Hi Jim,
Actually, the way I am controlling the H-bridge, I am "braking" the motor rather than allowing it to free-spin until friction stops it (maybe this is helping too). Certainly slowing it down should help you would think... but I am really trying to avoid using PWM unless I have to... causes excessive heat (MOSFETS switching) which I am also trying to eliminate. Actually, last night I added two (vice one) short 20ms reverse pulses just in case the first one does not work... Cycled at least 40-50 times last night with no binding... amazing. This weekend I hope to run more extensive tests... in an attempt damage the actuator if I can. I'm real curious what will fail first (any one want to take a guess? Maybe I should had made a poll as to what they think will breaks first... LOL). I plan to run it till something breaks, regardless! Should be fun! I wonder if Doug (Lado) performed this type of failure analysis/test before selling his gear?[8D]
As a side note, I am also messing around with a cool idea, using a optical reflective sensor that monitors the rotation of the shaft/coupler... this could also be used (along with current sensing) to determine the approximate position of the screw drive... but I think I might be over-engineering, something I tend to do all too often!
No, actually this makes sense (and the reason why I suggested the slow down just before the end). When you pause that 200ms with the motor power off, the motor is still spinning down and the gear train is still moving for another 200ms (which makes the it turn quite a bit). When you immediately back up the motor, you are preventing it from going too far by halting the movement and changing directions. Slowing down prior to the end will prevent any possibility of binding because of this ''overlap'' due to the gear train spinning down.