Heresy: Do ball bearings really make sense in headsets?

[johnbryanpeters]

A prior thread got me to thinking, which is always dangerous.

Why the hell do engineers use ball bearings in headsets? The application is an invitation to race failure since it's a high impact application with no rotation to speak of, so false brinelling seems inevitable.

Plain bearings would seem to make a lot of sense, with the added benefits that

They would be easier to maintain.
They would provide some damping, which is not a Bad Thing.

So, why the hell do engineers use ball bearings in headsets?

J

[Echo]

I remember a couple years ago I read a long writeup by Bontrager about it making more sense to use needle bearings in a headset... apparently there aren't enough people that think it's necessary though. Seemed like he had a lot of good points.

[D_D]

I think conical needle bearings are best.

Brinelling appears to be more of a problem on road bikes where the forks rotate very little creating lubrication problems.

You have also got to remember that for the vast proprotion of bikes sold a £5 ball bearing headset would never be replaced in the life of the frame.

Nobody buys a bike because the headset has needle bearings so where is the incentive to fit them on new bikes?

[Ridemonkey]

Onza used to make one.

I think the problem was that alignment had to be perfect or it would bind.

[D_D]

Originally posted by Ridemonkey
Onza used to make one.

I think the problem was that alignment had to be perfect or it would bind.

Fsa also make them but the seals are badly designed and don't work.

[johnbryanpeters]

And tapered rollers, while offering more contact area, have the same sort of problems balls do. No velocity, no planing on a lubricating film.

J

[oldfart]

Your right about ball bearing headsets. I used to go through even good Campy or Dura Ace headsets in less than a year on my road bikes. Small bikes=short headtube=more stress on the headset. I had a Stronglight roller bearing headset and it worked really well. It had floating races which alligned themselves. Now I find I need to use expensive Chris King or Syncros headsets. The Cane Creek hidden headset on the road bike seems the doing fine so far.

[ohio]

The stronglight on my road bike uses angular contact needle bearings (conical). Works great, and is super light. Requires more preload than most though.

Needle bearings don't deal well with contamination or misalignment. Ball bearings handle it well, are cheap and readily available in many different sizes, and their strength ratings are more than sufficient. It's just a question of lubrication.

Someone (blackspire maybe?) tried using a thrust bearing coupled with radial needle bearings... not sure if it worked, but it was probably heavy and had a big lower stack height.

If you could seal it well, high quality full contact bushings might work well.

[SwisSlesS]

A lot of FSA headsets use needle bearing in the lower cup. I've never seen one with needle bearings in the upper cup, which I've never really understood.

[GrahamKracker]

Originally posted by swiss_less
A lot of FSA headsets use needle bearing in the lower cup. I've never seen one with needle bearings in the upper cup, which I've never really understood.

There is no reason to use a thrust bearing in the upper cup. Baring a freak accident, upper headtube bearings do not have to with stand major thrust loads, whereas the rider's weight and every single impact are transfered to the frame through the lower cup's bearings. BTW, Woodman still makes and sells a headset w/ a thrust bearing setup.