rocketscientistnate wrote:
I reread the advert at homeshopcnc.com. Its 850 oz-in PEAK. 170 constant. Its definitly not a pulley slipping, I can watch the end of the motor shaft. I wrote a short program to just move the quill up and down, and watched it. It can go up and down a hundred times and then it just kinda stalls on one move and makes a nasty grinding noise but just creeps a little. When the controller sends it the other way, it takes off and goes fine again for a while. If it was something that happened everytime i did something in particular it would be easy to diagnose.
Hmmm, I wonder if you have a chip in your ballnut? I get these in my machine
every once in a while. My marginal quill motor will sometimes stall on this.
My EMC software is continuously reading the encoder position, and if it lags
more than a little, it will stop on a following error and let me know there is
a problem. What good are servos if it just blindly goes on mis-making the part?
(This is not a big problem here, it happens once every few years.) My ballscrew
is on the front of the quill (a retrofitted manual 1J head) and more prone to
crud getting into it. Yours is pretty hard to get to, so my trick with greasing and
cleaning may be the only way to get chips out of the ballnut or its roller bearings.
Are you using Gecko 320/340 drives? They should fault out when they get behind
by 128 steps.
I don't know if it helps, but everytime I run a program and it does this, it seems to lose steps in the same direction everytime it does it. Like in this program everytime it loses some it will be higher, or the next time I run it they will always be lower. Or say I'm running a 3d part where the x moves back and forth stepping .005 in the y every pass, (z axis making the profile) the whole part will be tapered from the lower to higher y values at a fairly consistent taper. Sometimes it tapers up, sometimes down, and ocasionally it works fine. If instead it lost steps both directions, it would be rough and jagged, but not consistently tapered.
This makes me think it might be an acceleration problem, or possibly something to
do with the timing between step pulses and changes in the direction signal.
Jon
