes. Moving the rim sideways far enough makes the rim unstable and it collapses to a “potato chip” or “taco” shape. That is, the whole rim changes shape, rather than just buckling near the ground.

  To say that another way, the bigger the radial load, the easier it is to move the rim sideways. Move the rim sideways far enough, and the wheel collapses. So the bigger the radial load, the easier it is to collapse the wheel sideways.

• Most strength failures are “potato chip” failures, where there is a big radial load, but not big enough to buckle the rim; and then also some lateral load which pushes the rim sideways and in to the “potato chip” shape.

  Lateral (sideways) wheel strength is improved if you can keep spokes from going entirely slack near the ground. If they have some remaining tension, they can support at least some sideways loads.

The above “how a wheel carries load” leads to some observations about what builds a stronger and more durable wheel:

• Many failures are durability failures rather than strength failures. Strength and durability are only somewhat related. In other words, you can build a wheel which is any combination of strength and durability:

  • stronger and more durable;
  • weaker and less durable;
  • stronger and less durable;
  • weaker and more durable.

  Often, “stronger and more durable” means a heavier wheel. Whereas you may be able to trade strength and durability: if your rims often crack at the spoke holes, you need more durability. Since they fail from cracking, they are not failing from buckling. In turn,you can probably give up some strength, to get durability without adding weight.

  Conversely, if you mostly collapse wheels, maybe you can give up some durability to get a stronger wheel. If your wheels never fail — or only fail because the rim’s brake track wears out — maybe you give up both strength and durability, to save some weight.

• For this discussion, I observe that mass-production bikes are used by a wide range of riders for a wide range of riding. Some riders will have no problems with the wheels; others will have strength failures; yet others will have durability failures.

  So my proposal is to add a little bit of weight (and cost), so both strength and durability are improved. Overall, that should give more riders more trouble-free service.

• A stiffer rim tends to share load among more spokes at the same time. Notice this is stiffness and not strength. A rope can be strong but not stiff. A dry stick can be stiff but not strong. For better load sharing, we want more stiffness, but do not need any more rim strength.

  More sharing means each spoke is worked less on each wheel revolution, which is good for wheel durability.

  A stiffer rim also means there are more spokes sharing high loads. So it takes a bigger load before spokes start going completely loose — so a stiffer rim is good also for wheel strength, even without any more rim strength.

  You can have a rim which is stiffer but not stronger, or stronger but not stiffer. Things which improve one tend to improve both, but often unequally. You often get more wheel strength and durability benefit from rim stiffness then from rim stregth, hence the emphasis here on rim stiffness.

• A thinner and more springy spoke has similar effects to a stiffer rim. It’s not exactly the same, but can help in some of the same ways. The spoke is a spring: when you put a load on the rim, the spokes get looser because the rim moved slightly. If the spokes are thinner and springier, then you have to move the rim further before the spoke loses tension. Because the rim has some stiffnes, when you move the rim further, it also shares load with more spokes.

  This is a nice “cheat”: a thinner spoke weighs less, but can be more durable. And, maybe stranger, it takes a higher rim load before the spoke goes completely slack, so a thinner spoke sometimes helps resistance to “potato chip” failures.

  There are limits, though. And thinner spookes may make the wheel weaker against local buckling. Most strength failures are potato-chip, and not local buckling, so it is often a good trade-off. But it is not entirely free.

• Thinner and more springy spokes mean more of the spokes pointed at the ground lose tension; but since the load is spread over more spokes, each spoke has less loss of tension.

  That helps reduce “working” of the rim’s spoke bed, so can help improve rim durability.

• More total spokes also helps share a given load among more spokes at one time. That also means each spoke is worked less per revolution.

  More spokes also means that it takes more total wheel load before spokes go completely loose. So more spokes increases wheel strength.

  So “more total spokes” is good for spoke durability, wheel strength, and rim durability. Often for similar reasons as “springier spokes”, and “stiffer rim”, above.

• All else equal, adding spokes adds cost, weight, and also air drag. Whereas a stiffer rim often adds cost and weight; and thinner spokes often adds cost but reduces weight and air drag.

  Just making the rim stiffer seems nice because it avoids adding air drag. But the weight may go up faster than the strength and durability gains from also adding spokes. Just going to thinner spokes seems nice because it removes weight; but without a stiffer rim and/or more spokes, it weakens the wheel against some loads.

  So “all of the above” is often a good way to balance the various costs and benefits.

  That said, a wheel with more spokes takes longer to build, even for a robot. Thinner spokes also can increase wheel build time, and thin-in-the-middle spokes (see “swaged spokes”, below) are more expensive than straight spokes. Just going to a stiffer rim may not get you as much, but may be a good compromise for low-cost wheels.

• A stiffer rim, more spokes, and springier spokes can also help with hub flange durability. Most modern hubs have good flange durability, so this may be a small benefit.

The list just above should explain the recommendations. Some costs:

• A stiff rim has a cross-section which is taller and/or wider than a less-stiff rim. A stiff rim helps spread loads over more spokes, which helps both durability (many years of riding) and strength (strength against large loads). A stiff rim has a larger cross-section, which adds material and weight. One way to save weight is by using thinner walls. But it is somewhat more expensive to make a rim with thin walls.

  Thus, adding a lot of stiffness without adding much weight or cost can be hard. Even if cost is not a concern, it gets harder and harder to add stiffness without adding weight.

• A thicker and more durable spoke bed helps prevent rim cracking around the spoke holes. It also adds weight. Some of the added weight can be reduced by removing material between spoke holes, but removing material is more work, so adds cost.

  Note that wwith rim brakes, the brake track often wears out before the spoke bed fails. However wheels used mostly for dry riding may last much longer. The spoke holes are “worked” more, leading to spoke hole cracks.

  With disk brakes and other hub brakes, the rim has no brake track to wear out. The rim lasts longer, and spoke hole cracking is a common failure.

• “More spokes” shares load among more spokes. The main reason spokes fail is they are “worked” on each wheel revolution, and over millions of cycles, microscopic material cracks grow and the spoke fails. Using more spokes reduces the amount each spoke is worked on each wheel revolution. A small reduction in working can give a large improvement in spoke durability. This also helps spoke bed durability.

  “More spokes” also improves the rim’s support, which helps support high loads, so helps the wheel stay true under high load, and helps wheel strength against high loads.

  For wheels with a low spoke count, if one or a few spokes are damaged or come loose, the wheel can go out of true, and it also loses strength. “More spokes” helps keep spokes tight. Even if one or a few spokes is damaged or comes loose, “more spokes” means less loss of wheel trueness and less loss of strength.

  But: more spokes cost and weigh more; more spokes have more air drag; and more spokes means it takes longer to build a wheel.

• “Thinner spokes” stretch more under a given load. When built in to a wheel, that means a given load is supported by more spokes, which then means each spoke is “worked” less on each wheel revolution, which increases spoke durability.

  Spokes mosty fail from fatigue. And they usually fail at the head, elbow, and threads, because there are uneven stresses left from making the spoke. As the spoke is “worked” on each wheel rotation, the high-stress parts accumulate fatigue faster and so fail sooner.

  Thinner spokes are more stretchy, so the tension changes less on each wheel revolution. That reduces working, which reduces fatigue, and so spokes last longer.

  The cheapest approach to a thinner spoke makes everything thinner — head, elbow, shaft, and threads.

  A more durable approach is to thin only the middle. The parts with uneven stress — head, elbow, and threads — are thicker and stronger. However, those parts are worked less than either an all-thick spoke or an all-thin spoke. Because the ends are thick and the shaft is thin, the shaft so does most of the stretching, gets most of the “working”, and takes more of the fatigue. But since the shaft has even stress, it is still durable — spokes like this can last hundreds of thousands of kilometers.

  Spokes which are thick at the ends and thin in the middle are swaged. They are sometimes called “butted”, but the industrial manufacturing process to make the spokes is swaging [https://en.wikipedia.org/wiki/Swaging as of 2025-03]. Whereas “butting” is a way to shape tubes [https://www.reynoldstechnology.biz/faqs-on-reynolds-steel-tubing/what-is-butted-tubing as of 2025-03], but spokes are made of solid wire, not tubes. If you see or hear “butted spoke”, they mean swaged.

  Thin-in-the-middle spokes cost more to make. And, they are somewhat more time-consuming when building a wheel — both for humans and for robots.

• Another source of wheel problems is bearings which are damaged and loose. This can lead to “slop” which causes rub for both rim brakes and disk brakes. If you keep riding, the bearing can jam or fall apart, causing damage which is more expensive to repair.

  The main reason bicycle bearings fail is grit intrusion. Hub bearings are usually “sealed”, but dirty seals wear and eventually let in grit which damages the bearings. Cheap bearing seals can wear faster, and might not seal even when new [http://pardo.net/bike/pic/fail-036/index.html as of 2025-03]. Changing air pressure due to weather and temperature changes can also draw dirt outside the bearing past seals in to the bearing.

  Keeping seals clean is thus an advantage. Many hubs thus have secondary seals or shields. But doing it right is tricky, and many hubs do it wrong. Mis-designed secondary “seals” can let in dirt and then trap it. A good secondary seal can be cheap, but is not free. It usually takes a few years before bearings fail from bad seals, so worse seals are one way makers can cut costs and still get “several years” durability. But worse seals hurt long-term durability.

  Bearings also fail due to overloaded. Larger bearings can help, but add wight and cost. One reason for overloaded bearings is wheel clamping forces cause the axle to compress, which presses the bearing inner races closer together. But the outer races are in the hub shell, which does not change length. This can lead to high side loads that damage the bearings.

  One remedy is a thicker axle which compresses less under clamping loads... but that adds weight. Some hubs use bearings which are fit loose enough to slide under axle compression. However, you want “just enough” looseness to allow shifting, but so much there is rim or brake slop. Loose-fit bearings can also make clicking, creaking, and other sounds. However, accurate looseness is more expensive; and looseness can also somewhat hurt long-term durability.

• For riders using rim brakes, the brake track wears away. The rim eventually needs to be replaced, otherwise the tire will blow off the rim with a loud bang and sometimes with rider injuries. If the front tire suddenly goes flat, you can be thrown from the bike. One rider online reported he was stabbed by a part of the rim which came off.

  One option is a thicker brake track. Suppose a rim needs to be replaced when the brake track is 0.5 mm thick. Consider two new rims, one 0.6 mm thick and one 0.7 mm thick. The one which is 0.7 mm has double the wear (0.7 - 0.5 = 0.2, versus 0.6 - 0.5 = 0.1). But it weighs only about 10% more (0.7 ÷ 0.6 = 1.167 or 17%) because the brake track is only one part of the rim.

  In practice, a new brake track is often much thicker, but comparing two rims which differ only in brake track thickness, and one weighs 500 grams and the other 600 grams. The heavier rim will give about double the brake wear before it needs to be replaced [http://pardo.net/bike/pic/fail-026/000.html as of 2025-03].

• When spokes go slack, the nipple can turn a tiny amount, which loosens the spoke. If a wheel is under-built for its normal loads, this “tiny amount” is repeated enough times for spokes to go loose and for the wheel to come out of true.

  Put another way, if a wheel gradually loses true, that probably means it is under-strength for normal loads.

  Using thinner/lighter spokes spreads loads across more spokes, and so makes it harder for any one spoke to come loose, Thus, the wheel is more likely to stay true. Similarly, a stiffer rim makes it harder for any one spoke to come loose, so the wheel is more likely to stay true.

A durable wheel is somewhat heavier, but not lots. For example, a stiffer rim may add 200 grams. Going from 32 spokes to 48 could add 100 grams of spokes and nipples ... but may be partly offset by using lighter spokes, which could save 1 gram/spoke or 48 grams across all 48 spokes. So 32 heavier spokes to 48 lighter spokes may add 50 grams. A hub with larger bearings, a stiffer axle, and better secondary seals might weigh 50 grams more.

That adds 300 grams per wheel or 600 grams to the bicycle. It also somewhat increases the bike cost, but probably only a few tens of dollars.

For for ordinary riders, the modest increase in weight and cost can lead to many more years without needing service.

More-durable wheels just look like wheels, not “cool” like centerless/hubless wheels. But if problems with ordinary wheels are one reason people get interested in centerless/hubless wheels, then improvements to ordinary wheels is probably a good thing. Indeed, many people have wheel problems after a few years use. So improvements like the above are probably a good thing, even without considering centerless/hubless wheels.

A fair comparison

What is a “fair” comparison of two bikes? Made-up example:

bike with hub-type wheels	$500	12 kgs	1.0
bike with centerless/hubless wheels	$500	18 kgs	18/12 = 1.5

The bike with centerless/hubless wheels weighs a lot more. Somebody suggests a fix is to spend more for lighter materials and fancy manufacturing that saves weight:

Mistake: compare cheap versus expensive
bike with hub-type wheels	$500	12 kgs	1.0
bike with centerless/hubless wheels	$2500	13 kgs	13/12 = 1.08

It saves a lot of weight ... but comparing a $2500 bike to a $500 bike is not really fair.

Next, somebody suggests upgrading the $500 bike with $2000 wheels:

Mistake: spend on “wheels” instead of spend to save weight
$500 bike with $2000 hub-type wheels	$2500	11 kgs	1.0
bike with centerless/hubless wheels	$2500	13 kgs	13/11 = 1.18

But riders care about bikes, not about wheels. A fairer comparison is two bikes that cost the same:

Fair: two bikes that cost the same
$2500 bike with hub-type wheels	$2500	9 kgs	1.0
bike with centerless/hubless wheels	$2500	13 kgs	13/9 = 1.44

This is a made-up example, but I have seen mistakes like the above.

The same kind of reasoning applies for drag, cost, durability, etc.

Unsupported claims

It is common for product literature to make unsupported claims: “We use XYZ tires for better traction” but without providing any supporting evidence that the tires have better traction. And, indeed, better than what?

Sellers make unsupported claims for both good and bad reasons. Reasons can include:

• Space and time are limited, so get the main point across. Supporting info? More details on request, over here, etc.
• This page is a work-in-progress, supporting info will be posted soon.
• It seems right to me; I am lazy and have not actually checked.
• It seems right to me; I am confused, so it has not occured to me to check.
• The less I say, the harder it is for me to get caught.

Unsupported claims are a problem for many reasons.

• Unsupported claims are often right ... and often wrong.

  Often it is hard to tell which claims are good or bad, unless supporting data is (or is not!) provided.

• Even if nobody is trying to deceive you, if you believe an incorrect claim, you can wind up with something expensive which is useless to you; or you could pay and get nothing at all.

  One way to get in trouble is somebody honestly thinks they have a great idea, but have not thought it through. They offer a product along with their claims — which they believe, but have not themselves tried to support. Later, they find there is a serious problem, but by that time buyers have already lost money on it.

  For example, they may honestly believe their rim-on-rollers design is durable. So they never try to test it, and never even try to figure out what kinds of use are expected or what counts as acceptable durability.

  If it never occurs to them somebody will ride it in the rain, they will not even try to figure out if it is durable in the rain, or after being submerged in dirty water. If they believe it is durable, they never even stop to consider how other riders will use it, and what somebody else thinks is “durable”.

• Scammers often use unsupported claims to claim something is good and then deliver something bad... or nothing at all.

  If they publish an actual weight, then somebody reviewing it can measure and see if the actual weight is a lot different. If they make an unsupported claim “lighter than a spoked wheel”, then they can simply claim the spoked wheel is one they picked — which happens to be vastly heavier than most.

  Scammers also use unsupported claims as part of a psychological trick to get people enthusiastic and distract them from problems; or to convince them to put money towards something the scammer never even planned to deliver.

Here are some thoughts on how to approach unsupported claims:

• Example: consider the Reevo claim “The wheels are triple-sealed from the elements for superb long term reliability.”

  Okay that’s nice, Reevo, but... What are the triple seals, and how do they work? Reevo might mean triple shields and no contact seals, in which case rolling through a deep puddle could still get junk inside the bearings. What testing did Reevo do to check whether the seals work as intended? What are their standards for “long term reliability”? Some bike makers seem to think anything longer than the warranty period is “long term”.

  So a first comment is: read the words and see if they lead to specific numbers, such as number of hours service in a rain simulation chamber; and number of hours for hub-type wheel.

  Or, if no numbers, is the information detailed enough that you could in theory make one of your own? If they tell you enough about the seals, you can figure out whether water and junk are likely to get inside the bearings when the wheel rolls through a deep puddle.

• Supporting evidence can help answer a wide range of questions, beyond the specific test and results.

  For example, suppose the seller tests durability and explains how they did the test. Then somebody else can look at the study and figure out how much the tests and results apply to them.

  If, say, all the tests are dry, a potential buyer from someplace wet can figure out that they need more information to find out if the wheel is durable in wet use. The seller does not need to test all conditions, they just need to show what and how they tested.

• If there is not supporting evidence, a good first step may be to assume the worst and see if you care.

  “Our wheel is lighter!” If it turns out to be lighter than one horrible wheel, but much heavier than everything else. Are you still happy? If your main goal is art, then maybe you do not care much about the weight. If your goal is racing, maybe you do, and should ask for the specific weight — or give up on it and go looking for more reputable stuff.

• Thinking about unsupported claims may also lead you to think about things that you care about, but which are not being mentioned.

  For example, the Reevo Indigogo page mentions nothing about rolling drag, noise, or maintenance. If you start thinking about reliability (an unsupported claim), it may lead you to also think about these other things which are not mentioned — but which are probably important for most potential owners.

There is a risk that a scammer provides fake “supporting information“. However, it is often hard to make a realistic fake. And, high-quality fakes are more work, so if a scammer’s goal is to take your money without doing work, they will be reluctant to spend a lot of effort on a high-quality fake.

This section probably seem obvious: “Of course I will ask about any unsupported claims.” But look at comments from people disappointed by the Reevo they got, or by the Cyclotron they paid for and never got. It is clear things could have gone better. Careful thinking about unsupported claims is just one of several things a buyer can (should!) do; but it can help. It seems from comments at least some buyers got tripped by unsupported claims; and had they looked a little harder and thought a little more about those unspported claims, maybe things would have gone better.

Summary

TL;DR:

• The advantage of centerless/hubless wheels is: they look nice.
• Centerless/hubless wheels have more drag, cost more, weigh more, and fail faster.
• Every few years, somebody introduces a new version and uses big words to say they fixed all the problems. They did not.
• This page has some analysis of why it is hard to build good centerless/hubless wheels.

In a bit more detail:

• Centerless/hubless wheels look nice. But for most riders and most riding, they do not solve any problems better than regular hub-type wheels.

• Worse, centerless/hubless wheels add lots of problems: they slow you down, they cost more, and they wear out faster than hub-type wheels.

• And: the problems are hard to solve. And: you need to solve all of them to make a good “use it every day” product: Got rid of excess drag? Great! Still wears out fast and is expensive to maintain? No thanks, I will stick with hub-type wheels.

If something looks good and is more expensive, that is okay: it can justify the added price based on art. Fancy paint costs more without making the bike faster or stronger or more durable. But fancy paint also won’t slow you down.

Whereas centerless/hubless wheels can make the bike slower and weaker and less durable and more expensive. Fixing those problems is hard — which is why we do not use them, even though the idea is popular and keeps coming up.

Perhaps what is most MOBI about centerless/hubless wheels is the same basic idea keeps coming up, and each time many people seem to treat it as “This time for sure!” Without considering maybe there are some Very Good Reasons we are not using them already. Or, maybe worse, somebody proclaims they have solved all the relevant problems, and other folks seem to assume that is true, without really checking. And then it turns out: nope.

The problems might be solvable. Perhaps someday magnetic bearings [“Big Bearing” Alternatives] and and electric drive [Drivetrain types] can reduce the problems to a tolerable level. Or maybe there will be other developments. But for now, design sketches are design sketches — NOT a fast, affordable, and durable product. For now, when you see “We solved it!” you probably want to read that as “We have some ideas!”

Or, maybe, “We have some bad ideas.” Coming soon to a museum near you.