Main bearing failure

From some previous Norton twin custom billet crankshaft work I had done I found that the main journals are ISO Fit Class m5

http://www.engineersedge.com/calculators/mechanical-tolerances/table_shaft_tolerances.htm

So for a nominal bearing ID of 30mm, the shaft should be:
Max diameter - 30.0170 mm
Min diameter - 30.0080 mm

It may be too late to check your journals now that you have installed the inner races but I recall that there is some journal that stands proud of the inner race so maybe you can get a measure.

Any chance that burrs, nicks or gouges on the main shaft were giving you the difficulty? It does not take much to make a tight fit diifcult. Any chance the main shaft(s) are deformed and not supporting the inner race properly; this could explain a premature bearing failure though its is a real stretch of the imagination as to how the shaft(s) would get deformed. Were both bearing inner races difficult to install?

jug said:

Jim, these were not snug. These were ridiculously tight. I even contacted the supplier to see if they had had any other customers have the same problem, if they did they weren't admitting it.

Click to expand...

It's not hard to remove a bit of material from the ID of the race if needed. I have even done it with a wheel cylinder hone.

Just be careful. I have not seen problems come from a bearing that was too tight, but I have seen quite a few cranks damaged when the drive side bearing didn't have enough interference and moved in use.
If the drive side race can move it will eat the crankshaft pretty quickly. Jim

Service Release No. N2/9
Nature of Release Crankshaft main (roller) bearings
Models Affected 1972 Commando (all models)
Distribution Worldwide


Explanation
In order to extend main bearing life, a change of roller main bearing specification has now been authorized and fitted on production engines from Engine No. 211891. The new roller bearing is designated 'Superblend' with an increased load carrying capacity and is supplied under Part No. 063906 (manufacturers part number R&M 6/MRJA30).

Action:- Where engine units are stripped for examination, or the need for main bearing replacement, fit only roller main bearings Part No. 063906 in place of the previously specified 063114. The new bearings are fully interchangeable, and should be fitted in pairs for maximum reliability.

Note:- The new bearings may also be fitted to advantage to 1971 and earlier Commando machines, provided care is taken fitting the inner spool to the right side crankshaft journal (was 1.1807" - 1.1812") diameter, now 1.1812" - 1.1815" diameter. Also crankshaft fitted end float should be checked, and shimmed where necessary to provide 0.010" minimum, 0.020" maximum end float using shim Part No. NMT2196A as necessary.

August 1972

Click to expand...

(30.00248 -30.0101mm)

comnoz said:

Just be careful. I have not seen problems come from a bearing that was too tight, but I have seen quite a few cranks damaged when the drive side bearing didn't have enough interference and moved in use.
If the drive side race can move it will eat the crankshaft pretty quickly. Jim

Click to expand...

+1 What Jim says. Tight is right.

This happened to me with a billet crankshaft where one of the main journals was a tad out of spec (too loose an interference fit). Only one journal gave me trouble where the inner race burnished the main journal to a very nice slip fit within a very short amount of time. It sounded like a main bearing failure. Not easy for a permanent fix; in my case I bought some time with Loctite (green) bearing compound.

Again, tight is right.

Tight.

Someone on this list fairly recently mentioned that their newly fitted bearings didn't spin well - too tight.
Don't recall if they had the C3 spec though...

Never heard nor seen it happen. Not the issue with regards to the OP.

Dances with Shrapnel said:

Never heard nor seen it happen. Not the issue with regards to the OP.

Click to expand...

Doesn't matter whether you've seen it or not, its happened to someone here.
(although don't recall discussion whether the bearing was C3 spec - which allows for slightly tighter fitting).

How do we know its not a problem with the OP - if even one spot was slightly tight,
it could preferentially wear that spot, with that result (?).
20/20 hindsight AND x-ray vision needed to rule that out...

Yes, you may be right...someone.

Rohan wrote;

although don't recall discussion whether the bearing was C3 spec

Click to expand...

Yes it was C3 spec.

Say Reggie, do you recall if there was any hint of rolling resistance, binding or roughness when you installed the C3 bearing? What about the most recent install where you mentioned the inner race(s) went on tight?

Was there any hint of snagging or binding of rollers when you split or assembled the cases with either set of bearings?

Less likelyhood of binding with C3 as it has more internal clearance than a CN bearing (which is the size Norvil will send), it is only a few microns difference but enough to reduce the problem of skating rollers on a cold engine which will cause premature bearing failure due to the oil film being breached. Even a C3 will bind if the inner ID is too small.

kommando said:

Less likelyhood of binding with C3 as it has more internal clearance than a CN bearing".............Even a C3 will bind if the inner ID is too small.

Click to expand...

Which gets back to the question of whether the OP recalls binding, tightness or rough rolling of the newly installed C3 bearings or their replacements. I know it may be subjective but the observations could hint to possible problems I had outlined earlier.

Aside from my experience, others have experienced the problems that arise when the inner race fit to the main shaft is not tight enough.

One or two weeks ago a friend raided my loft and took away very probably my last new old stock Ransom and Marles 6MRJA30 'superblended' Commando main bearingleaving me with a few new old stock RHP ones.... I have just phoned asking if he had fitted it and he had so I cannot retrieve it to measure. He stated it fitted perfectly giving a very slightly tight rotating crank but that with a touch of heat applied to the cases it was perfect and exactly as he had built engines for around 40 years for himself and friends which have never had main bearing problems. He also stated that the crankcase housing and main shaft measured standard size. he has a special plug guage for cheking the cases..... He believes that as the cases heat up and expand as does the bearing outer raceway there is enough internal clearance for the oil to lubricate the bearing correctly but NOT enough to allow the crank to bounce up and down resulting in vibes and having ridden one of his 650ss lumps they vibrate a LOT less than any other 88/99/650 /Atlas lumps I have ridden but he does go to great lengths to assemble the crank correctly and balance the rods and pistons etc etc......
Using my very olde memory banks yet again.....many decades ago when a Gentleman at the old Ransom and Marles factory in Newark went down into the cellers and came back up with the 6MRJA30 drawings I am certain the bearing was shown to be a C2 fit bearing but that when I used my jig to check one it gave an internal clearance of 0.0012 inch making it a low CN fit bearing and at the time I did wonder qabout quality control......
Wonder if those drawings still exist?? If it were a British company that took over they would of been burnt long ago in a skip just like so many BSA drawings were.......except those some people saved that is.....
Just in case anyone is actually interested a very old (1973) Hoffman bearing book of mine shows the internal clearances of roller bearings haveing inner diameters of 24 to 30 mm as being.
Values first in 0.001 mm (and then in 0.0001 inch for our U.S.A. friends and old Brits like me who think still in inches.....)
C2 10 - 25 ................4 - 10
CN 25 - 35 ...............10 - 14
C3 40 - 50 ...............16 - 20
C4 50 - 60 ...............20 - 24
These are as printed in the book. I have marked the pages with slightly different values for some but decided to give the printed word values only. I must of queried the printed values at some time and received correct values but I cannot now remember what occured all those decades ago.......Anyway some over paid useless non productive European official has probably changed the system again since then......
The FAG bearings used on Commandos were and are a bog standard production bearing..simply a high load capacity 306 bearing with a brass cage with rolling elements that are crowned......as I am sure the Fag bearing designers will tell you if you take the trouble to contact them and ask......

I am still waiting to read the results of the original failed bearing being returned to the manufacturer for examination..... or have I missed it???

I came across this interesting line in the FAG book " The modified line contact between rollers and raceways of cylindrical rolling bearings eliminates edge stressing and ensures a certain aligning capability" so maybe there is something to all this after all, just remember we are talking microns here so any measurement may be rather difficult without very specialised equipment. Also it seems that the SKF bearings have a higher static and dynamic load than the FAG ones. The internal clearances are the same between SKF and FAG (although different to what is posted above) and the angular miss alignment is also the same.

Can you quote which FAG booklet that is, or scan the cover and relevant section.
Is there a year printed there anywhere.

Over on the NOC Forum, they are saying we are measuring worn FAG bearings, hence are not seeing this magical property. !!!
That doesn't explain how the bearings go on to have a long and happy long life though, if the magic bit has rubbed off....

On the pic of the R&M Superblend that was posted on the NOC a while back, you could actually see a tapered area near the end of the roller.
Must have been expensive to manufacture it like that...

Back when Nortons switched to FAGs, FAG said at the time this was a stock off the shelf heavy duty bearing,
nothing special about it for use in Nortons... (?).

But we diverge, slightly.

I think if the property is there to any extent you may not have the resolution to measure it, I would be trying a Talysurf or similar. Although it is all rather academic it is kind of interesting. The other thing to think about is that it is impossible to make a roller that is perfectly cylindrical with no taper or convexity/concavity but I doubt we will ever see the manufactures tolerances for the rollers, I would not be surprised if it called for parallel to convex .

It may be that the modified line contact I mentioned is from the radius to parallel section that Jim talked about. The book is FAG Rolling bearings WL41 520/2 EA December 1996

How slight is this magic supposed to be, if you can't see or measure it, but it makes a difference ?!
The slight taper on the end of the R&M bearings on the pic on the NOC looked to have a slightly etched look to them for a few mm on the very ends.
Much more so than the rounded looking ends of Jims FAG.

The C3 clearance well may give some (extra) scope for a slight 'self alignment' over a stock tighter bearing. (?).

Cheesy said:

The other thing to think about is that it is impossible to make a roller that is perfectly cylindrical with no taper or convexity/concavity but I doubt we will ever see the manufactures tolerances for the rollers,

Click to expand...

Likewise, perfectly circular ball bearings are probably difficult to achieve.
But supposedly its the improved quality that has allowed some japanese makers to offer an unconditional 100,000 km warranty on their sealed automotive bearings, and thus on cars. Toyo Kogyo (Mazda bearing division) first did this a while back now, and it doesn't seem to have come back to bite them...

Rohan said:

How slight is this magic supposed to be, if you can't see or measure it, but it makes a difference ?!

Click to expand...

It is not magic. It is material science and engineering. You are dealing with mitigating stress concentrations and microstrains.

Rohan said:

How slight is this magic supposed to be, if you can't see or measure it, but it makes a difference ?!
The slight taper on the end of the R&M bearings on the pic on the NOC looked to have a slightly etched look to them for a few mm on the very ends.
Much more so than the rounded looking ends of Jims FAG.

The C3 clearance well may give some (extra) scope for a slight 'self alignment' over a stock tighter bearing. (?).

Cheesy said:

The other thing to think about is that it is impossible to make a roller that is perfectly cylindrical with no taper or convexity/concavity but I doubt we will ever see the manufactures tolerances for the rollers,

Click to expand...

Likewise, perfectly circular ball bearings are probably difficult to achieve.
But supposedly its the improved quality that has allowed some japanese makers to offer an unconditional 100,000 km warranty on their sealed automotive bearings, and thus on cars. Toyo Kogyo (Mazda bearing division) first did this a while back now, and it doesn't seem to have come back to bite them...

Click to expand...

That 100K warranty sounds a bit off, do you have a link to it? How does the bearing manufacturer know what the design loads are to be able to give a warranty on it? That said there is no reason that you could not design your machine to have a 100K bearing life, thats why the bearing books have the dynamic, static loads and thermal reference speed so you can calculate bearing life.

You may be able to have slightly more misalignment with a bigger clearance bearing but it would have to be lightly loaded. I would not me surprised if the Norton crank flex is actually outside the limits of the bearing misalignment (which is 4 minutes)

And then there are these......