Having finally stripped my 1938 16H basket case engine, I find I didn't realise it has two 'main' bearings on the drive side. Unless my memory has really failed me, I'm sure the 16H engine I built 30 years ago had just a single ball-race (thrust bearing) on the drive side.
This 1938 one has (had) a pair of ball-races on the drive side (separated by an outer race spacer) and a roller on the other side.
Generally I consider this layout to be bad news - radial alignment, thermal fights etc.
So why was it adopted? The only reason I can come up with is to combat crankshaft failures due to bending (fatigue). Was that it, or was a single thrust bearing of that time not up to the task?
Reminds me of a story I heard about early Austin A-series engines in the 30's - the original 2-bearing arrangement led to crankshaft failures in the competition cars. One cure was to use lower compression pistons in cylinders 2 and 3 (1 and 4 were 'supported' by the two mains). Desperate.
Anyone shed any light here?
The 3 bearing layout was used on other makes too. Before 1930 a Rudge did 200 miles in 2 hours.with this setup. The ball races did the locating and the roller took the drive loads.
.. 2 bearings were used to share the load on the drive side. A bit like the double row roller in the back wheel.
Just like Russ I've pondered over the wisdom of two bearings side by side, and the likelihood that one would lever against the other. I can only suppose that the crank cases are more flexible than the crankshaft, so bending of the sides of the cases ensures similar bearing loads.
.. the double bearing was adopted in 1934 along with the train of gears driving the inlet camshaft from the exhaust.
Thanks for the comments. Double roller bearings work well to reduce shaft bending, and share radial load in a small diameter like a wheel hub. For a location bearing, one thrust race back to back with a roller will work well, as it controls both axial and radial loads, so I reckon I'm going to build mine like that....after all these are big bearings for what, 15 BHP?
Yes the correct set-up is a lipped roller bearing nearest the flywheel and a fixed ball-race for the outer bearing. Your engine was probably re-built by an in-experienced mechanic or on a budget as ball bearings are half the cost of roller types. On the timing side the bearing on the pre 1948 bikes was often a self aligning ball bearing, mainly to overcome inaccurate machining of the crankcases. This was also the reason for using 3 spot clearance bearings. All post '47 engines were fitted with a lipped roller on the timing side and it is the shoulder on these that take the side thrust as the main-shafts are a relatively loose fit in the inner races.
I think the Austin engine you refer to is the 750 side-valve rather than the A series, and failure was due to crankshaft flex due to high revs, especially when supercharged.
Thanks Richard - confirms my 'back to basics' thinking.
On the Austin story, you're correct - yes, it was lack of crankshaft stiffness combined with (for those days) high cylinder pressures driving the crank bending fatigue failures.
So, to close this one out, I took the advice and put the thrust bearing nearest the drive side, and a pretty useless semi-roller next to it, without any outer-race spacer, because the one I found in the rebuild was about 8mm width and I needed nothing more than about 1 mm. I think the floating roller will have an easy life now.
For the historians, the WD repair manual specifies 3x ball races for the 16H main crank support assembly. This works is it also says "make sure both outer and inner races are free to slide". That's a poor design solution from my point of view.
The WD16H used the 1936 pattern engine and the Spare Parts lists, both civilian and military from 1934 to 1937 inclusive list three ball bearings (Part no.9033) for the SV and OHV engines (other than the 350s). There is a note in the factory records mentioning 'Hoffmann' bearings to be fitted to WD models. These always seem to have been of the self-aligning type.
From 1938 onwards, two rollers and one ball race were listed.
Whatever the theory, the main bearings on 1934 - 1947 engines don't seem to have been a problem or low-life area on road going machines.