Why is my T35 crank twisting

[aggettd]

It's happened again. For the fourth time this season, the crank in my T35 has twisted and seized. I was certain that my bodge (thanks Doug) of welding all the components into one rigid mass would prevent the crank from pretzeling, but no.

Yesterday I got the bike reassembled for the 4th time this summer. It started third kick, ran fine except for a knocking sound at low revs. Is this normal? I balanced the carbs as best I could, took it for a short ride and put it away for the evening. Today I fired it up again, second kick, let her warm up while I fetched my helmet, and went for another short ride. After about a kilometre I had slowed to negotiate the end of a cul-de-sac and she just lost power. I pulled the clutch and she stalled. When I tried to kickstart again, the motor was solid. I put her in third gear, rocked back and the motor just turned over. She will just barely turn with the kickstart lever now.

What could possibly be happening inside my crankcases to make the crank twist like that even after I have welded all the crankpins? The welds were good, with probably 3/32 inch penetration. (I ground out one section to check before rewelding just to check)

Fortunately I have another crank to refurbish (many thanks to Trevor Pickett), but don't see the point of ruining a good crank if I haven't divined the cause before installing it.

What should be the interference fit of the crankpins to the web and counterweights? Should I chrome the pins to increase their diameter, bore the web and counterweights, and turn the pins to size? How about stelliteing (or other process) the pins, rather than chrome? Would I be better off building a crank from scratch from modern materials?

Surely my experiences aren't unique. Any suggestions?

David Aggett
Nova Scotia
Canada

[Doug]

David,

No, there should not be a knocking sound at low rpm.  Use a long screwdriver as a stethesscope and try to isolate where it is coming from.  If it stops when you pull in the clutch you likely have an end-float issue with the crankshaft. 

It sounds like the holes in the crank throws and center web have stretched and you are not achieving the requisite press fit.  Douglas honed the holes out +0.004" when they factory refurbished crankshafts and fitted oversized pins.  An old fella that refurbished a number of postwar Dougie crankshafts for both the Mark series and the Dragonfly said he used a 0.0025" interference fit.  Quite a lot for that diameter and needing some skill and a bit of luck to press up without damage.  And right at the limit, and not surprising it would stretch after being pressed together and apart. 

Tack-welding the crankpins to the throws is a bodge, but it is surprising they did not stop the crank from slipping under any normal circumstances.  I assume the weld cracked; through the center or alongside the weld?  Also a thought, do you have matched crankcase halves?  Douglas was not known for interchangeability of case halves, so you could have a bearing missalignment that is causing the crankshaft to work and flex, thought I would expect the front bronze main bearing bush to first complain by seizing. 

If you chrome plate the crankpins, you will need to take care not to plate the area where the rollers run.  The plating will not hold up under the pressure of the rollers.  Also chromium plate has a very low coefficient of friction, so you will have that working against you.  Also you will need to bake the parts immediately after plating to prevent hydrogen embrittlement. 

I would make new pins from SAE 9310 alloy steel case hardened 0.040-0.050" deep.  You can do like Douglas an make these oversize to reclaim the used crank by honing the holes in the throws and center web.  Do not forget to hone the cages accordingly.  With a 0.0025" interference, the finish in the holes and on the pin has to be very good, 4 micro inch or better, else you may have trouble with galling. 

-Doug

[aggettd]

Doug:

Many thanks for the detailed (as usual) reply. When I have some time during the winter I'll see of I can do a proper refurbishment of a spare crank I now have. In the meantime I thought I'd give a short post mortem on the twisted crankshaft. You were right, the weld cracked very clearly right up the middle. On closer examination, I also discovered that, although the fellow who last straightened the crank got the timing end and drive end of the crank perfectly concentric, the centre web was out by 2 or more millimetres. I had not thought of inspecting that feature before reinstalling it. I assumed he knew what he was doing.

Some quick calculations of the force imposed by 2mm  of eccentricity assuming 4000 rpm and 2 kg of mass (including the web, journals, and big ends, I know perhaps an exaggeration) I still came up with a force of 71.8 kg. Coupled with the fact that there is no central bearing in those cranks, means the crank was probably whipping significantly, resulting in knocking (although I could see no marks inside the crankcase), perhaps some pinching in the front crank bushing, and voila! instant rebuild!!

After grinding out the welds, I tried my hand at truing the crank myself. Piece of cake. Holding the crank in my hand, and using a 5/8 bar for reference through the holes in the counterweight plates, I applied a brass hammer to the offending plate and brought it into line. Repeat procedure on other end. I then put it on my vee blocks and found the crank wobbled 5 thou on the timing end and only 2 thou on the drive end. That suggested that the twist was now located on the timing end. Three more wacks, interspersed with visits to the vee blocks, and the wobble was down to less than 1/2 thou. Good enough for government work, or should I go closer?

To case harden 9310, do you need to use Kasenite, or would an oil quench (or three) be enough? What do people use for the plugs in the end of the crank journals? What should the interference fit be for these? How do you get them out, press them right through, drill them out? Given the hardness of the journals, would it make sense to use a slightly tapered and oversized plug to try to lock the journals in the counterweight plates, by expansion, or would they be too brittle?

Do you know of any reference books which would give specs like interference fits for cranks, bushings etc?

Thanks again for your insightful reply.

David A

[Doug]

...or should I go closer?

Two dozen whacks with the brass hammer and you got it within half a thou, you best leave it well alone and consider yourself lucky.  If fact you ought to consider hanging up a shingle for crankshaft repair!  Half a thou is more than good enough.  One thou would be good and one and a half would be the limit where I would worry about trying for better. 

You should probably consider re-doing the tack weld.  The holes are probably so stretched by now that they lack the requisite interference fit to hold the crank true by themselves.  Of course, running out of true as it was, does make it more prevalent that it would twist again.  I doubt it was due strictly to imbalance that should have just made it vibrate like a darn single!  I found a stainless steel welding wire often works best when welding high carbon steels.  It is still a brittle weld, but seems less prone to cracking than a mild steel 6013 or even a low hydrogen wire like 7018.  Disclaimer: we are not talking perfect welding practices here; it is still after all a bodge repair.  The fact that it cracked up the center of the weld is encouraging in a way.  Using a more ductile filler material might do the trick.  If it cracked at the weld/parent metal interface that would have shown a problem getting the weld to 'hang on' to the high carbon alloy.  Do not v-notch it out too much; you do not want a large, wide puddle of weld that will contract excessively on cooling.  It will just create stress in the weld, right up the center, and there it will crack again. 

The knocking sound was probably the rods working back and forth or hunting, if the center section was out of true by as much as you say.  Either at the crank end or the piston end, such an eccentricity causes the rods to flex mightily as they try to maintain alignment with the bores via the pistons.  Something has to give.  Look for signs of rubbing/heat on the sides of the rod big end eye and the crank throws.  Also look for bright spots on the piston/bores to the fore and aft orientation, indicating the pistons were being force to run slightly out of kilter.  Also see if you have not developed more play in the big end or wristpin bearings than you recall previously.  If the crank were apart I would check to see if the rods are still straight. 

You are not going to achieve 0.040-0.050" case depth on 9310 alloy with Kasinite.  Send it out to a heat treatment firm that knows how to carburize, quench and temper 9310.  It is a little more fickle than your average 8620 most places are use to dealing with.  I also have 9310 dry cryro tempered to -300 (there is only one outfit in the USA I trust for this.)  In a pinch you can use a liquid nitrogen temper, which most heat treat firms have available.   

I have seen aluminum and steel crankpin plugs.  Also there seems to have been several different designs of plugs used, in the way the oil circuit was arranged.  Some are two separate short plugs, others go all the way through end to end with grooves for the oil, and some have the center reduced in diameter.  You will not know which type you have till you go to knock them out with a drift.  On a used crank this will extrude sludge and wear debris trapped in the crankpin, out into the big end bearing.  I am told some Villers cranks use a heavy press fit plug of hardened alloy to expand the end of the crankpin.  But given the wall thickness of the Douglas crankpin it is highly unlikely you will be able to get enough of an interference fit to actually expand the pin a measurable amount.  So I would not bother.  I would use steel, nominally hardened, if only because it is less prone to galling when pressing in.  Aluminum plugs are very prone to this; especially if the short, tablet-like, variety.  The surface gets shaved, leaving the fit loose, and they subsequently fall out with loss of oiling and likely get jammed betwixt rod and crank on the way to the sump. 

-Doug

[aggettd]

Doug:
Again thanks for the prompt and comprehensive reply. Alas, dumb luck with a brass hammer is not a particularly firm foundation on which to build a career. I think I'll stick with what I know.

I wish I had thought to use the stainless rod for the tacks, as I have a small box of 312 just sitting there. I've kept the v-notch within reasonable size using 6014. The idea is that with sufficient size (not too big), there will be enough give in the weld to stretch a bit rather than crack. We'll see. Fortunately I have a spare unit to refurbish correctly this winter. Now I have just a few weeks to work out any other bugs in the bike.

Upon reassembly, the engine runs beautifully. For now.

The misalignment that caused the crank to be out of balance wasn't at either end, it was in the centre web. Both crank journals had rotated so that although both ends of the shaft were concentric, the centre web as not. If you held the crank so that the ends of the crank and both journals were in one plane, the centre web was dropped down by about 2mm. It never should have been so if the mechanic had used the 5/8 alignment bars to set it up. Live and learn.

If I understand correctly, the journal plugs are a loose enough fit to extract with a drift. I assume this means an interference fit of something less than 1 thou. I'll probably use a press anyway to see if I can get some meaningful measurements off them.

Thanks again,

David A.

[Doug]

David,

Yes a press would be a little more refined than using a drift to ease out the crankpin plugs. 

If the center web was out of alignment, at one or both crankpins, the main shafts had to be out of concentricity.  They might still have been parallel, but they could not be co-axial.  Think about it, if you held the center web stationary and rotated both throws about the crankpins till they were at 90 degrees to the plane of the crank, the main shafts would swing up and away from each other in the path of an arc.  The only time they would converge is when the throws ‘point’ towards each other.  Even if one was twisted up and the other down, it would still be the same situation.  Only if the stroke on the center web and the throws were un-even (and the center web would have to be the shorter of the two) could you have a situation where center web would be off center and the main shafts co-axial.  Or, both main shafts are bent where the join the throws… gulp!

-Doug