bikemonkey
I live for the CABE
If you read deeper into Brandt's second link I posted above, you can see where he states the slippage builds up from the canting of the pinion pins and axle deformation. That is probably why yours slips intermittently as the slippage continually increases under torque for x number of revolutions.
It is an inherently poor design that SA refused to address whilst blaming all the injured riders for improper cable adjustment.
...Two features that cause disengagement occur only under continuous hard
torque. The four pinion pins, that fit loosely in the planet cage,
cant slightly under pressure from the clutch (driver cross), to an
off-perpendicular angle so that the bearing surface with the clutch
slopes toward disengagement.
Meanwhile, chain loads on the hub cause
the axle to bend slightly so that the active parts (planet cage and
clutch) displaced along the axle, rotate about separate skewed axes to
each other. During rotation, the four bearing faces of the clutch
each experience reciprocating motion of the planet pins that, through
their slant, generate disengaging creep. The engagement spring is no
match for these forces.
In addition to the disengagement forces, caused by pin skew and axle
flex, the clutch faces develop indentations from the slanted pins that
enhance disengagement. This is clearly visible on any used clutch.
This problem could have been resolved by putting a slight inward taper
to the ends of the planet pins and a similar matching slant on the
bearing faces of the clutch, giving their engagement a preferential
retaining force instead of the opposite. Most motorcycle gear boxes
use such features, especially in older non-synchronized sliding gear
boxes... the classic clunk of BMW boxes for instance.
It is an inherently poor design that SA refused to address whilst blaming all the injured riders for improper cable adjustment.
...Two features that cause disengagement occur only under continuous hard
torque. The four pinion pins, that fit loosely in the planet cage,
cant slightly under pressure from the clutch (driver cross), to an
off-perpendicular angle so that the bearing surface with the clutch
slopes toward disengagement.
Meanwhile, chain loads on the hub cause
the axle to bend slightly so that the active parts (planet cage and
clutch) displaced along the axle, rotate about separate skewed axes to
each other. During rotation, the four bearing faces of the clutch
each experience reciprocating motion of the planet pins that, through
their slant, generate disengaging creep. The engagement spring is no
match for these forces.
In addition to the disengagement forces, caused by pin skew and axle
flex, the clutch faces develop indentations from the slanted pins that
enhance disengagement. This is clearly visible on any used clutch.
This problem could have been resolved by putting a slight inward taper
to the ends of the planet pins and a similar matching slant on the
bearing faces of the clutch, giving their engagement a preferential
retaining force instead of the opposite. Most motorcycle gear boxes
use such features, especially in older non-synchronized sliding gear
boxes... the classic clunk of BMW boxes for instance.