How tightly should I do up the cone caps/dust covers on my axles? I read somewhere that there should be a bit of play but is that correct? I did them up tight enough to just be in contact with the bearings and then secured the locknut.
Should there be play? I don't like the sound of leaving things that wobble about on a wheel but maybe when the wheel is installed on the bike the play is largely eliminated - if so why leave play in the first place - surely it just gives water an excuse to get in doesn't it?
JUST tight enough to remove the free play, but no tighter... You don't want free play though, this will get worse over time.
In my experience, tighten so they run smoothly, and then back off a tiny smidge, as doing up the locknut takes out any play.
If you get them just so and then do up the locknut, they tend to be slightly too tight.
You generally need to leave a little play as fitting the wheel will remove it. Really you need to use trial and error, generally if there's zero play with the wheel off the bike it will usually mean slightly rough bearings when you install the wheel.
What @dc1988 says.
You want a small amount of free play which the qr will take out when you put the wheel back in the frame.
As the last three posts have observed, it takes a level of skill to get it right. I've spent bloody hrs doing it. An engineer mate told me to remember that all bearings have some free play, otherwise they are too tight. So I always end up with some free play, the skill is in trying to minimize it.
You also need to check you are doing the we up enough to remove the play - I don't know of it's the same with modern hubs.vs old school qrs.
But at the end of the day you don't need to get it perfect - lots of hubs don't have the play and so are too tight in use - I've not seen that do any harm.
Thanks all - sounds like I was on the right lines then. I spent a good hour messing about with it and I think it's OK. Too much snow and ice for a test ride atm.
Ive got a thru axle that just screws up to 12-14nm (rather than a QR lever) and I do it up to 13nm each time, so the wheel is always clamped in the frame with the same force.
I just serviced a set of Alivio hubs today (that are far superior to the XT hubs I was running 25 years ago). Get them about right, clamp the wheel in the dropouts. Side to side play? Nip it up a quarter turn. Not spinning freely enough? Back it off. Repeat. Might take a couple of tries. A bench vice helps.
@oceanskipper - Old tip from The Sheldon era was to use a set of old dropouts with the wheel out of the frame but clamped to ensure the tiny bit of play gently disappears.
With your through axle, could you try similar with in terms of spacing?
The tiniest smidgen of free play is taken up by tightening the QR.
As above - the axle is slightly compressed when tighten by the QR, the aim is replicate this when adjusting the cones rather than the guesswork of leaving a small amount of play that is taken up by the compression of the axle.
I used to use a spare QR with some thick washers that replicated the axle/frame setup and compressed the axle whilst I was twiddling the cones.
Surely the free play will worsen once the grease has settled in? Would it not be better to have them fractionally too tight to start off with?
THIS is one of the reasons I HATE cup and cone.
surely it just gives water an excuse to get in doesn’t it?
Tightness of cones has no effect on the amount of water that can get in.
It's a long time since I've done cup/cone and I think they are a pain in the rear now I'm used to cartridge... But there is a wonderful feeling upon getting them "just right". I do kind of miss that a bit.
If looked after, they can also last a ridiculously long time too.
I put more muddy miles in, in the mid 90's than I ever have since and it was all on cup/cone bearings. With a bit of regularish care they last forever... I'm not sure I'd say the same for cartridge.
I do prefer cartridge these days though.😉
On quick release wheels the compressive force of the QR squeezes the bearing together, removing some play. Is the same true of through axle systems?
If you use washers or spacers as dummy dropouts you can use the QR to compress the axle while still leaving access to the lock nut on the side you are adjusting. This is sometimes called a hub vise. You can then adjust them precisely. This eliminates the guesswork and speeds up adjusting them a lot.
This method is described by Sheldon brown in his page describing cup and cone hub adjustment.
I've always just got them perfect, spun the lock nut on and tightened it by hand. Then snugged the cone up to the lock nut. And chucked it in.
I used to have to do cup and cone hubs quite a bit and not take too long over it (group trips, cheap hubs). Similarly (the same?) to joshvegas, I used to wind the cone on as tight as I could with my fingers, snug the locknut up behind it and then tighten the cone and the locknut up against each other. The slight backing off the cone would do generally gave a free running but not loose bearing. If not, it was close enough to be an easy tweak second time.
Disclaimer- cheap hubs, threaded (not qr) axles and I haven’t done it for years!
That might be fine if you are well practiced and know how much play to leave so that the QR removes it, leaving it in perfect adjustment. I never got that good and ended up taking the wheel in and out of the frame several times to readjust. People who are not so familiar with the process are likely to adjust the hub so it is running perfectly in the hand (out of the frame) resulting in it being too tight when clamped by the QR. Excess preload will lead to premature wear.
Using the hub vise removes the uncertainty. A stack of 10 M5 washers taped together is all you need for a spacer and you can be certain that it is not running too tight. I thought it was bit of a faff from Sheldon; description but really it is very easy to assemble and use.
If you get them just so and then do up the locknut, they tend to be slightly too tight.
Am I right in saying that you are just using 2 cone spanners, for this?
If so, then this is because the axle, cone and locknut all move independently. When you apply force to the spanners, the cone often moves a little, before the locknut takes effect. This is why an axle vice helps, as it stops the axle from moving and, providing you hold the spanner which is attached to the cone, completely still, the only thing that moves is the locknut.
You still need to allow a little bit of play, to allow for the force applied by the QR
Can someone explain to me how tightening the QR increases the preload? You're not compressing the axle with the force from the QR, and if the locknuts are tightened against the cones, the force from the QR isn't going to move the cone down the thread. What am I missing?
It must compress slightly, how else does the play disappear when a qr is tightened
Can someone explain to me how tightening the QR increases the preload? You’re not compressing the axle with the force from the QR, and if the locknuts are tightened against the cones, the force from the QR isn’t going to move the cone down the thread. What am I missing?
You are compressing the axle with force from the QR. It's not moving the cone down the thread, it's moving both cones a fraction of a millimeter towards each other.
https://www.sheldonbrown.com/brandt/cone-adjustment.html
Fair enough. I had assumed that the force developed by a QR skewer/lever wouldn't be enough to compress the axle (moreover, by more than the skewer itself stretches) an appreciable amount, but if Sheldon says it is...
I don't buy into that. I suspect you only get that effect if the locknut is not tight enough against the cone so that they both move marginally on the play they have on the threads. I always made sure that the locknut and cone were tight enough so that they were't affected by QR tension and so that they didn't come out of adjustment, as that's the recipe for damaged bearing surfaces.
You don't have to buy into anything because you can easily demonstrate it for yourself in about thirty seconds.
You don’t have to buy into anything because you can easily demonstrate it for yourself in about thirty seconds.
In a similar way to "the laws of aerodynamics should mean that a bee can't fly"?
<blockquoteYou don’t have to buy into anything because you can easily demonstrate it for yourself in about thirty seconds.
I've not found that to be the case when locknuts are properly tightened. Spannered in shops for many years in 90s, so have plenty of experience of this.
I don’t buy into that
Living up to your username there. I expect you didn't put a hub with any play into a bike? That's what you'd need to have done to see the effect.
It really is tiny, possibly ~1mm at the rim (and possibly also not that important). Fell free to.come round and I'll show you mine (ooh er etc).
Thanks for your input.
I’ve not found that to be the case when locknuts are properly tightened
They're properly tightened when the QR removes the last trace of play. HTH.
In a similar way to “the laws of aerodynamics should mean that a bee can’t fly”?
Exactly that, never argue with someone about something that can be easily demonstrated by a simple experiment. If someone is too fing lazy to put a QR and clamp it on the axle to see for themselves then no amount of forum posts is going to convince them.
...and this is what makes cup and cone adjustment a black magic art.
It used to be a deep mystery for me. That ended when I started compressing the axle prior to adjusting the preload.
…and this is what makes cup and cone adjustment a black magic art.
I really don't think it's that hard. I speak as an average mechanic who is less than fastidious about maintenance.
I don't think it's that hard once you have done a few,but I can see why it annoys the hell out of some folk.
Even worse if it's those cheap and nasty cup and cone units.
How tight should a QR skewer be?
The vice idea sounds good, preclamp the whole thing so you can set the tension as it would be.
Go and have a search on the CUK forum if you really want to go into this.
as that’s the recipe for damaged bearing surfaces.
Equally over tightening them as you have been doing is also a great way to damage bearing faces.
So from that video and a quick look on the DT Swiss website it would seem that DT Swiss hubs certainly DO make use of the clamping forces. I'm not convinced that cup and cone bearings do though as I would have thought that the whole lock nut business means that it doesn't move.
Or am I completely missing something? I can't seem to find ANY reference on the Shimano website on how tight to do up the cones it just says "After assembling the necessary parts and adjusting the rotation,
tighten the lock nut to double-lock the assembly."
So are cup and cone bearings really the work of the devil? Lots of people seem to hate them.
I have ordered some Shimano Deore XT wheels (which have the dreaded cup and cone bearings albeit a version called Labyrinth???) )and I am thinking I made a mistake now. Trouble is I needed to make sure whatever wheels I ordered would take my centre lock disc rotors and microspline 12 speed cassette with a 12x148 axle and I couldn't see any DT Swiss wheels that would fit...
I could try and cancel the order...... 🙁
I’m not convinced that cup and cone bearings do though as I would have thought that the whole lock nut business means that it doesn’t move.
If it doesn't move, how is the play taken up?
If it doesn’t move, how is the play taken up?
It isn't and therefore there shouldn't be any play. My guess is they need to be done up tight enough so there is no play but not so tight that they wear out... Hence the difficulty I guess. 🤷♂️
Next time I take my wheel to bits I'll try it but trying to determine if any play is taken up will be tricky me thinks...
The vice idea sounds good, preclamp the whole thing so you can set the tension as it would be.
That's not what an axle vice is for...
...it just stops the axle from rotating, while you adjust the cone and locknut...
(The beginning big that video shows a bodge, but gets onto the axle vice, later)
It isn’t and therefore there shouldn’t be any play
I must be imagining it then.
As mentioned above, the Cycling UK forum loves this topic.
Cup and cone bearings are good. You can maintain them easily and all you need is grease and a rag and a few simple tools.
I think QR Shimano hubs are brilliant. Tough flanges and big bearings in the right place. I am not sure about through axle ones as I don’t have any Shimano through axle hubs. Generally, as axle sizes increase bearing sizes decrease which I don’t like.
And the thing about setting cones so they are slightly loose is true. Stick them in the frame, tighten the QR and the play is taken up. People say it’s the axle compressing. I wonder if it the axle bending slightly. Whatever it is, the system works.
I think QR Shimano hubs are brilliant.
Admittedly, I only got 20 years out of my last set.
I’m not saying that you have to do the loose then tight thing. It may help but then again many don’t do it with no problems.
When I say it’s true, I mean it’s true that slight play will usually be taken up when you tighten the qr.
@oceanskipper you say you don't know about c+c hubs and yet you know enough to say this?
Next time I take my wheel to bits I’ll try it but trying to determine if any play is taken up will be tricky me thinks…
Adjust a bit of play into the hub, put it in the frame, feel the play, tightener the qr, see that happens.
well I’ve only taken a wheel apart once ( well it ended up being about 5 times as I did them up obviously too tight the first time and too loose the second and third!) so clearly that won’t make me an expert in c+c, hence the start of the thread. I have done a lot of reading since then though. 👍
EDIT I did think I had left the teeniest, tiniest amount of play I at one point but it did not disappear, whether it was reduced or not I couldn’t tell but I don’t think I could have made much of a finer adjustment on the cone to retest under the exact same conditions.
I had assumed that the force developed by a QR skewer/lever wouldn’t be enough to compress the axle (moreover, by more than the skewer itself stretches) an appreciable amount
I've just tried to work this out. (Disclaimer - I am not an engineer.)
Young's modulus specifies how much the axle will compress - dL = F x L0 / A x E.
Force F = 3000 N (from the Peak Torque video linked to above)
Axle length L0 = 130mm
Axle area A... I think a QR axle is a 10mm axle with a 7mm hole for the skewer - so A = Pi * ((0.005)^2 - (0.0035)^2) = 4 x 10-5 m
Young's modulus E = 200 GPa (from Wikipedia)
So change in length dl = 3000 N * 130 mm / (0.00004 m2 * 200,000,000,000 N/m2) = 0.05 mm
The thread pitch on the axle is perhaps 1mm so that's 1/20th of a turn.
That's in the right ballpark - it's too low by about a factor of 3 on the observed effect but I've likely got some of the numbers wrong. So I'd say this is inconclusive on proving or disproving axle compression as the cause.
I think that's actually pretty close to the observed effect IME, somewhere in the region of tenths of a turn?
A good recent thread that also links to many of the other relevant past discussions at the cycling UK forum. https://forum.cyclinguk.org/viewtopic.php?f=5&t=143125
I've always thought it wasn't the axle compressing, but the threads flexing, being as they are perpendicular to the force you're applying when you tighten down on the lock nut, and transferring that force to the axle.
you get used to how much play to leave, depending on how much the lock nut grabs the axle and turns the whole lot when you tighten, essentially loosening the cone off a bit. I'm so used to it that I had to re-learn a bit when I first used an axle vice.
XT hubs are a bit different though. adjust the same, but different as you can't really see the axle turning when you tighten the locknut, as it threads over the end of the larger diameter alu axle (for the QR version).
Right - 1/10 of a turn sounds like a good starting point and gives a really good insight into the kind of tolerances we are talking about. Thanks.