6mm is about 0.236" and 15/64 is 0.234375", both very close to the figure given by Rex. I know nothing about this clutch but assume that there was some end float in the original so that the odd thou and a bit wouldn't be enough to bind anything. Unless 1/4" (0.250") long rollers have been used and one or more washers have been left out?
Very hard to help Antoni without having a look - correctly set up there should be essentially no wobble. The answer is to measure everything (bearing inner and outer and the rollers with vernier and micrometer) and inspect the bearing tracks under a bright light. The diameter of the inner track + diameter of rollers should be only 0.001" or so less than the inside diameter of the outer track (in the basket). Something's badly amiss if it can wobble enough to hit the primary case!
Leon
Just had a good look over the rocking clutch problem with calipers and mic.
The previous owner replaced the basket during a (failed) gearbox repair and has admitted to being unhappy with the way the clutch went together. He thought it would 'sort itself out'.
I dismantled the clutch to get the gearbox out, not expecting to see any problems with the clutch, but did notice probably recent wear marks on the case cover. So before refitting everything the first time I placed the basket on my Goldring record deck, set 16 RPM and offered a bench grinder's wheel to the edge of the basket. Turning on the cheap. Re-assembled, the clutch functioned but rocked.
Getting the old basket out of the spares pile it looks like by chance I ground it back spot-on. Original 1.635", replacement 1.636".
The rollers' bearing surface on both baskets has a diameter of 1.873". The width of that bearing surface is different; 236th original, 239th replacement. The difference might be put down to wear on the original.
I can see now the reason for the rocking. Looking at the photo, the inner roller bearing surface's width on the 'mainshaft flange' controls the position of the driven bearing hub (pictured resting on the the mainshhaft flange). The hub looks clean and sensible. When the hub is placed over the splines it rests with 269th space left for the rollers and clutch basket flange. So regardless of quarter inch wide rollers or approx 6m ones, the clutch basket is free to float by 269th minus 239th = 30th. So it rocks.
Has the surface on which the roller sits in the picture worn badly? The flange thickness is 176th, very clean and flat but has 'Spirograph' wear marks.
Advice needed!
In the absence of said advice I will try to turn back the width of mainshaft clutch flange roller bearing surface on my small lathe, so that most of the 30th is taken up, and buy new correct rollers. I expect the material will be hard carbon steel but will sacrifice a turning tool or two to get the bike runing.
...runs under the nearest table and dons tin hat...
