The best approach for what you want to do is always some type of crank mechanism. A crank turns rotary motion into sinusoidal linear motion, is energy efficient because energy is recovered during return motion of the pedal, and allows for the highest reciprocation frequency with the lowest mechanical forces.
Compare a crank turning at 200 rpm to a cylinder doing 200 strokes per minute.

hello
i made that similar machine see the video now i do not know exactly how mutch forces i will apply to get the good results

Link

I mean this in the nicest possible way,
but
it's a stupid test.

I have to assume you were forced to do it, and waste a fair amount of money and time, by a clueless superior.


The exercise is stupid and futile because:

You don't have 100+ cables under test, to provide statistically significant results.

You have no idea what the test criteria might be, i.e. you don't know how to call a test a pass or a fail, or how many cycles under what load would be a minimum acceptable service life, nor how the life distribution of a population of cables would be distributed.

You have failed to simulate the clutch cable's actual environment as follows:

Your test rig does not angulate the cable ends against the bushings/nipples at the cable housing ends, as the clutch pedal lever and the throwout lever do.

Your clutch cable UUT does not duplicate the installed shape of cable in your own illustration, so the friction between cable and housing is not representative.

You are testing against what appears to be a very light return spring, not transmitting the force necessary to throw out a real clutch.

The velocity profile of the test stroke does not correspond to the way a real driver uses the clutch pedal; have you ever driven a car with a clutch?


There are probably more things wrong with your quest than I have outlined above, but basically, you are just jerking off. I hope you are being paid well for it.

Mike Halloran
Pembroke Pines, FL, USA