Losing My Way

Hmmm... that's no doubt a misunderstanding of the units on velocity. I guess I assumed it was inches per minute. Through my experiments I start to stall at around 350 ipm and so I chose 120 ipm as a maximum velocity. I guess that's 120 ips though.

I will take your suggestion on the acceleration and see if that gets me anywhere.

A lift from the docs

MAX_VELOCITY = 1.2 - Maximum velocity for this axis in machine units per second.

MAX_ACCELERATION = 20.0 - Maximum acceleration for this axis in machine units per second squared.

Your figures would be high for a metric machine, but when you multiply by 25.4, they are astronomical.

For strong Nema 23s (3 Nm +) with a good high voltage supply, I would probably use somewhere between 25-30mm/sec max velocity and 80-120 mm/sec/sec acceleration
All depends on how many pulses you can get, which depends upon the base-thread being fast and without hiccups and the amount of gearing, multi-stepping and axis weights etc.

That equates to approx 1.2 in/sec and 4.8 in/sec/sec respectively using the higher end of the figures, so you will see that almost exactly matches my divide by 100 and 10 recommendation

Regards base-thread, I usually multiply the best latency figure by 3 and try working down once the speeds are sorted.

regards

With the acceleration and velocity lowered as you suggest the problem becomes even more pronounced. I can stand at one side of the gantry and watch the other side begin to move before the side that I am standing at. I believe I am observing this only when there is a change in direction... almost like backlash.

Any thoughts on that?

No, I am not sure about the drive strength of the parallel port on this motherboard. However, I do know that this problem is persistent through two other parallel ports (one on another motherboard and one on a PCI card). That means nothing, I know. I was wondering if the output voltage of the parallel port might be too close to Vin min on the gecko. I was thinking maybe I'd try to wire in a small board with some optocouplers and a 5V supply. That's a bit of work so I was leaving that until I get absolutely desperate.

The G540 does not support a slave configuration. I'll return it to two separate step/dir pins on the same step generator and make the modifications to the step timing as you have suggested. Fingers crossed....

I sent the G540 into Gecko and they did find that two of the drives had damaged opto-isolators. They did not provide any information so I can't tell if it was the input or output side which was damaged. I just assume these are the opto-isolators for the step and direction inputs too. Anyway, I was happy about that but after installing I found the problem still exists. This time I decoupled the motors from each side of the gantry and ran them with no load. The problem is quite obvious. Have a look for yourself...

I think this pretty well rules out any kind of mechanical issue. I also believe it points to the G540 still. If I swap the motor, the problem follows the drive, not the motor. Keep in mind that this is using one step and direction signal applied to both motors so they should both be receiving the same input.

Next I plan to do some testing to see if one is simply lagging or if one is moving when it shouldn't (noise?). Regardless, I think this implicates the G540 still.

www.dropbox.com/s/1xrnv6fybi8wvy8/2015-0...%2010.59.08.mp4?dl=0

was it always making that horrible noise?
It sounds like the stepper is stalling, but much worse.

You can check the resistance on all the motor coils and compare them to another motor.

No, the noise you hear is a vibration caused by the temporary arrangement. If I were to tighten the motor frame to the bracket and hold the bracket with a clamp you would not hear that.