Mark, Ted et al,
Most commercial precision bearings like those found in autos, trucks, machinery and so on are made of SAE/AISI 52100 high chrome, high carbon alloy bearing steel (DIN 100Cr6, BS 534A99/535A99). These kinds of bearings are precision ground and the alloy can be hardened to at least 63 Rc.
I have no idea what kind of steel is used for the cones and cups in bicycle bearings. A lot depends on the brand and quality level.
The loose ball cup and cone style bearings used on bikes allow for a lot of misalignment of the bearing components.
The cups and cones on most cheap bike components are usually "case hardened" and made of low grade low carbon SAE/AISI 1018 steel (DIN CK15). After machining the parts are heated to red hot in a carbon rich atmosphere were a certain amount of carbon is absorbed in the surface of the parts.
When quickly quenched the surface (or case) of the parts reaches a hardness 60 Rc to a depth of between .005" and .050". The steel under the surface remains relatively soft and tough.
Once the bearings wear through the hardened surface layer the cups or cones wear out fast.
There are several alloy steels, for example 8620 and 9310, that are used for making case hardened precision bearings for specific applications. The heat treatment and case hardened depth is much closer controlled than with 1018 steel.
Last year I wrote about testing 18 headsets back in the mid 1970. Most of the cheap ones were obviously case hardened. Zeus and Stronglight P3, V4 and S5 headsets were all over 60 Rc while all of the Campy NR headsets we tested between the mid 50s Rc and as low as 47 or 49 Rc.
I have a bag with half dozen Campy headsets and lower races that I need to check the Rc on. I'll report the results.
So, to answer your questions about through hardness, I don't have an answer. The only way to tell would be through destructive testing: cut the parts in half and measure the hardness on the surface and in the center.
Here are several links that discuss bearing failures. Bear in mind that these are precision bearings under the kind of loads that you would rarely encounter on a bike.
Check out the section on corrosion in this link:
http://tinyurl.com/
look at flaking in this site:
http://www.actionbearing.com/
I don't think that there is any one cause for bike bearing failures. It usually a number of contributing factors.
Chas. Colerich Oakland, CA USA
Mark Bulgier wrote:
> Ted Ernst wrote:
>> I've been musing on the regrinding and had a few Q's about the
>> process/results.
>> How much materiel has to come off before the race is thru its hardness
>> depth? Enough to make a problem if ridden?
>> Are the cones/races/cups hardened all the way thru?
>> Would that make the fork fixed cone easy to split/crack if too tite on
>> fork crown/base position.
>> If too much materiel is removed does that increase the radius/radii of
>> the paired combination and change the position enuf to maybe cause
> some
>> very close tolerance parts to have contact/friction?
>
> Ted, yes those questions are haunting me too. I had one lower headset
> cup/cone bearing re-ground, and I think the bike it's on will be a
> rider, so I'll be able to report back after some miles. Unfortunately
> it will take me a while to get it built up, and then it won't be my
> highest-mileage bike so it will take a while to get real miles on it.
> Probably a year before I can report back meaningfully -- at least I hope
> so!! Maybe it'll die inside of 10 miles?! I don't expect it to though.
> I have a feeling (based on zero evidence) that Campy headsets are fairly
> tough throughout.
>
> The clearances may have tightened up on my headset between the crown
> race and the frame, but it still clears, and a miss is as good as a
> mile, yeah? And less room for splash off the tire to get into the
> bearing. Since the grinding just went to the bottoms of the deepest
> dimples and a tiny bit more, and the bottoms of the dimples is where the
> bearings were sitting already before grinding, I don't think the
> grinding reduced the clearances much beyond what the dimpling (wear)
> did. And I've never heard of a headset begin to rub on itself from the
> wear from riding -- you?
>
> Oh and another point about the balls being at the bottoms of the dimples
> -- if that was enough to go through the case-hardening depth, then we
> still haven't harmed it any by re-grinding, we've just taken the rest of
> the race down to where the balls were sitting anyway. I hope Campy
> hardens their ball races to a depth deep enough that you can't wear all
> the way through it in normal use -- but maybe that's just wishful
> thinking.
>
> I don't know how this guy chooses (or maintains) a profile to grind the
> curve into the races, but he did ask me what size balls it uses (in this
> case 1/8"). And though I'm no expert, it sure looks like a race for a
> 1/8" ball now that it's done -- I don't think he changed the radius
> much, or not enough to tell by eye anyway.
>
> This particular crown race on my bike is so damn beefy that there's no
> way the grinding could cause it to split. I've seen more than one Super
> Record crown race split on a tight-fitting crown, _without_ any
> re-grinding, so removing any metal off one 'o those might well be risky.
>
> Mark Bulgier
> Seattle, WA USA