So all roller bearings since WW2 had barrel shaped rollers did they?? Well you learn something new every day on this web site which appears to contain as much MISinformation as other Norton / British bike web sites and Club magazines etc!!
As far as I am aware the drive side roller bearing employed up to Commando models was a bog standard Ransome and Marles ( R & M from here on) 8MRJA30 bearing which had 11 rollers of approx width and O.D. 11.12 mm They had a 10 mm wide parallel section with a very slight radius at each end. In use on the more highly tuned Commando Combat motors the crank flexed so much that the edges of the rollers were the only bits in contact with the inner and outer raceways which resulted in bearing premature failure often at around 4000 miles DRAMATICALLY helped by the advance retard units falling apart and staying fully advanced at all times,,, their construction was later improved!! To help overcome this rather serious bearing problem a plain fag packet was obtained and someone suggested slightly altering the shape of the rollers such that when the crank flexed more of the roller than previously was in contact with the inner and outer raceways. R & M then produced a new bearing with a designation 6MRJA30 with the MRJA30 stamped onto the side of the outer raceway with the 6 etched on. The inner raceway was stamped MRJ30. The bearing employed 13 rollers of width and O.D. approx 9.53mm but the parallel portion of the roller was approx 6 mm with a slight taper on each end reducing the diameter to approx 9.46 mm and on the edge a slight radius as previously. I believe the 'barrel shape' is known as logarithmic profiling. This is the 'Superblend' bearing originally employed for Commandos.
Interestingly the load capacity of the earlier bearing was greater than that for the 'superblend' bearing....
8MRJA30 Dynamic 41,900 Newtons .....Static 35,200 Newtons
6MRJA30 Dynamic 35,800 Newtons .....Static 31,000 Newtons
Some of us actually talk to bearing manufacturers to learn the facts ........... and have removed the rollers from various new bearings and measured them 9with micrometer and OLD knackered eyes to try to find out what was actually going on. Well it kept me amused for a while......
Now another bearing company was FAG... who had an office just down the road from Norton.....and could, so I was told by several ex NVT Gentlemen friends, supply cheaper bearings....... FAG produced the roller 306 bearing in a std form and a high capacity load form and by that time had changed their roller shape (as did various other bearing manufacturers) This high load capacity FAG bearing had a designation of NJ306E the 'E' indicating EXTRA load capacity which were Dynamic 50,000 Newtons and Static 48,000 Newtons. This bearing employed 12 rollers of approx 12 mm width and 11 mm in diameter with an approx 8 mm parallel portion reducing at each end to approx 10.97mm. Funny thing is that every Fag bearing I have looked at taken from motors has wear markings on the rollers right out to the edge of the superblending whereas on the R & M bearings it stops just short of the edges......Perhaps its all down to the very slightly greater barrel shaping of the R & M rollers????
I believe Norton tried another manufacturers bearing at the same time as they did the R & M bearing because in a cellar in Birmingham there were a pile of SKF steel caged roller bearing that we were told had come out of Norton.....There were two types , one a std designated NJ306 and the other a high load calacity version designated NJ306E. All the boxes wer marked 1972. The rollers were all the older non superblended version.
The tale of the Commando main bearing failure problem was told by someone CLEARLY involved at the factory in Motor Cycle Sport January 1977 on pages 27 and 28 The heading on page 27 is ' VICIOUS CIRCLE... A cautionary tale in which it is shown that in Engineering, at least, nothing is as simple as it seems. The following story is based on a series of events that took place not long ago, at a motor cycle manufacturers somewhere in Europe. Only the names have been changed to protect the guilty.' On page 28 the heading is ' For life and reliability there is a threshold which cannot be crossed with impunity. Most bikes are there already'.
I would suggest you read it and I believe one of the many people I have over the years sent a copy actually published it on the web but I have not bothered to see if it is stil on the web.
A WARNING. Many years ago a certain British Bike spares emporium in London bought a VERY cheap pile of new brass caged boxed FAG NJ306E bearings ..... opened one box and removing the bearing inner I looked at the rollers expecting to see the slight barrel shape...there was non....I removed a roller and measured it proving my eye sight was still working......I phoned FAG who suggested they must of been a batch held by a dealer for a few years and that they had the old NON barrel shaped rollers. I assume everyone checks such things before fitting them ?? The dealer sold them all to Commando owners and I bet NON of them shecked.........
I even checked the RHP main bearing used on the D.S of later Triumph twins NUP306ETN which employed 13 rollers 12 mm wide and 11 mm O.D. with a 9 mm mm parallel portion reducing to 10.96 mm at the edges. Load capacity was Dynamic 57,000 Newtons and Static 53,000 Newtons.
All my measuring took place decades ago and the bearings could well of changed since then.