Hood, I tried the pulse width, along with, or without sherline mode, but then the motor just turns very rough. Gecodrive says to use a pulse width of 2, which does seem to be the best.
You know, the more I work on this, the more I think I've got some sort of voltage drop, or signal noise. While a jog key being held does cause minor twitching of the motor, there's still something else. I thought we had it figured out, until I let it run some straight line programs, 300mm one way, 300mm back. Sometimes it's within 5 seconds of starting the program, sometimes it takes 30 seconds or more, but I can hear the motor 'twitch' once in a while, and it usually stalls if there's any load on it. If I put the motor on the table and run the program, when the twitch happens, it actually jumps, like it tried to change direction, or stop instantly for just a split second.
I've been reading looking for how to hook up a capacitor to filter the power supply, but the numbers I was coming up with weren't in the realm of reality. Geckodrive says:
(80,000 * I) / V = C
Example: Using a power supply of 65V and 5A, the equation would look as follows:
(80,000 * 5) / 65 = 6153µF
So, for me it's like this:
(80,000 * 19amps) / 48volts = 31666µF capacitor ... I think I might have a hard time finding a cap like that around here ..
The last thing I thought of is line noise, but I have a spare motor with the factory wiring, hooked right up to the driver, and the twitch is still there.
EDIT: I wanted to mention that this problem has existed ever since my first setup of this machine. I started with a hobby cnc drive, and that did it too. I've tried a completely different computer with a fresh install of windows xp and mach 3. I bought this new G540 drive. I have a couple power supplies, ranging from nothing more than a rectifier and capacitors on a transformer to an vending machine power supply, and several others. I've tried unipolar, bipolar, parallel and series wiring.
I may go bald by the time I figure this out >.<