Steering Head Bearings

davidpnorton · Feb 28, 2011

Any good alternatives to the standard sealed steering head bearings out there? Going in '73 850.

grandpaul · Feb 28, 2011

Old Britts has an excellent value priced substitute.

Talk to Ella or Fred, I'll bet they have 'em on the shelf.

builder · Feb 28, 2011

davidpnorton said:
Any good alternatives to the standard sealed steering head bearings out there? Going in '73 850.

Going from memory here but they should be 6205-2RS. ANY bearing shop will have them in stock...usually less than $5 each. Just use the cheap ones, it doesn't matter what country they were made in for this application.

You could go with tapered rollers, but they would require to be greased on occasion...the standard sealed are totally fine for 99.5% of the riders out there.

DogT · Feb 28, 2011

If you use the standard bearings, make sure you have the proper spacer between the bearings, standard bearings are not meant to take perpendicular loads. You can get annular load bearings, but they are about $25 ea. There was a thread on this a while ago. Lets see post88014.html?hilit=annular#p88014

I think there was even another earlier that went into the tapered rollers.

Dave
69S

SteveMinning · Feb 28, 2011

When I built my Manx I labored a good bit over how to configure the steering head bearings. I never trusted the ball bearing (wheel bearing) approach because of the radial load (or is it axial load?) issue and have found tapered rollers to be a pain to keep properly adjusted. I spent too much time varying the adjustment of another race bike that I had experimenting with the ideal adjustment for tapered bearings. Probably my ignorance, but I found it a pain.

I spent some quality time with my local bearing supplier explaining my dilemma. He ultimately provided me with a flanged ball bearing (one race was actually a cup) that was designed to handle load perpendicular to the bearing face. The bearing was pretty expensive (about $85 discounted) at the time so I only used one of them at the bottom of the steering head, and a conventional sealed ball bearing at the top. My thinking was that the lower flanged bearing would handle the load.

The only downside is that it was not available as a sealed bearing, so I used an alloy cover as usually used on top of the steering spindle with a greased felt gasket between it and the bearing. It has worked superbly on the racetrack; I've never had any steering head adjustment issues and it has lasted very, very well. I take the steering head down every other year and clean/re-grease the bearing and have found no issues when I've done so.

I don't remember the number of the bearing off the top of my head, but can probably dig it up if anyone is interested.

Tintin · Mar 1, 2011

SteveMinning said:
When I built my Manx I labored a good bit over how to configure the steering head bearings. I never trusted the ball bearing (wheel bearing) approach because of the radial load (or is it axial load?) issue and have found tapered rollers to be a pain to keep properly adjusted.

Standard ball bearings are not intended to take axial load above a certain level, the ballpark no there is 10% which for the bearings Norton uses in the steering head gives roughly 800N IIRC - that means the bearing is stressed higher than this as soon as the bike is off its main stand. This is just a ballpark no but for me it still leaves a bad taste.

The problem with tapered roller bearings is that they are very easily overstressed by static pre-load. I had a lengthy discussion with Joe S. aka ZFD about this some days ago - he doesn't like tapered rollers from experience and doesn't see a problem with the ball bearing setup. Although I don't fully share this opinion I understand his reasoning and this discussion triggered a bit of thinking on my side about the setup.

IMHO the only proper way to set the pre-load on a tapered roller bearing is with a distance tube of very narrowly tolerated length - very similar to the Norton "wheel-bearing" approach just that the length is far more critical. This way you can torque up the nuts sufficiently and have correct bearing pre-load. Just setting the pre-load by adjusting the nut, yoke and top nut will not really do the trick in the end.

He ultimately provided me with a flanged ball bearing (one race was actually a cup) that was designed to handle load perpendicular to the bearing face.

I'd be tempted to assume that this refers to an angular contact bearing - for the Commando this apporach would lead to e.g. this bearing, a 7205-2RS. It would need the same setup as the taper roller bearing in terms of a narrowly tolerated spacer.

Tim

grandpaul · Mar 1, 2011

So many millions of trouble-free miles have been racked up with standard bearings on Nortons, that it's not even funny.

Then again, even more millions of miles have been racked up on old Triumphs with tiny loose balls in open cup races...

builder · Mar 1, 2011

grandpaul said:
So many millions of trouble-free miles have been racked up with standard bearings on Nortons, that it's not even funny.

Then again, even more millions of miles have been racked up on old Triumphs with tiny loose balls in open cup races...

Exactly...fixing a problem that doesn't exist.

DogT · Mar 1, 2011

True, but I'm not sure I would replace annular thrust bearings with standard ball bearings without the appropriate spacer to take up the thrust. That's what I was about to do. I think that's where the issue comes to play.

Or are you saying it is fine to put standard ball bearings in the head without the spacer?

Dave
69S

grandpaul · Mar 1, 2011

Definitely use the spacer; int keeps the distance between the bearings in check when you tension the yoke stem.

DogT · Mar 1, 2011

That was my trouble, I didn't have a spacer and didn't want to make one. So I re-used the annular thrust bearings I had.

Dave
69S

grandpaul · Mar 1, 2011

Well, the neck has shoulders, but I would think they may never fully seat well under stress.

Tintin · Mar 1, 2011

DogT said:
Or are you saying it is fine to put standard ball bearings in the head without the spacer?

That is probably the worst possible setup as this inevitably means static pre-load on ball bearings or a very lose nut. The reason for the spacer is that this way you can bolt up the bottom nut against the bottom yoke with reasonable torque and a properly tensioned-up chain of contact surfaces. The bottom yoke pushes against the inner race of the lower bearing pushes against the spacer pushes against the inner race of the upper pairing pushes against the top yoke. The outer race of the bottom ball bearing then pushes against the shoulder of the steering neck whereas the upper bearing is free axially. Classic fixed-bearing-loose-bearing setup - the only problem I see is that a properly set-up paired angular contact bearing (roller or ball doesn't really matter as long as the tightening torque is partially decoupled from pre-load by the spacer) is just better. Off course if a ball bearing set-up works one can argue that it is fixing a non-existing problem but to me this doesn't mean there ain't no room for improvement. Apart from the higher load bearing capacity it also means a wider support for the steering head forces (whether that is needed or beneficial on a Commando frame is another discussion but it is certainly worth the effort on a Featherbed).

Tim

nortriubuell · Mar 1, 2011

Well, I'm with "Tintin' on this. Great to hear he talked with "ZFD" (who I highly respect) and added to this post. I've since decided to stick with my tapered rollers, (like my Triumphs, CB750, and XS650 have) ... and I agree with Tim's points completely. But, I would like to know what dimension distance tube to put between my tapered rollers, as I currently don't have a distance piece intalled. ( see the Norton 750 in my avatar) Then again, my Triumphs, the CB750 and the XS650 don't have any distance tubes either ?!?!?! Perhaps this subject (tapered rollers vs. the stock setup) can be almost like ... dare I write it ... an "oil thread" ??? :shock:

Tintin · Mar 2, 2011

nortriubuell said:
But, I would like to know what dimension distance tube to put between my tapered rollers, as I currently don't have a distance piece intalled.

Take a look on what SKF says about this subject. Regarding the actual length and given that you only have to do this ones I'd go for this apporach (which I'm doing with my Atlas at the moment): Measure the distance of the outer races as installed in the steering neck. It doesn't really matter which surface you use for this measurement as the tolerances on the races are pretty tight already. The bearing dimensions are available on the manufacturer's homepage usually so you can derive the distance of the inner races from that. I'll start with an aluminium tube just a few 1/100 mm longer than that, install it without torquing it up and measure the axial play. Then grind the tube down by that dimension minus one or two 1/100mm so that without torque on the nut there will be very little play. I'll then torque up the nut to reasonable level (remember I'm talking about the old Atlas-Dommie-early Commando setup in my case) and see what the play is. Probably this process has to be repeated one or two times or maybe it makes sense to use shims for this but then sourcing these comes into play.

An alternative would be a more theoretical approach calculating the axial force on the yoke from the torque and the "spring" force of the aluminium tube but given the accuracy of the seats in the steering neck this is probably useless and all the calculation methods and recommendations usually apply to rotating assemblies anyway.

( see the Norton 750 in my avatar) Then again, my Triumphs, the CB750 and the XS650 don't have any distance tubes either ?!?!?!

I don't know the setup there but the less laborious (but also somewhat inferior) solution is to set the pre-load by the nut and then use a counter nut to get the tension up to the level need for fixation. This works too but the danger of setting the static pre-load too high is qute big there and I'd assume that most of the f*cked-up tapered roller bearings in steering necks come from exactly this problem.

Tim

bpatton · Mar 2, 2011

I like the stock sealed ball bearings. They seem to last about as long as a fork rebuild, so you can replace them while you're at it. And they are cheap and don't need to be adjusted. One problem specific to Commandos is that the headstock is not perfectly round. It gets warped when it's welded up to the frame. I've checked a couple frames and they had it. A guy I knew that straightened frames for a living said all Commandos have it. Both of my frames were about 5-7thou oval. The cups in a tapered roller bearing are thin and distort to the shape of the headstock. The outer race on a ball bearing is heavier and may have a better chance at keeping it's shape. The frame guy said it's no good to put the tapered bearings in without fixing the headstock and the benefit just isn't there when you do.

cash · Mar 2, 2011

Tintin,

Standard ball bearings are not intended to take axial load above a certain level,

Do these bearings really take a full/high axial load? Personally I don't know of any wear failure of the standard bearing set up, can they be highly stressed?

Cash

Tintin · Mar 2, 2011

cash said:
Do these bearings really take a full/high axial load?

Yes, they take a high load. They take most of the vertical component (the cosinus part of that) and a bit of the longitudinal force (sinus there). As the steering neck is pretty much parallel to the fork tubes they take a similar amount of load as the one which compresses the fork legs. It is a substantial load and the Norton setup certainly doesn't follow std engineering routines.

The problem with axial load is that is is usually derived as a function of the radial load bearing capacity. SKF states 50% of the static radial load bearing no. for "purely axial load" meaning no additional radial load. There is a rather high radial load on steering neck bearings, especially under braking - which coincidentally is also the situation under which the axial load becomes pretty high due to dynamic wheel load shift. Show the Norton setup to any engineer who has to deal with roller bearings and IMHO they will probably all say the same. The one real argument that speaks in favour of the original Commando setup is the fact that it works - but it does this IMHO only because the abuse isn't high enough.

Regarding the imperfections on a Commando head stock: Sorry, but the argument that a ball bearing outer race is stiffer than the one of tapered roller bearing and therefore less prone to deformation sound pretty much as creating an argument for the sake of an argument to me. a) just judging by the geometry of the races and the load carrying capacities I strongly doubt the relevance (if not the truth) of this and b) I love to see the measurements on an installed inner race in terms of roudness, not the headstock. If it is deformed by welding (which it undoubtedly is) guess what happens when you press in a pretty sturdy and dimensionally far more "correct" steel ring ....

Tim

kraakevik · Mar 2, 2011

On a related topic, does anyone have advice on freeing the steering stem from the upper yoke? I've been dosing the joint with PB Blaster over the last few days, from the top down--many thanks.

Tim Kraakevik
kraakevik@voyager.net
Two Commandos

Cheesy · Mar 2, 2011

The fundamental design of any head (or swingarm) bearing arrangement will always have the potential to be problematic. Rolling element bearing load-life calculations are all based on complete revolutions of the races, in small angle movements the rolling elements continuously run on the same part of the race and will die from sub surface fatigue a lot faster than predicted by any of the bearing life calculations.... That said the deep groove balls seem to work and are cheap. When mine goes back together it will probably have taper rollers in it.