One question: What about angular contact bearings? You see these being used in bottom brackets on occasion, for instance by WheelsMFG. What's the advantage? Are they more durable? Are they easier to preload right? Why do we rarely see them in hubs nowadays when - to my understanding - all the loose ball bearing assemblies that used to be common in hubs are in principle angular contact, right?
This question keeps coming up so I'll pin it. In my opinion we went away from cup and cone bearings in hubs because, and some of these reasons are why we don't use cartridge AC bearings in hubs; 1. Commercial. Only Shimano has/had the ability to in-house produce cup and cones at high enough volume and quantity to make them viable. 2. Weight: cup and cone and angular contact setups are generally heavier. 3. Drag: AC bearings normally have a higher dynamic friction than norma deep groove ball bearings. 4. Packaging: AC bearings are less readily available in as many sizes, so it can be hard to find one to neatly fit into an extremely lightweight/complex hub design. 5. Tolerances. As we know bike industry tolerances are shit. Preload tolerances are crucial in AC bearings as they can only really take axial load properly in one direction. So a pair need to be 'matched' which is highly costly. 6. Radial load. For a given size envelope an AC bearing has a lower radial load rating than a deep groove ball bearing. This is where most of the load is coming from (rider weight, bump input etc). There are more reasons im sure!
@@PeakTorque Excellent explanation. Thank You very much! I'm surprised to read that drag is higher for AC bearings. My completely unscientific subjective impression has always been that nothing spun as freely and easily as a properly set up and preloaded cup and cone hub. In fact I had a front wheel with a Shimano hub that virtually spun forever.
I currently have a Wheels Manufacturing BB with Enduro's angular contact bearings (steel) to see how it's like living with a pair. (this is also my first quality BB in 10 years of riding) Since this BB should be easily rebuildable, I can swap to radial bearings on any future full service.
@@sebastianm2381 The drag on an angular contact bearing "may" be higher when loaded radially, however it's significantly lower when loaded axially. Different horses for different courses. Of course the design of a particular angular contact bearing could influence that as well, since "angular contact" does not imply a "fixed in stone" angle of the contact face.
Much of what the bike industry favors and tends to trend to is cost and ability to get the general consumer hyped up to buy it. For example in 20 years gravel bikes (I'm betting that road bikes too) are going to be largely dual suspension but consumers typically buy designs without suspension. Suspension designs are more complicated to engineer and more costly to manufacture but consumers aren't going to readily spend the extra money to make it worth the extra hassle to make. In fact most likely they would make more money selling a bike without suspension (despite it being easier to design, manufacture and market). It's a similar idea to angular contact cartridge bearings used in hubs and bottom brackets. My old mountain bike has some wtb new paradigm hubs on it and they've been brilliant for the past 25 years. Angular contact is expensive to do half decently but hard to market to consumers. Heat treated super steel is lighter weight than straight gauge titanium. If you tell an average consumer that a frame is ti (especially 20+ years ago) their eyes will/would light up and they'll get excited. Easy sale. Try explaining the differences between different types of cartridge bearings and you're going to see their eyes glaze over in a minute. It will result in the salesperson/bike shop wasting their time and the customer walking out of their shop to buy a bike at the other store down the street - nearly every time. Good radial contact bearings are good such as dt or bitex. Good angular contact is better. That said how much better and how much more expensive to manufacture and for what extra profit?
Man! This should have been a lecture for mechanical engineering design. Really! Why none of my lecturers ever brought a bicycle wheel in the class and just talk about bearings? I would have listened for much longer than 20min. Things make so much more sense now!!
As a non-professional when I hear "pre-load" I know what it means in a general sense, your explanation had made it much clearer. I assume the same principle applies to headsets.
Excellent presentation skills, the illustrations and practical examples were fantastic. More of this please. I think you should pursue more of this type of content. Engineering is clearly your skill and it sets you apart so you should push more of it. Well done.
I finally made use of this video servicing some DT swiss hubs. They preload the 6802 bearings just like your wheels. I have a set of micrometers and did my best to measure the preload and it's about .3 mm. DT swiss advertises that they over preload the bearings with the intention of having some reduced by tightening the skewer. This is absolutely true. The wheels have noticeably high rotational drag when spun by hand and even with a lightly clamped skewer. When the skewer is tightened down with significant force the drag disappears. Most people are probably not tightening their skewers enough with this amount of preload built in however.
@@basengelblik5199 There is nothing to torque as the bearings are press fit. I suppose there could be a torque spec for the skewer but this would never be used in actual practice
I am glad that you had the same thought I had that the axle should be shorter instead of longer in order to make the end caps keep the whole assembly locked in place. Had a bit of a stress moment there, thought I had misunderstood the theory for so many years. Another point is that with a longer axle pushing in the bearings will not work with my toolset. They push on both races so the bearing won't be complelty set and preload is not achieved.
Really brilliant detailed explanation which transcends bike wheel assemblies. If there’s one thing I think it may have been worth adding, it would be the type of load case this bearing arrangement is subjected to, and the effect on bearing fit preferences on the shaft and housing, which of course would change according to the kind of preload design you have too.
Educational and still enjoyable, I really aprieciate that you took time to explain how to read axial clearence chart and show example of application of this knowledge.
Quick comment: Around 3:30 in, you mention that wheel bearings in cars, preload doesn't matter as much as there can be some slop. As an engineer for wheel bearings for cars, preload makes a huge difference in lots of the different attributes of the bearing, from drag, life, and seal integrity/sealability. Most passenger vehicles use a double row ACB and are designed in a way to ensure the proper preload. When a bearing goes out of preload, this can lead to vibrations, poor life, and reduced stiffness.
Yes to be fair you are right, i was getting brake judder on the front right (in the video), tightened the nut 1/10th of a turn and it fixed it. I am always worried about thermal effects so i tend to leave a tiny tiny amount of free play rather than doing it too tight!
@@PeakTorque for this reason, most manufacturers have gone to an orbital roll form, essentially a big rivet, which eliminates any adjustment of the preload. Mercedes is one manufacturer that has stuck with a very old design that still relys on a nut to determine preload. The theory of most manufacturers is, eliminate this variability as who knows how the quality of some mechanics torque wrench (or if they even use one) to determine all the final preload. Overall good video. The wind space axle length bit is slightly funny. Always enjoy your videos. Keep up the good work
@@collinsnyder8682 cheers! If you happen to know what the torque should be on the w204 mercedes front do let me know! I just nipped up finger tight so there was no play when grabbing the top and bottom of the rotor!
Wheel bearings in struts have a circlip to mitigate any drift when loading up the hub. The old tapered roller bearings were preload critical. You could do it to feel but preload dial was advised if you had one. I've done loads of bearings on my old classics by feel.
Nice work. 👍 Good info Ive only ever preloaded cup and cone wheel bearings with spacers taking up the fork gap under a quick release during adjustment. It was part of Paconi blueprinting at the time in the early 90’s. It was amazing how the qr would compress the axle.
as a mech engineer myself l have to agree very hard. the stiffness differrence between X and O bearings may not be that much, but the discbrake heat aspect could be quite important. you can maybe check that with unequal bearing wear drive to nondrive side
As long as i cant feel noticeable drag I'm pretty happy. Some wheel definitely have too much preload and its either the preload tube is out or the bearing housing are out. You can shim a seat to correct the bad machining in some cases
just a quick thought of why it's face to face is that the dust cap rotates with the spindle. it should't contact the outter race . plus then it would become a possitional preload system which could lead to exessive preload and wear on the bearings?
I wouldn't wanna correct an engineer but I think they did it right when the wheels are tighten to the frame . When the end caps pushes in the inner race toward the axle. You would get the back to back loading you want. And there would be no load path to push the bearings out once installed onto the bike with the thru axle tightened. Left outer race lateral outward force would be stopped by the right outter race towards the hub body when axle wanna shift visa versa both side.
This is an awesome demonstration and answered a lot of questions as I was putting new bearings in my dt240 hubs. I do have one question in relation to preload and the length of the axle vs the hub shell. Now I didn't do any extensive measurement so take this with a grain of salt...just my observations. It seems that the front hub on my wheelset is set up similarly to the Winspace wheels you were demonstrating where the hub shell is shorter than the axle. I noticed this due to the fact that when the bearings were fully seated in the shell the bearings were fairly stiff and didn't roll very smoothly. They felt a bit bound or overly preloaded and it made me worry about bearing life. My moderately scientific method to test how freely they would spin was simply to put the valve stem at 90 deg relative to vertical and see if it would fall to the bottom. When the bearings were seated in the hub shell the valve stem would not move and the wheel remained stationary. However when I put the wheel in the fork and loaded up the inner races by tightening the skewer the wheel spun much more freely based upon my method above. The valve stem rotated downward solely based on gravity. So after all that here is my question. Could the reason for the axle being longer than the hub shell be due to a small displacement that occurs in the flanges of the axle when under load from the skewer? If there is enough displacement (or in other words a slight shortening of the axle between the flanges) due to load could it potentially bring the bearing into proper alignment? Anyway...just a question from my observations. Great content and thanks for the video.
This very interesting, thanks. We have lots of bearing setups that we rebuild on our rotary motion industrial machines. Learning as we go, there are lots of questions and mysteries we have about these topics. What is the design intent, where are angular contact bearings used, ahouls they be back to back or face to face, what clearance ratings mean and which to select, how important this might be. Why a bearing might be press fit vs sliding fit, why apparently identical assemblies run at different temperatures, what size race spacers might be and are they worn, when to use wavy springs and shims, circlips. Its interesting to be able to imagine how a few microns might affect our results in practice
Very interesting topic. One thing that you didn’t discuss is the method of assembly. There are various tools (read sockets) out there which variously apply pressure on both of the races simultaneously, or just the outer etc. In my testing, I’ve found that the assembly method can have just as much if not more of an effect on the final preload of the bearings as the tolerances in the bearing and the shaft lengths.
Do you torque your pressfit bearing? Or am I misunderstanding. Because if the axle is shorter you can hardly over torque the pressfit bearing during assembly. A stip is a stop right? Honoust question.
Would you agree that a easy way to check for back to back tolerance is if the wheel spins freely when the axle/QR is tensioned but then feels a bit draggy if the end caps are held in your hands and the wheel spun? Also your hub tolerance must be deliberate as even Chinese CNC turning is more than capable of single micron tolerance, even with anodising allowances.
Negatory on the GXP crank preload. With the Sram crank, only the NDS bearing supports the crank axially, the DS bearing is there only for radial support. This is achieved by the spindle having a step down from 25 to 24 mm (I think...) in the diameter before going through the NDS bearing where it is then captured by the NDS crank being tightened down on it. This works on threaded BBs, but with PF BBs, the DS cup could get loose. THAT is why the PF GXP interface uses a wavy washer. Sure, it preloads the bearings, but only in the PF varant, not in the threaded variant.
Yes only the nds takes axial loads on gxp agreed, the other bearing is floating and only does radial loads as you say. However with both PF and Threaded gxp, the wave washer placed on the DS does indeed preload the nds bearing as it pushes the crank assembly away from the bike, 'pulling' on the spindle which is locked to the NDS inner race.
@@PeakTorque that's the catch, the threaded variant of the bearing, in my experience, does NOT come with a wavy washer. EDIT: www.sram.com/globalassets/document-hierarchy/user-manuals/sram-mtb/drivetrain/mtb-cranksets-and-bottom-brackets-eeu.pdf Page 11 and 15.
@@PeakTorque If I remember correctly (transfering a crank from a PF BB bike originally to a BSA BB bike) there is actually a gap on the DS between the crank and the bearing. Imagine my surprise at not getting a wavy washer and me not being able to find the old one, not having a spare, etc., only to look up the manual and see it's not needed. I'm also fairly certain I saw somewhere written that it's there to prevent the PF cup from wiggling out from the frame, like I mentioned. Which I don't see as impossible given the design and given how BB30 and DUB from Sram (and Shimano stuff) works preload wise.
I made the same observation when I changed the bearings of my Hope Pro 5 rear. The bearings preloaded substantially in the same manner, showing that the axle was shorter than the hub shoulders -- which I found weird. This preload should get reduced when the wheel is installed though (since the axle gets compresses a tiny bit). Next bearing change I plan to play with 0.1mm calibrated washers to try and set the contact angle the other way around. I'm a heavy rider so I won't be waiting for long:)
Hi, I'm an engineering student and I'm trying to understand how to calculate my bearing resistance to axial loads. I have a double row ball bearing with a radial load resistance of 4000N, my radial load is 2700N and my axial load is 800N. What should I look for in a datasheet to be sure the bearing can resist my axial load? What calculation should I do? I cannot find any specific explanation to get a value of max axial load for angular bearings, if you could help me out that would be amazing, thanks
You can use this tool to calculate the bearing life www.skfbearingselect.com/#/bearing-selection-start The equations behind that tool are on the skf site.
Great video! What's the optimal preload for bearing life then? When setting up Shimano hubs the usual advice is that a tiny bit of slop is better than too tight as the QR will squash it up a bit. But idk if this is just bike shop myth and lore.
Not a myth. Park Tool also recommend this. I also do it, and you can actually feel how tightening the QR skewer makes the slop in the bearings go away. Some newer Shimano hubs use an aluminum axle, so it will behave a bit different, but still should squish slightly under the QR load. I have one of those newer hubs with aluminum axle as well, but haven't tried it out yet.
Thanks, interesting to watch, watched all through, wasn't that nerdy at all, I will eventually measure my hubs out of curiosity, please make more videos like this.
One explanation for the weird paaring of Hub Shell and axle could be the additional tension from e.g. a classical quick Release ? With this design the tension of the quick release does not Impact the wheel bearings at all.
I would like to see a demonstration that a difference in lateral wheel stiffness between two different bearing preload arrangements can be meaningfully measured. I suspect that it cannot, due to all the other components that make up a wheel (spokes, rim, nipples) being in the load path.
Doesn't the internal axle get compress? more so if it a disc brakes wheel; wouldn't that come into play for the angular effect you described? Great video!
Where did you get that chart to convert radial clearance to axial clearance from please? I've been designing gearboxes for 13 years and never seen that before
Doesn't the axle get compressed by the through axle, shortening a bit and unloading the bearings? I noticed this on my DT350 hub, with the caps off the bearings are stiff and notchy, but when the wheel is clamped on, it's very smooth. In my case the inner axle sits between the inner races (pushing them out) and the end caps press onto the inner races when the wheel is clamped by the fork.
To improve that moment arm from the wider stance of outer race, you would need to redesign the the hub's preload system to provide preload externally rather than the current internally provided preload. By reducing the length of the axle collars, the bearings would no longer be preloaded? There would then only be two ways to apply that preload, either through the end caps, or via a threaded collar. Since preload provided through the end caps would vary depending upon axle torque, on top of that being too much preload, that would then require a redesign of the preload assembly. Please let me know if I have misunderstood, just trying to learn (Mech ENG student/bike mechanic) :)
Actually to correct myself, I suppose the thru axle would provide the preload, and then the hub axle collars would prevent any excessive preload. Would feel a bit weird off the bike but if it works who cares. I guess the only issue that you may encounter would be the large amount of force exerted on those hub axle preload collars, which if aluminium, may not be strong enough.
@@jdodd1258 Yep you got there in the end! When the axle is shorter, the end caps provide the preload when clamped and lock everything to the inner races. Preload does not depend on the thru axle load, but the thru as axle is needed to set it. If that makes sense!
question, are stiffer wheels always faster? I'm not an engineer, but I could see a slighty flexier wheel absorbing road imperfections better, and thus you as the rider can put down more power cause you're more comfortable. And thus this increased power might overcome the efficiency loss of the flexier wheel
The stiffness of the wheels concerning the bearings is axial stiffness, you wouldnt want to absorb bumps in the road by your wheel going from side to side. And its very hard to engineer vertical compliance and torsional and axial stiffness. Also its debated if vertical compliance in a wheel is actually something you want since your tires can deflect much much more than the wheel itself and do so more efficiently. TL;DR: stiff wheel good. Adjust tire size and pressure to road condition.
Theoretically, yes, stiffer wheels can make you faster if your body can endure it. We are human with different body strength, ability, mental strength, skills and size. If we put a robot on a bike, stiffer bike/wheels will accelerate first and save some watts. If your objective to get more comfort, just don't buy wheels with carbon spokes.
I rememer bearing manufacturers recommending a loose fit h6 on the inner ring when the inner ring is standig still an a tight fit on the moving outer. maybe thats what winspace is following.
At LAST an intelligent exploration of the role of preload. For best load distribution across contact areas. Stops slap damage. Slowly polishes itself to a better fit. Too much? Damage. None? Short life & off center damage.
Where do you get those bearing clearance charts? Any links available? Thanks. As for the preload direction, who in the hell would make the inner tube longer?!?! Of course you make it shorter for exactly the reason you mention - the endcaps ensuring the preload once the wheel is mounted to the bike. The way this industry works there is no way in hell the bearing inserted into the hub shell will stay there, unless it's not supposed to stay there and there way a mistake made (too tight of a tolerance in production or something similar).
Thank for this, super! Question out of the blue, from a non-mechanical engineer. The design choice for ball groove bearings in wheels vs cone shaped ones, why? I mean, in a headset, usually cone shaped ones are used (load angle differ, of course...) But is the groove ball bearing "always" the best option, given what types are available?
So if the bearings in wheel hubs are effectively being used as angular contact bearings, is there a good reason not to just use angular contact bearings in the first place?
The contact angle is more extreme with a proper AC bearing thus the drag will be higher. Radial load rating (rider weight, bumps) will also be reduced for a given size.
Interesting, didn't get lost. I am concerned that when I next service the land bearings life will get more miserable 😁 At the recent Europeans I was discussing bearing choice, and the suggestion was ceramics were best. I would like to point out not my tests and I and sticking with my skf and fag steel bearings I will be double checking in inner spacers.
If the correct preload is so important why do Shimano have a plastic knob on their BBS that you tighten by hand, rather than a precise torque setting like most of the other bolts on a modern bike?
I could be wrong here so go gently, but I vaguely recall reading in a Shimano document that the torque specification for the NDS crank/axle retainer was 1.5 NM.
I believe you missed the portion where PT discusses the Hollowtech II crank adjustment as a "position preload" system. The downside of position-based preload is you can never really measure how much preload force you actually applied, unlike with a wave washer where you can measure force via its deflection. So Shimano tries to make up for this by providing a torque spec for the preload adjustment bolt - a very low torque, such as 1.5 Nm
@@TypeVertigo Correct. In a shimano BB, Hambini may verify, but i think they actually use proper angular contact bearings which dont mind quite of a lot of axial preload, but yes setting it by hand is rather arbitrary.
Interesting. So when you say “Shimano BB”, would this be a Shimano branded BB or any Shimano type BB? ie: Would say, a Praxis T47 BB also use AC bearings?
@@BikePappy Normal Praxis stuff is not angular contact. If it is they specifically advertise as such, eg "Ceramic Version:Enduro 2437 Angular contact, Grade5 balls, ABEC5 Assembly". Cost is $240 vs 90 for the regular version.
Buy a new hub axle, turn down the shoulders 0.2mm on each side and then back to back test it! I'm sure the difference could be measured but I'd love to see if the difference your describing can be perceived. You could even have someone else assemble each setup without you knowing which axle was in it to eliminate any potential bias.
fyi there is no need for another axle: set of standard shim rings will do. They come in thicknesses from 0,05. Making the assemly as a blind test is smart though...
there is something wrong with calculation you said i get something in the middle but it depends on the fitting not only chart.before you have to calculate fitting and you have to measure shaft and inner ring than you have to calculate how much clearance left for example inner ring in 19.97 mm and shaft is 20.00 that means you have 3 microns fitting than you have to subtract from maximum celarance than you find the correct celarance.chart is just for in run clearance not depends on stok tolerances
Never knew such simple topic would take 20 minutes... But boy how good has your presentation become! If you had to develop these skills at work, I'd consider firing your colleagues Also we call these O and X arrangements (put Stroheim meme here)
Thanks for this video. Very interesting. I have a set of DT Swiss disc wheels with 240S hubs. The original front hub bearings felt rough off the bike, so I pushed in a new set of NTN/NSK bearings (I forget which) and to my dismay they also felt rough. I started to suspect that the shoulders on the thru axle had been machined too wide for the hub, creating excess axial load. I then found a maintenance guide on the DT Swiss site that advises beginning by pushing the non-drive side bearing fully home against the shoulder, before inserting the axle and then finally pushing the drive side bearing only as far as needed to eliminate bearing play (i.e. NOT fully home against the hub shell shoulder). DT Swiss claims that the bearings will always feel a bit rough when off the bike and will only run smoothly once the wheel is fitted to the bike and the axle compressed by the end caps. Does this smell fishy to you? Obviously, I have no way of testing how smooth/freely the bearings are running once the wheels have been fitted to the bike.
Dare I say it, that's not just fishy, but double bullocks at least: - DT are known to mess up radial tolerances, thus axial fiddling won't help - when the bearings are overly preloaded in the manner the manual advises, extra pressure when the wheel is secured will make things even worse
@@feedbackzaloop Thanks for your thoughts. I considered getting some shims to sit behind the outer bearing races, to pad them out laterally. For now, I have pushed one bearing fully home - the other, I pushed in bit-by-bit, testing the axle for any play and pushing a little bit further until no play remained. The bearings now feel OK outside of the bike. God knows what they are doing when fitted into the bike and the thru axle tightened 🤷♂️
@@trevekneebone369 It certainly does not sound like good practice. The final bearing position should be up against the shoulder, so it sounds like they messed up!
@@Hambini Thanks. So what is the fix in my case? Machine off a slither from the axle shoulders? Place a shim between the hub shell shoulder and the bearing outer race? Thanks!
One question: What about angular contact bearings? You see these being used in bottom brackets on occasion, for instance by WheelsMFG. What's the advantage? Are they more durable? Are they easier to preload right? Why do we rarely see them in hubs nowadays when - to my understanding - all the loose ball bearing assemblies that used to be common in hubs are in principle angular contact, right?
This question keeps coming up so I'll pin it. In my opinion we went away from cup and cone bearings in hubs because, and some of these reasons are why we don't use cartridge AC bearings in hubs;
1. Commercial. Only Shimano has/had the ability to in-house produce cup and cones at high enough volume and quantity to make them viable.
2. Weight: cup and cone and angular contact setups are generally heavier.
3. Drag: AC bearings normally have a higher dynamic friction than norma deep groove ball bearings.
4. Packaging: AC bearings are less readily available in as many sizes, so it can be hard to find one to neatly fit into an extremely lightweight/complex hub design.
5. Tolerances. As we know bike industry tolerances are shit. Preload tolerances are crucial in AC bearings as they can only really take axial load properly in one direction. So a pair need to be 'matched' which is highly costly.
6. Radial load. For a given size envelope an AC bearing has a lower radial load rating than a deep groove ball bearing. This is where most of the load is coming from (rider weight, bump input etc).
There are more reasons im sure!
@@PeakTorque Excellent explanation. Thank You very much! I'm surprised to read that drag is higher for AC bearings. My completely unscientific subjective impression has always been that nothing spun as freely and easily as a properly set up and preloaded cup and cone hub. In fact I had a front wheel with a Shimano hub that virtually spun forever.
I currently have a Wheels Manufacturing BB with Enduro's angular contact bearings (steel) to see how it's like living with a pair. (this is also my first quality BB in 10 years of riding) Since this BB should be easily rebuildable, I can swap to radial bearings on any future full service.
@@sebastianm2381 The drag on an angular contact bearing "may" be higher when loaded radially, however it's significantly lower when loaded axially. Different horses for different courses. Of course the design of a particular angular contact bearing could influence that as well, since "angular contact" does not imply a "fixed in stone" angle of the contact face.
Much of what the bike industry favors and tends to trend to is cost and ability to get the general consumer hyped up to buy it. For example in 20 years gravel bikes (I'm betting that road bikes too) are going to be largely dual suspension but consumers typically buy designs without suspension. Suspension designs are more complicated to engineer and more costly to manufacture but consumers aren't going to readily spend the extra money to make it worth the extra hassle to make. In fact most likely they would make more money selling a bike without suspension (despite it being easier to design, manufacture and market). It's a similar idea to angular contact cartridge bearings used in hubs and bottom brackets. My old mountain bike has some wtb new paradigm hubs on it and they've been brilliant for the past 25 years. Angular contact is expensive to do half decently but hard to market to consumers. Heat treated super steel is lighter weight than straight gauge titanium. If you tell an average consumer that a frame is ti (especially 20+ years ago) their eyes will/would light up and they'll get excited. Easy sale. Try explaining the differences between different types of cartridge bearings and you're going to see their eyes glaze over in a minute. It will result in the salesperson/bike shop wasting their time and the customer walking out of their shop to buy a bike at the other store down the street - nearly every time. Good radial contact bearings are good such as dt or bitex. Good angular contact is better. That said how much better and how much more expensive to manufacture and for what extra profit?
Well now my voice has broken, I must have hit puberty.
That opening soundtrack is CLEVER GREAT! for a moment I thought I misclicked to hambini video
Hambini PTSD kicking in
I freakin love it!
Man! This should have been a lecture for mechanical engineering design. Really! Why none of my lecturers ever brought a bicycle wheel in the class and just talk about bearings? I would have listened for much longer than 20min. Things make so much more sense now!!
Cheers
As a non-professional when I hear "pre-load" I know what it means in a general sense, your explanation had made it much clearer. I assume the same principle applies to headsets.
Excellent presentation skills, the illustrations and practical examples were fantastic. More of this please. I think you should pursue more of this type of content. Engineering is clearly your skill and it sets you apart so you should push more of it. Well done.
Finally a Hambini Pilgrim collab.
That hambini voice opening though
Very nice. Great that the pen is working.
5 seconds in the video and you already have my like earned. Congrats.
Incredible content! Thanks for trying to bring the rest of us along!
I finally made use of this video servicing some DT swiss hubs. They preload the 6802 bearings just like your wheels. I have a set of micrometers and did my best to measure the preload and it's about .3 mm. DT swiss advertises that they over preload the bearings with the intention of having some reduced by tightening the skewer. This is absolutely true. The wheels have noticeably high rotational drag when spun by hand and even with a lightly clamped skewer. When the skewer is tightened down with significant force the drag disappears.
Most people are probably not tightening their skewers enough with this amount of preload built in however.
Does DT Swiss provide torque settings for the axle assembly?
@@basengelblik5199 There is nothing to torque as the bearings are press fit. I suppose there could be a torque spec for the skewer but this would never be used in actual practice
All I can say is AMAZING!!! This was a master class in clarity without simplification or dilution 😀
I am glad that you had the same thought I had that the axle should be shorter instead of longer in order to make the end caps keep the whole assembly locked in place.
Had a bit of a stress moment there, thought I had misunderstood the theory for so many years.
Another point is that with a longer axle pushing in the bearings will not work with my toolset. They push on both races so the bearing won't be complelty set and preload is not achieved.
Really brilliant detailed explanation which transcends bike wheel assemblies. If there’s one thing I think it may have been worth adding, it would be the type of load case this bearing arrangement is subjected to, and the effect on bearing fit preferences on the shaft and housing, which of course would change according to the kind of preload design you have too.
Educational and still enjoyable, I really aprieciate that you took time to explain how to read axial clearence chart and show example of application of this knowledge.
Wonderfully deep machine engineering talk but interesting. Felt like home ❤️
Quick comment: Around 3:30 in, you mention that wheel bearings in cars, preload doesn't matter as much as there can be some slop. As an engineer for wheel bearings for cars, preload makes a huge difference in lots of the different attributes of the bearing, from drag, life, and seal integrity/sealability. Most passenger vehicles use a double row ACB and are designed in a way to ensure the proper preload. When a bearing goes out of preload, this can lead to vibrations, poor life, and reduced stiffness.
Yes to be fair you are right, i was getting brake judder on the front right (in the video), tightened the nut 1/10th of a turn and it fixed it. I am always worried about thermal effects so i tend to leave a tiny tiny amount of free play rather than doing it too tight!
@@PeakTorque for this reason, most manufacturers have gone to an orbital roll form, essentially a big rivet, which eliminates any adjustment of the preload. Mercedes is one manufacturer that has stuck with a very old design that still relys on a nut to determine preload. The theory of most manufacturers is, eliminate this variability as who knows how the quality of some mechanics torque wrench (or if they even use one) to determine all the final preload.
Overall good video. The wind space axle length bit is slightly funny. Always enjoy your videos. Keep up the good work
@@collinsnyder8682 cheers! If you happen to know what the torque should be on the w204 mercedes front do let me know! I just nipped up finger tight so there was no play when grabbing the top and bottom of the rotor!
Wheel bearings in struts have a circlip to mitigate any drift when loading up the hub. The old tapered roller bearings were preload critical. You could do it to feel but preload dial was advised if you had one. I've done loads of bearings on my old classics by feel.
Nice work. 👍 Good info
Ive only ever preloaded cup and cone wheel bearings with spacers taking up the fork gap under a quick release during adjustment. It was part of Paconi blueprinting at the time in the early 90’s. It was amazing how the qr would compress the axle.
Magnificent, can you please provide the link to the graph to calculate axial clearance
hey, great video!
where did you find that radial clearance X contact angle chart? [10:54]
thanks
Doesn't the 0.2 value apply to a single bearing, so with two in the assembly the target value should be 0.4 rather than the 0.19 you measured?
as a mech engineer myself l have to agree very hard. the stiffness differrence between X and O bearings may not be that much, but the discbrake heat aspect could be quite important. you can maybe check that with unequal bearing wear drive to nondrive side
For me it's always the disc side bearing which fails first in the front hub and the drive side bearing on the rear.
As long as i cant feel noticeable drag I'm pretty happy. Some wheel definitely have too much preload and its either the preload tube is out or the bearing housing are out. You can shim a seat to correct the bad machining in some cases
Peak Torque = my state when you upload a video.
just a quick thought of why it's face to face is that the dust cap rotates with the spindle. it should't contact the outter race . plus then it would become a possitional preload system which could lead to exessive preload and wear on the bearings?
I wouldn't wanna correct an engineer but I think they did it right when the wheels are tighten to the frame . When the end caps pushes in the inner race toward the axle. You would get the back to back loading you want. And there would be no load path to push the bearings out once installed onto the bike with the thru axle tightened. Left outer race lateral outward force would be stopped by the right outter race towards the hub body when axle wanna shift visa versa both side.
This is an awesome demonstration and answered a lot of questions as I was putting new bearings in my dt240 hubs. I do have one question in relation to preload and the length of the axle vs the hub shell. Now I didn't do any extensive measurement so take this with a grain of salt...just my observations. It seems that the front hub on my wheelset is set up similarly to the Winspace wheels you were demonstrating where the hub shell is shorter than the axle. I noticed this due to the fact that when the bearings were fully seated in the shell the bearings were fairly stiff and didn't roll very smoothly. They felt a bit bound or overly preloaded and it made me worry about bearing life.
My moderately scientific method to test how freely they would spin was simply to put the valve stem at 90 deg relative to vertical and see if it would fall to the bottom. When the bearings were seated in the hub shell the valve stem would not move and the wheel remained stationary. However when I put the wheel in the fork and loaded up the inner races by tightening the skewer the wheel spun much more freely based upon my method above. The valve stem rotated downward solely based on gravity.
So after all that here is my question. Could the reason for the axle being longer than the hub shell be due to a small displacement that occurs in the flanges of the axle when under load from the skewer? If there is enough displacement (or in other words a slight shortening of the axle between the flanges) due to load could it potentially bring the bearing into proper alignment?
Anyway...just a question from my observations. Great content and thanks for the video.
Hey PT, at 15:43, what part ensures that the bearings are preloaded and how is it designed? Is the axial force only applied on the inner bearing ring?
Great vid man!
Wouldn't the spoke tension change the interference between the bearing and the hub and thus changing the preload setting?
Cheers!
Love the technical stuff here. Can you explain why the bearings are different sizes from left to right?
This very interesting, thanks. We have lots of bearing setups that we rebuild on our rotary motion industrial machines. Learning as we go, there are lots of questions and mysteries we have about these topics. What is the design intent, where are angular contact bearings used, ahouls they be back to back or face to face, what clearance ratings mean and which to select, how important this might be. Why a bearing might be press fit vs sliding fit, why apparently identical assemblies run at different temperatures, what size race spacers might be and are they worn, when to use wavy springs and shims, circlips. Its interesting to be able to imagine how a few microns might affect our results in practice
Very interesting topic. One thing that you didn’t discuss is the method of assembly. There are various tools (read sockets) out there which variously apply pressure on both of the races simultaneously, or just the outer etc. In my testing, I’ve found that the assembly method can have just as much if not more of an effect on the final preload of the bearings as the tolerances in the bearing and the shaft lengths.
Do you torque your pressfit bearing? Or am I misunderstanding. Because if the axle is shorter you can hardly over torque the pressfit bearing during assembly. A stip is a stop right? Honoust question.
Would you agree that a easy way to check for back to back tolerance is if the wheel spins freely when the axle/QR is tensioned but then feels a bit draggy if the end caps are held in your hands and the wheel spun? Also your hub tolerance must be deliberate as even Chinese CNC turning is more than capable of single micron tolerance, even with anodising allowances.
haha, that hambini teaser intro! "HELLO" love it
Great video! By the way where can i find the graph you are using to find the axial clearance? Thanks!
Negatory on the GXP crank preload. With the Sram crank, only the NDS bearing supports the crank axially, the DS bearing is there only for radial support. This is achieved by the spindle having a step down from 25 to 24 mm (I think...) in the diameter before going through the NDS bearing where it is then captured by the NDS crank being tightened down on it.
This works on threaded BBs, but with PF BBs, the DS cup could get loose.
THAT is why the PF GXP interface uses a wavy washer. Sure, it preloads the bearings, but only in the PF varant, not in the threaded variant.
Yes only the nds takes axial loads on gxp agreed, the other bearing is floating and only does radial loads as you say. However with both PF and Threaded gxp, the wave washer placed on the DS does indeed preload the nds bearing as it pushes the crank assembly away from the bike, 'pulling' on the spindle which is locked to the NDS inner race.
@@PeakTorque that's the catch, the threaded variant of the bearing, in my experience, does NOT come with a wavy washer.
EDIT: www.sram.com/globalassets/document-hierarchy/user-manuals/sram-mtb/drivetrain/mtb-cranksets-and-bottom-brackets-eeu.pdf
Page 11 and 15.
@@Primoz.r That is very odd. Maybe the threaded version has a zero clearance bearing for NDS. Who knows.
@@PeakTorque If I remember correctly (transfering a crank from a PF BB bike originally to a BSA BB bike) there is actually a gap on the DS between the crank and the bearing. Imagine my surprise at not getting a wavy washer and me not being able to find the old one, not having a spare, etc., only to look up the manual and see it's not needed.
I'm also fairly certain I saw somewhere written that it's there to prevent the PF cup from wiggling out from the frame, like I mentioned. Which I don't see as impossible given the design and given how BB30 and DUB from Sram (and Shimano stuff) works preload wise.
I made the same observation when I changed the bearings of my Hope Pro 5 rear. The bearings preloaded substantially in the same manner, showing that the axle was shorter than the hub shoulders -- which I found weird. This preload should get reduced when the wheel is installed though (since the axle gets compresses a tiny bit). Next bearing change I plan to play with 0.1mm calibrated washers to try and set the contact angle the other way around. I'm a heavy rider so I won't be waiting for long:)
credit to dark install for a dark hambini ua-cam.com/video/5mPyMYMNh2s/v-deo.html
An excellent video, even by your high standards. Fair play
You've got me wanting to rip my hubs apart and dig around too
Hi, I'm an engineering student and I'm trying to understand how to calculate my bearing resistance to axial loads. I have a double row ball bearing with a radial load resistance of 4000N, my radial load is 2700N and my axial load is 800N. What should I look for in a datasheet to be sure the bearing can resist my axial load? What calculation should I do? I cannot find any specific explanation to get a value of max axial load for angular bearings, if you could help me out that would be amazing, thanks
You can use this tool to calculate the bearing life www.skfbearingselect.com/#/bearing-selection-start
The equations behind that tool are on the skf site.
that was super clear and interesting, thank you!
Great video! What's the optimal preload for bearing life then? When setting up Shimano hubs the usual advice is that a tiny bit of slop is better than too tight as the QR will squash it up a bit. But idk if this is just bike shop myth and lore.
Not a myth. Park Tool also recommend this. I also do it, and you can actually feel how tightening the QR skewer makes the slop in the bearings go away. Some newer Shimano hubs use an aluminum axle, so it will behave a bit different, but still should squish slightly under the QR load. I have one of those newer hubs with aluminum axle as well, but haven't tried it out yet.
superb intro! superb content! superb video!
Thanks, interesting to watch, watched all through, wasn't that nerdy at all, I will eventually measure my hubs out of curiosity, please make more videos like this.
One explanation for the weird paaring of Hub Shell and axle could be the additional tension from e.g. a classical quick Release ? With this design the tension of the quick release does not Impact the wheel bearings at all.
I would like to see a demonstration that a difference in lateral wheel stiffness between two different bearing preload arrangements can be meaningfully measured. I suspect that it cannot, due to all the other components that make up a wheel (spokes, rim, nipples) being in the load path.
It can easily me measured by not doing up the thru axle tight (so no preload)
Doesn't the internal axle get compress? more so if it a disc brakes wheel; wouldn't that come into play for the angular effect you described?
Great video!
Where did you get that chart to convert radial clearance to axial clearance from please? I've been designing gearboxes for 13 years and never seen that before
Send me an email and il send you the graphs.
I am just going to pop over to Only fans for some bearing insertion 🤣
Where can I hear the full Hambini/Zach Hill project please
Excellent vid!
Doesn't the axle get compressed by the through axle, shortening a bit and unloading the bearings? I noticed this on my DT350 hub, with the caps off the bearings are stiff and notchy, but when the wheel is clamped on, it's very smooth. In my case the inner axle sits between the inner races (pushing them out) and the end caps press onto the inner races when the wheel is clamped by the fork.
There is another video from him on that topic and he takes a dive in to the DT350 special case
When are we gonna see the next episode of PEAK TORQUE HAMBINI AND CHINA CYCLING LIVE?
next week!
Where did you find the chart if you don't mind me asking
reamaction video from hambini coming?
To improve that moment arm from the wider stance of outer race, you would need to redesign the the hub's preload system to provide preload externally rather than the current internally provided preload. By reducing the length of the axle collars, the bearings would no longer be preloaded? There would then only be two ways to apply that preload, either through the end caps, or via a threaded collar. Since preload provided through the end caps would vary depending upon axle torque, on top of that being too much preload, that would then require a redesign of the preload assembly. Please let me know if I have misunderstood, just trying to learn (Mech ENG student/bike mechanic) :)
Actually to correct myself, I suppose the thru axle would provide the preload, and then the hub axle collars would prevent any excessive preload. Would feel a bit weird off the bike but if it works who cares. I guess the only issue that you may encounter would be the large amount of force exerted on those hub axle preload collars, which if aluminium, may not be strong enough.
@@jdodd1258 Yep you got there in the end! When the axle is shorter, the end caps provide the preload when clamped and lock everything to the inner races. Preload does not depend on the thru axle load, but the thru as axle is needed to set it. If that makes sense!
@@PeakTorque Thanks for the clarification!
question, are stiffer wheels always faster? I'm not an engineer, but I could see a slighty flexier wheel absorbing road imperfections better, and thus you as the rider can put down more power cause you're more comfortable. And thus this increased power might overcome the efficiency loss of the flexier wheel
The stiffness of the wheels concerning the bearings is axial stiffness, you wouldnt want to absorb bumps in the road by your wheel going from side to side. And its very hard to engineer vertical compliance and torsional and axial stiffness. Also its debated if vertical compliance in a wheel is actually something you want since your tires can deflect much much more than the wheel itself and do so more efficiently.
TL;DR: stiff wheel good. Adjust tire size and pressure to road condition.
Theoretically, yes, stiffer wheels can make you faster if your body can endure it. We are human with different body strength, ability, mental strength, skills and size.
If we put a robot on a bike, stiffer bike/wheels will accelerate first and save some watts.
If your objective to get more comfort, just don't buy wheels with carbon spokes.
I rememer bearing manufacturers recommending a loose fit h6 on the inner ring when the inner ring is standig still an a tight fit on the moving outer.
maybe thats what winspace is following.
At LAST an intelligent exploration of the role of preload. For best load distribution across contact areas. Stops slap damage. Slowly polishes itself to a better fit. Too much? Damage. None? Short life & off center damage.
Epic intro!
have you pay any royalty to Hambini?
Could you please review Vel 50mm RL from sigma sport? I got them really cheap at 330pounds, but are they any good? Thanks.
Sure, send them in!
Needed a second for the Sam reference. Please do your Moni-o'clock montage too.
Im only into monnies because of sam, kind of addicted now!
Would bei interesting what DTswiss has here ...since they are building Hubs for decades
Where do you get those bearing clearance charts? Any links available? Thanks.
As for the preload direction, who in the hell would make the inner tube longer?!?! Of course you make it shorter for exactly the reason you mention - the endcaps ensuring the preload once the wheel is mounted to the bike. The way this industry works there is no way in hell the bearing inserted into the hub shell will stay there, unless it's not supposed to stay there and there way a mistake made (too tight of a tolerance in production or something similar).
Thank for this, super! Question out of the blue, from a non-mechanical engineer. The design choice for ball groove bearings in wheels vs cone shaped ones, why? I mean, in a headset, usually cone shaped ones are used (load angle differ, of course...) But is the groove ball bearing "always" the best option, given what types are available?
See pinned comment
So if the bearings in wheel hubs are effectively being used as angular contact bearings, is there a good reason not to just use angular contact bearings in the first place?
The contact angle is more extreme with a proper AC bearing thus the drag will be higher. Radial load rating (rider weight, bumps) will also be reduced for a given size.
the pen is working correctly only with the right preload some would say 😂
lol, peak torque, hambini and sam pilgrim in the first ten seconds, i like that
THE PEN IS WORKING!!! 🙌🙌🙌
Preload goes through the roof when locknuts slip on a track hub tightening axle nuts! Scorched toast!
More of it, please :-)
Interesting, didn't get lost. I am concerned that when I next service the land bearings life will get more miserable 😁
At the recent Europeans I was discussing bearing choice, and the suggestion was ceramics were best. I would like to point out not my tests and I and sticking with my skf and fag steel bearings I will be double checking in inner spacers.
Not preload but minimum play.
Only big steering real bearings can be set with a little preload.
since u can't adjust preload on a thru axle I think qr is better
Finally... preload
If the correct preload is so important why do Shimano have a plastic knob on their BBS that you tighten by hand, rather than a precise torque setting like most of the other bolts on a modern bike?
I could be wrong here so go gently, but I vaguely recall reading in a Shimano document that the torque specification for the NDS crank/axle retainer was 1.5 NM.
I believe you missed the portion where PT discusses the Hollowtech II crank adjustment as a "position preload" system. The downside of position-based preload is you can never really measure how much preload force you actually applied, unlike with a wave washer where you can measure force via its deflection. So Shimano tries to make up for this by providing a torque spec for the preload adjustment bolt - a very low torque, such as 1.5 Nm
@@TypeVertigo Correct. In a shimano BB, Hambini may verify, but i think they actually use proper angular contact bearings which dont mind quite of a lot of axial preload, but yes setting it by hand is rather arbitrary.
Interesting. So when you say “Shimano BB”, would this be a Shimano branded BB or any Shimano type BB? ie: Would say, a Praxis T47 BB also use AC bearings?
@@BikePappy Normal Praxis stuff is not angular contact. If it is they specifically advertise as such, eg "Ceramic Version:Enduro 2437 Angular contact, Grade5 balls, ABEC5 Assembly". Cost is $240 vs 90 for the regular version.
Buy a new hub axle, turn down the shoulders 0.2mm on each side and then back to back test it! I'm sure the difference could be measured but I'd love to see if the difference your describing can be perceived. You could even have someone else assemble each setup without you knowing which axle was in it to eliminate any potential bias.
fyi there is no need for another axle: set of standard shim rings will do. They come in thicknesses from 0,05. Making the assemly as a blind test is smart though...
So to make the wheels stiffer simply put one .05 shim on outside of each bearing on the rear axle? Does this apply to front axle as well?
haha - nice homage to the other guy with the pen quip! 🤣
there is something wrong with calculation you said i get something in the middle but it depends on the fitting not only chart.before you have to calculate fitting and you have to measure shaft and inner ring than you have to calculate how much clearance left for example inner ring in 19.97 mm and shaft is 20.00 that means you have 3 microns fitting than you have to subtract from maximum celarance than you find the correct celarance.chart is just for in run clearance not depends on stok tolerances
Bearings are Ok but what do the stickers on the carbon look like? That's the important part!!
That intro 🤣
So just machine 0.2mm off each of the axle faces and you get your prefered arrangement = simples !
Surprised not to see the orange beaver in comment section till now.
Fuck that's a good intro
Build your own then!
Never knew such simple topic would take 20 minutes...
But boy how good has your presentation become! If you had to develop these skills at work, I'd consider firing your colleagues
Also we call these O and X arrangements (put Stroheim meme here)
I can never remember which is 0 and which is X so I don't use those names! But you're right!
@@PeakTorque why remember when you have a drawing of forces in correlating shapes... Leave learning and remebering to undergrads)
That ‘hello’ is becoming cycling meme?
i am not sure, but your intro voice saying hello, sounds more like Hambini.🤥
17:08 comical genius hahaha
Nice intro 👌🤣
Very good magnificent
hambini's ears must be burning
Oh.... no double wishbone :(
Thanks for this video. Very interesting. I have a set of DT Swiss disc wheels with 240S hubs. The original front hub bearings felt rough off the bike, so I pushed in a new set of NTN/NSK bearings (I forget which) and to my dismay they also felt rough. I started to suspect that the shoulders on the thru axle had been machined too wide for the hub, creating excess axial load. I then found a maintenance guide on the DT Swiss site that advises beginning by pushing the non-drive side bearing fully home against the shoulder, before inserting the axle and then finally pushing the drive side bearing only as far as needed to eliminate bearing play (i.e. NOT fully home against the hub shell shoulder). DT Swiss claims that the bearings will always feel a bit rough when off the bike and will only run smoothly once the wheel is fitted to the bike and the axle compressed by the end caps. Does this smell fishy to you? Obviously, I have no way of testing how smooth/freely the bearings are running once the wheels have been fitted to the bike.
Dare I say it, that's not just fishy, but double bullocks at least:
- DT are known to mess up radial tolerances, thus axial fiddling won't help
- when the bearings are overly preloaded in the manner the manual advises, extra pressure when the wheel is secured will make things even worse
@@feedbackzaloop Thanks for your thoughts. I considered getting some shims to sit behind the outer bearing races, to pad them out laterally. For now, I have pushed one bearing fully home - the other, I pushed in bit-by-bit, testing the axle for any play and pushing a little bit further until no play remained. The bearings now feel OK outside of the bike. God knows what they are doing when fitted into the bike and the thru axle tightened 🤷♂️
@@trevekneebone369 It certainly does not sound like good practice. The final bearing position should be up against the shoulder, so it sounds like they messed up!
I would agree with the above, this is not good practice. it's basically a bodge to leave one of the bearings in "free air"
@@Hambini Thanks. So what is the fix in my case? Machine off a slither from the axle shoulders? Place a shim between the hub shell shoulder and the bearing outer race? Thanks!
You know someone what they're talking to if they can compress it to the fullest extent (not BSting some University Prof.)
whats up with that hambini voice?
and i thought people who like to discuss these races and those races are bad company. /jk excellent video well done
HELLO!