Wheel-building – replacing a hub

[Samuel D]

P.S I had to buy 11 speed (5800) as I couldn't find 5700 now in 32 hole versions.

I looked long and hard without success for a silver, 36-hole, FH-5700 before I needed it, knowing that I would eventually want it, because I don’t want an 11-speed hub (with the right flange farther inboard) for my 8-speed cassettes … now or in the distant future. It was only happenstance that it had become available on eBay in the meantime.

Bike-Discount in Germany are selling off 32-hole FH-5700 hubs for cheap, if you’re still interested. A little less cheap after shipping.

I also bought the 105's because I think they have better bearings than the Tiagra.
The advert for the 105 say "Cubic Boron Nitride treated cone races" which to me implies that Tiagras aren't ?

I think that was the case in the past, but the FH-RS400 has CBN-treated cones.

Whether that makes any difference I don't know and is it just the cone or is the cup treated as well ?

The difference I don’t know either, but it’s just the cones. According to Shimano’s specifications PDF, you have to go all the way to Dura-Ace to get CBN-treated cups. The rest down to FH-RS400 just have a “polish” finish; and below FH-RS400 you don’t even get that or CBN cones (although surely they must be polished anyway?).

531colin: thanks for that tidy explanation. I’ll give that a go when I receive the eBay hub.

[Brucey]

....I confess this is my own rear wheel this time. It’s a Shimano 105 FH-5700 hub. For several months there has been a light clicking once per wheel revolution when freewheeling at certain speeds. Having tried other things, I thought I’d replace the freehub body to see if that was the problem. I greased the fixing bolt per Shimano’s instructions, screwed it in, and promptly stripped the female threads off the alloy hub. (I need to get a big torque wrench soon to see what these torques feel like.)

The threads continue for 15 mm inboard of the stripped length, but I don’t know where I’d find a much longer freehub bolt to reach them.....

You can buy the bolts separately, but I don't think you will find a longer bolt; however you might be able to make one (eg by welding parts of two together).

I don't think you should beat yourself up about having stripped that thread; I suspect it was already damaged.

Shimano have changed the way they make those hollow bolts; older ones were clearly made from wrought tube and had a rolled top. The most common thing was that the bolt yielded near the top and the rolled top unrolled itself (and just pulled through) if you overtightened the hollow bolt; I have seen several go this way. However I have only ever seen this happen when an allen key was used to tighten the bolt, and the allen key was difficult to remove afterwards. I think that if you use a proper 12pt spline bit,



inserted fully so that the shank helps to support the top of the bolt, this mode of failure is much less likely. The newer bolts (of which there are several different styles) appear to be drop forged and have more substantial tops. They also seat differently in the freehub body. I have not seen a revised bolt fail no matter how it is tightened, although it may be possible to strip the thread I suppose.

The reason I think the threads were already damaged is that you report a noise when freewheeling only. If the bolt isn't quite tight enough the torque of pedalling may partially stabilise the position of the freehub body but when freewheeling the freehub body may move more easily, making a noise as you describe. The tell-tales that the freehub body was loose are that a) the bolt came out easily and b) that there are signs of rubbing in the interface; a kind of grey sludge is common. if there is a freehub body washer, this can make a noise as it hits the side of the groove it sits in.

Obviously any movement tends to wear the thread in the hubshell; again sludge is quite likely.

FWIW I always try to remember to check the freehub body bolt whenever I have a freehub axle out. They commonly go up a fraction of a turn; you don't tend to find one that is actually loose that hasn't also caused some damage to the parts, so I think it must go bad pretty quickly if the bolt works loose past a certain point. It seems to me that some movement at the interface is inevitable; the steel parts settle into the aluminium hubshell and if a body washer is fitted, you may find that it has burrs on it that prevent it from seating properly first time. I have deburred many body washers, so as to make them seat better.

BTW my preference is (if it is snug to start with) to leave the freehub body on the hub when overhauling it. This means that it isn't disturbed and it doesn't have to reseat itself in a slightly different place when it is refitted. I suspect that it may settle all over again if it is refitted with a different spline timing, which increases the chances of the bolt working loose perhaps.

cheers

[Samuel D]

7) clean the nipples; e.g. put the nipples into a small sealed container ( a 35mm film container used to be handy for this), give them a generous squirt of GT85, and then give the container a good shake. Usually the nipples come out looking like new, if the plating is still intact.

Thirty seconds of vigorous shaking worked amazingly well. Look at these beauties:



And pwa wants me to throw them out!

[pwa]

7) clean the nipples; e.g. put the nipples into a small sealed container ( a 35mm film container used to be handy for this), give them a generous squirt of GT85, and then give the container a good shake. Usually the nipples come out looking like new, if the plating is still intact.

Thirty seconds of vigorous shaking worked amazingly well. Look at these beauties:



And pwa wants me to throw them out!

I'll send you mine

[Samuel D]

Ha! Spokes nipples are minor miracles of the industrial age. How can things so intricate and precisely made (where it matters) cost so little? Iron Age people would have treated these things with the respect they deserve.

[Brucey]

I also bought the 105's because I think they have better bearings than the Tiagra.
The advert for the 105 say "Cubic Boron Nitride treated cone races" which to me implies that Tiagras aren't ?

I think that was the case in the past, but the FH-RS400 has CBN-treated cones.

Whether that makes any difference I don't know and is it just the cone or is the cup treated as well ?

The difference I don’t know either, but it’s just the cones. According to Shimano’s specifications PDF, you have to go all the way to Dura-Ace to get CBN-treated cups. The rest down to FH-RS400 just have a “polish” finish; and below FH-RS400 you don’t even get that or CBN cones (although surely they must be polished anyway?)..

the manufacturing steps for bearing surfaces (in bicycle hubs) typically vary roughly as follows

a) machine/form to shape, then harden.
b) machine/form to shape. harden, then grind
c) machine/form to shape, harden, grind, then polish.

A complication is that the materials used (and the forming/hardening processes) vary with the parts used. Dura-Ace (and posh campagnolo) cup and cone parts are machined from solid, using steel that is hardened to a considerable depth (it is usually hard right through in fact). The less expensive hubs used pressed steel cups and various flavours of cone material which usually have a hard layer on the outside but are somewhat softer on the inside. [When cheaper cones suffer pitting, it appears as if there is a whole (hard) layer waiting to come off the surface; this is often exactly the case.] Shimano also mix and match; for many years there didn't seem to be much difference in the bearing cups across most of their range, whereas variations in cones were more apparent.

Boron Nitride is an 'electron analogue' of carbon; it therefore manifests different forms (depending on how it is made) which have remarkably different properties. Thus 'Hexagonal Born Nitride' is structured somewhat like graphite and has similar properties too. In fact it is often called 'white graphite' and is used as an ingredient in some lubricants. Cubic Boron Nitride is structured as diamond is; it is incredibly hard (hard enough to polish hardened steel with) but very much cheaper to buy, of course.

So without details it is impossible to say for sure but most likely 'CBN treated parts' are just polished a bit better.

All cup and cone bearings will -if correctly adjusted and lubricated- tend to get smoother as time goes on. Likewise all cup and cone hubs will -if they are adjusted badly and/or get water in- suffer damage to the bearings. Adjusting posh hubs is, if anything, more difficult than adjusting cheap ones; excessive preload on the bearings is not so evident, but can be equally destructive to the bearings themselves of course.

[In case anyone needs reminding; the correct adjustment for a cup and cone hub with QR is a little free play that just disappears when the QR is used to secure the wheel in the frame. You can check for any binding/roughness at the service preload condition (i.e. with the QR load on) by tightening the QR onto a few M10 washers when the wheel is out of the bike. You should be shooting for an adjustment accuracy of just a few degrees of cone rotation; to go from 10 microns slack through to 10 microns of preload - which would equate to the tolerances in a good cartridge bearing hub, you would need to set the cones with an angle tolerance of +/- three degrees. If you choose to ignore the QR preload, you can easily impose at least 80microns of preload on the bearing, which could amount to several hundred kgs of extra loading. ]

In shimano hubs, excessive preload may more quickly damage the cheaper bearing parts but if they are set up correctly they can have a service life of many tens of thousands of miles. In most cases I would suggest it is sufficient to adjust the hubs once, then to run the hubs for a few hundred miles (to run the bearings in), then to clean/regrease/adjust them properly. Once this is done the bearings will last a very long time before they need doing again, and cheaper models of hub will likely soon be running as smoothly as more expensive ones; leastways you won't easily be able to tell the difference.

If you want to make a head-start on running-in or to improve a cheaper hub, you can polish the ballraces yourself. This is quite difficult in the case of a cup (you need a shaped polishing mop and good access to the cup, and even then most of your effort is wasted; the balls only run on a narrow track and polishing the rest of it is a waste of time....) but polishing is rather easy with a cone; just use a short length of old axle to mount the cone in a static electric drill, and polish away. About 1000rpm or more is about right. If the surface is ground already, you can start with 600grit paper or even a polishing compound. If the cone hasn't been ground, you may need to start with a 400grit paper and go from there.

Don't remove too much material; if you do, you may do more harm than good; the cone may lose its exact profile and the hard layer can be reduced in thickness. The latter may not be apparent to start with but the hard layer may fail in service via subsurface fatigue; it needs to be about 1mm thickness or so, and it is usually barely that on a new cone. Cheap cones often have a surface that, even when polished, still has a few tiny pits in the surface. This doesn't seem to cause big problems, and if the surface is smooth between the tiny pits the bearing can still run smoothly.

Nice work with the nipples, BTW; it is quite a good dodge that, even if I do say so myself.....

cheers