Some time ago I recall someone who replaced the rear roller bearing with a ball bearing to change thrust to the rear of the engine. Has anyone tried this?

Hi Kenneth,

I’m doing this mod, finding a replacement for the rear isn’t a problem, there’s a deep grove ball race that is a straight swap. The problem I had was finding a self-aligning roller for the front. In the end I asked a friend of mine, who is an excellent engineer, to make a sleeve for the crank and machine a flanged sleeve to go inside front bearing housing. Fingers crossed I’m not too far away from assembling the bottom end and will post a pic when it’s completed. The mod has caused a lot of delay and cost a lot more than I hoped it would. I have also spoken to quite a few people who abuse the hell out of their engines and can’t see the point of moving away from the original when they have failed to blow up an engine.

I have no idea how well the new bearing set up will work but I did run the idea past a crank designer who said he couldn’t see why it wouldn’t work, in fact he wondered why he’d never thought of it as he had built a 7 special, small world.

Tom

Location: Eye on the Norfolk / Suffolk border

Just a thought - if you have ball races at both ends then would there not be a problem with the differing expansion rates between aluminium crankcase and steel crank? I haven't calculated it but somewhere in the book "Design for Competition" it indicated that the crank would expand by 12 thou less than the crankcase.

Location: Wellington, NZ

Tom,
What's your theory re deep groove ball at the rear and SA roller at the front? I was thinking a self aligning roller at the rear would be capable of coping with a higher radial load than the deep groove ball. Bearing loading at the flywheel end is surely higher than the front? If there is no room for expansion it might cause problems,such as overloading the thrust faces of the rolling elements and inner races spinning on the crankshaft. You have probably thought about these possibilities and I am talking through my er... hat. Tony Press who is a contributor on here will have a better idea than me I think.
Good luck,
Dave.

Location: Sheffield

Sorry didn't read properly - I see you are proposing a self-aligning ROLLER for the front.

Location: Wellington, NZ

Hi Ivan and Dave,

The flanged steel sleeve that fits inside the front of the crankcase has been ground to a fit that will allow the SA bearing to move to cope with expansion. Things like specifying fits is reason why I asked my friend to make the parts.

The theory behind putting a deep grove ball race at the back is to limit the crank deflection as the original front bearing do. Putting an SA bearing at the back would allow the rear of the crank to do whatever it wants but it can’t because the gear box will be trying to stop it. So I thought constrain the back and take the thrust load, fit the SA bearing at the front and allow the front to do whatever it wants to, worst case I might end up with variable valve timing!

The crank designer I spoke to did some rough calculations for me said a ball race at the back was more than man enough to cope with the loading.

Tom

Location: Eye on the Norfolk / Suffolk border

I went down the "restraining the crank from the back" route after discussions with Frank Hernandez and wouldn't really consider any other way now.
Ball race at the rear, twin ball races at the front. The only complexity is that you need to put the rear bearing on the crank whilst it's held in a vice. Tighten the flywheel to a known repeatable torque (either with a massive torque wrench or a repeatable angular torque- it needs to be effing tight). You then accurately (to the thou) measure the gap between the back of the flywheel and the bearing (I used slip gauges) and make a spacer to fit with a thou or two clearance. Put it all together, retighten the flywheel to the same setting and bobs your uncle.
No front flange to fail.
No pressing the clutch and loading the crank still further when it's suffering maximum whip at 6,000 rpm.

Charles

I retained the rear bearing,making a spacer to the exact width as you have done Charles on 3 engines now.I know others using a belleville washer to do the same thing.
My advice came from a well respected and clever A7 engine builder who claims he never broke a crank in 20 years of racing in days when new cranks were not available.
Using self aligning double row ball races front and back to allow the crank to flex rather than trying to stop it.
On 2 engines I bored the front housing out and fitted a steel sleeve to allow the front bearing to float.
Again as you say Charles I can't imagine the stresses imposed on a crankshaft pressing down on the clutch as 6,000!

Location: Channel Islands

I've had an email query about "how tight is 'effing".

My use of the vernacular was simply because since I don't have a 3/4" drive commercial vehicle torque wrench or torque multiplier, I couldn't measure it.
I used a 3ft tube on a 1/2" wrench and heaved it as far as I absolutely could up to a repeatable position that I then marked on the flywheel. Then I measured the space, made the spacer, fitted it and heaved up to the same tippexed mark on the flywheel.
Measured no. Sophisticated, no. Repeatable, yes. Tight, absolutely.

A repeated response to a repeat topic.

Most cranks crack on the inside radii so perhaps should keep foot on clutch permanently! Has anyone noticed if revved cranks fail from the outside near centre, more as expected?

In view of modern ratings, and the bearings in motorcycles, the Seven crank bearings seem generous. In their day usually corroded long before traditional spalliing failure. Is the latter ever found? A ball race would seem very adequate. The load rating of expensive self aligning balls is less than a simple race.

All the texts advise against restrained races at both end of shafts.

Any deflection is only about gasket thickness. It is hard to imagine the Seven crankcase preventing crank deflections, even with preloaded ac races. With these, does the front flange routinely fail on revved or supercharged engines?

I suspect the inertia of the spinning flywheel restricts firing deflection more than the bearings do (which possibly accounts for frequent failure adjacent flywheel, where bending moment otherwise low).

If using a ball race at rear a tolerance with some tilt allowance as specified for semi floating axle shafts would seem appropriate. With a roller at the front and a (worn) roller or mt race as spacer.

The BSA car had a ball at rear with the taper fit flywheel not abutting directly. Cannot discern spacer arrangement. I vaguely recall a print of Seven based Jap car was similar. Anyone encountered these?

I suppose the very front bearing moderates hammering of timing gear and a ball race may dampen shortening and lengthening of the crank under firing?

Location: Auckland, NZ