I'd like to probe the accumulated knowledge bank out there.

recently picked up a late model box (has the cast lugs for paralever design, not drilled) from a guy who said he had personally broken the bike for parts - i.e. gearbox was fine.

I put it in one of my 75 C bikes to test it out (stupid - I now realize, but I've been spoiled by the performance of the $220 box I bought from Rick G). Anyway, it went through all the gears beautifully on the first test ride but problem emerged when I started giving it a bit more 'stick'. Now I'm told this is typical of a bike that has been red lined in first before a change into second.

Question - why is it only 2nd that is prone to being damaged in this way?

Do BMW acknowledge this design flaw and have they/how have they, got around the problem in later designs (if indeed they have)?

How do equivalent Jap bikes avoid the same problem?

When I stripped the box down, I found wear on the thrust face of the appropriate selector fork (although the fork didn't appear to be bent or deformed). Problem obviously comes from loss of the wedge shape of the dogs that tend to hold the corresponding drive faces 'in mesh'. The resulting 'unmeshing' force that was wearing the selector fork was also translated into a dimple/deformation on the contact point between selector fork roller and selector barrel groove. The accumulative effect of wear points in the selector train was obviously sufficient to allow the dogs to temporarily unmesh under load and immediately remesh under load removal. A very disconcerting effect!
I think I begin to understand why a guru I consulted, insisted that when one encounters this problem, a fix requires that the selector drum, fork and both the rotating elements affected by worn dogs, need to be changed out.

I believe it is that the box has been treated to some pussyfooting and has come back out of 2nd on several ocasions which is where the problem starts. The speed difference between 1st and 2nd is the biggest difference you get in a constant mesh gearbox so in a clunky box like these getrags it will be the one that suffers most.
Has it been jumping back out of 2nd or have you been having trouble with the 1st - 2nd shift being worse than usual.
The gears are flat faced gears and as I found by backcutting the faces it can make downshifting very difficult on any gearbox.
I found that if I missed a shift that if I left the gearbox alone I would continue to mis that shift so I would replace the drive and driven gears and the selector fork weather or not I could see damage. on the Getrag boxes you would also need to replace the drum as the hard anodizing would probably get chipped.
Margs K75 used to come back out of 1st when she first started to ride it and the gears, fork and drum were replaced despite no obvious damage and all was good.  I had it apart several times before I bit the bullet and bought new parts.

Cheers Rick

I had noticed that 1st to 2nd is generally the most clunky on the five Ks that I have had running on and off. Your explanation in regards to the ratios makes sense. Also interesting that you think that incomplete engagement can trigger this condition.

Shifts between 1st and 2nd felt fine if done at low revs - the problem came on opening the throttle in 2nd gear. Initially I thought I had bumped the gear change lever by mistake with my foot and that it had changed up a gear (gear position indicator not working on this bike). I later verified that the 'skip' involved no gear change - just a horrible momentary disengagement followed immediately by re-engagement....

Also interested to hear you talk of 'back cutting' the dogs - with a file I guess? Does this involve re-establishing the original plane-ness and relative angle between faces of each dog resulting in increased backlash of the meshing?

I used a small tungsten burr to do the cutting and usually took a full week of machining to do a complete box but the mob I worked for were good sponsors and donated their time.
Mind the gearbox was very different to the way Kawasaki made them.

I think Ben Kingham amongst others has experienced this also . it might have something to do with his racing I guess ...but the jist of it was that the dogs didn't all engage properly on second gear ...I believe he also said something about it being more prevalent - likely to happen on his higher spec engines.

another racer member confirmed this a while ago (with his detailed exploration into salt flat speed attempts , and higher power, gearbox mods etc  )

apparently the fix as you guessed is to reface the dogs ...I recon that would be worthwhile but fiddly thing to do ..unless as rick says, you have some experience in such matters ...

would love to do this also whilst changing the bearings in the gearbox ...planned for not to far off ...but delayed by the new car rebuild

I would imagine that confirming or reapplying case hardening on the appropriate parts would need to happen also after the modification ?..not sure exactly of that process details for that job ...something about heating to a certain temperature and oil in a cooling bath ?

good luck Floyd

Good point about the likelihood of case hardening Charlie. That would explain Rick using his tungsten burr as opposed to a file. Redoing the case hardening would involve getting carbon back into the molecular structure which if I remember correctly, requires the metal to be held at a certain temperature for a particular length of time with the surfaces in contact with the right grade/particle size of carbon powder. Next step would involve heat treatment of the high carbon content layer - all way beyond the private individual after a result as opposed to playing around. Better just to collect dead boxes that have come to grief by a range of different means! Then cannibalize as appropriate...