Brinelled - on the turbo?

[T42]

Moving my old Ti warhorse the other day I was surprised to notice that the steering was quite badly brinelled. I've only done one 90k ride on this bike this year, and I noticed nothing amiss. I used it on the turbo during lockdown, which is when I might have expected it to happen, but the 90k was done soon after lockdown ended.  The bike's been sitting unused on the turbo ever since. Could four months without moving the bars have done this?

FWIW, the bearing dates from 2004.

[mzjo]

Moving my old Ti warhorse the other day I was surprised to notice that the steering was quite badly brinelled. I've only done one 90k ride on this bike this year, and I noticed nothing amiss. I used it on the turbo during lockdown, which is when I might have expected it to happen, but the 90k was done soon after lockdown ended.  The bike's been sitting unused on the turbo ever since. Could four months without moving the bars have done this?

FWIW, the bearing dates from 2004.

Other possibilities (to my mind) would include condensation in the bearing causing rusting at the ball contact points (due to inactivity) or simply the grease going hard and soapy due to the inactivity. A mixture of all three causes is possible!

[Brucey]

there will have been tiny movements even on the  turbo, and that might have help create a fretting situation, especially if the grease has dried.  The slightest sign of conditions that might promote corrosion will greatly increase the erosion rate.

cheers

[T42]

Thanks, both. Oh well, a new bearing after 16 years isn't exorbitant. I don't feel like messing about with it so the LBS can do it.

Cheers.

[giropaul]

When I was involved in teams we noticed that the spare bikes somehow ended up with rough headsets despite rarely being ridden.
On a race car they are carried with the front wheel in on the car roof. It turned out that the small vibrations , left and right, due to airflow and road surface was the cause.

[Asterix, the former Gaul.]

Apparently the ball bearings very soon become ovalised and they start to become roller bearings.  So long as they are left in position they carry on behaving as normal even though worn but if something disrupts the plane* within which they operate they don't like it and they cease to behave nicely.


* this refers to the mathematical plane, not the carpenters' plane, the spiritual plane or an aircraft.

[Brucey]

it has been my observation that bikes on roof racks are prone to what aerodynamicists sometimes call 'flutter', i.e. the bike is getting flapped side to side all the time at certain speeds.  This could indeed tend to cause headset bearings to start fretting.

cheers

[Kim]

Seems to me that a bike kept attached to a turbo won't have the usual large steering movements of one that's being used in the real world, which presumably play an important role in distributing grease.

I discovered last month that the headsets on barakta's trike are in a shockingly bad state.  It's only done about 2000 miles, but has spent a fair time in turbo mode in recent years.

[hulver]

Sweat? My headset looks in a fairly shocking state, the top cap bolt is pretty rusty. Mostly I think due to the vast quantities of salt water I drip on it while I'm on the turbo.

If it ever comes off the turbo I'll have to fit a new headset I think.

[hatler]

Brinelling is different to fretting.

Brinelling is plastic deformation of the bearing surface as a result of an overload.

Fretting is erosion, probably through lack of lubricant.

[LittleWheelsandBig]

Well yes but most headsets die by fretting, which is known in the bike trade as Brinelling because of the appearance. Plastic deformation of headset cups and cones is almost impossible, short of whacking them with a sledgehammer.

[Brucey]

Well yes but most headsets die by fretting, which is known in the bike trade as Brinelling because of the appearance. Plastic deformation of headset cups and cones is almost impossible, short of whacking them with a sledgehammer.

I know that Jobst Brandt thought that, and that this has become received wisdom amongst cyclists.

 However I think he was wrong.

 He assumed that the headset always allowed the balls to share the loads between the balls to some extent.   This is certainly not the case, especially if you have a long (flexy) 1" steerer (as he did).  Loads can then get into the hundreds of kg per ball range during normal riding.

He also argued that if the material were displaced (rather than removed) you would see it as an upstand around the indent.  However if you actually do the sums you will very quickly conclude that an annulus of upwardly displaced material surrounding an indent would only have to be a small fraction of the height/depth of the indent in order to account for the displaced volume; in other words it might be so slight that it wouldn't be detectable without specialist equipment.

I think that damage on a roof rack or on  a home trainer is most likely fretting damage. However to assume that true brinelling damage cannot occur I think is wrong.

In support of this line of thought, I offer the fairly simple observation that overtightened headsets can be damaged by riding  just a few miles; this is just not enough cycles for a fretting mechanism to do much damage.

cheers

[LittleWheelsandBig]

Excessive contact pressure removes lubrication. Fretting becomes unavoidable and occurs quickly.

[Brucey]

Excessive contact pressure removes lubrication. Fretting becomes unavoidable and occurs quickly.

fretting in bearings can occur in the absence of a significant contact pressure. In near static-bearings the lubricant film is routinely displaced at the ball contacts; it is the nature of the beast. If such conditions invariably caused what you call 'fretting' (which might not be the same thing as other people call 'fretting') then pretty much every headset would be damaged, not just some.

So it certainly isn't as simple as A follows B.  Nor does fretting occur quickly; not as quickly as overtightened headsets can get damaged. Fretting damage in industrial bearings usually occurs when the assembly is static and  exposed to vibrations over a lengthy period of time; millions of cycles are required.

Classical fretting almost invariably includes a chemical erosion element; most usually the debris can oxidise and an assembly which has been fretting is often seen to be full of powdery rust.  If there is grease present the wear debris from fretting would soon discolour the grease.

As further evidence as to why the fretting hypothesis is wrong most of the time in failed headsets  is given by the ball bearings themselves; for the fretting process to occur the balls have to cold-weld themselves to the other bearing surfaces to some extent. This should damage and/or deposit material on the ball bearings as well as wear the raceways. However plenty of headsets get damaged and the balls are unmarked (even when examined under a microscope), and the grease is still clean.   These headsets cannot possibly have fretted.

No, Brandt's fretting hypothesis doesn't stand up to more detailed examination.  I'm not saying that fretting  can't ever happen, more that most headset damage that is attributed to fretting by cyclists who have read what Brandt wrote isn't really caused by fretting at all, and that Brandt's reasoning why "it can't possibly be brinelling damage" is easily proven to be false.

cheers

[LittleWheelsandBig]

Have you checked the surface hardness of headset contact surfaces compared to ball bearings?

[Brucey]

Have you checked the surface hardness of headset contact surfaces compared to ball bearings?

there is more to it than that alone;  when balls indent into surfaces at high loads, yield occurs subsurface.  This means that a ball that appears to be softer than a substrate surface can indent into that substrate if the hardness is not uniform with depth. 

This illustration isn't exactly correct for several reasons but it shows more or less what is going on



You can see that the material is stressed and may yield to a depth (and radius) beyond the indenter  that is comparable to the diameter of the indenter. One effect  of this is that when doing hardness testing, you need to space the indents apart using a pitch that is at least x3 the diameter of the indents, else the yielded regions can overlap and this can affect the results. It also means that indenting into hardened surfaces (like ballraces) can produce some weird effects; for example the 'pile up' region can be larger in diameter and less high that you might expect, because the plastic zone beneath the indenter is the shape of an upside down mushroom.

Bearing surfaces are commonly graded in hardness with depth and (depending on the diameter of the indenter (ball)) the surface stress at which deformation occurs will vary, largely because the stresses can easily extend downwards into considerably softer material, and cause subsurface yield.

IIRC Brandt's argument was more or less that it couldn't be brinelling because the loads per ball were not high enough (about which he was mistaken on several counts)  the marking on the raceways was the wrong geometry (again a flawed supposition for several reasons) and he therefore inferred that fretting must be occurring by a process of elimination. The fact that the number of cycles for true fretting to occur was usually completely different, and there were typically  none of the usual signs of fretting (such as wear debris) in most cases didn't seem to worry him in the slightest.

So yes fretting is certainly possible (on the turbo, or a roofrack for example) but on the road, meh, the arguments are a lot less convincing.

cheers

[LittleWheelsandBig]

I have tried hitting ball bearings in a cheap steel headset cup with a hammer and punch. The ball bearings developed flat spots. The cup didn’t brinell. YMMV