Travel moves are not straight. Is this by design?

dc42

Thanks for that last report. I suspect that both issues are caused by using a combination of high steps/mm (resulting in high step rates) along with extended step pulse timing for the external stepper drivers. What steps/mm (M92) and driver timings (M560) have you set in config.g?

Please execute a few high speed travel moves, then run M122 and report the MaxReps figure.

janbbeck

@dc42 here the results:

M92
Steps/mm: X: 1000.000, Y: 5120.000, Z: 5120.000, E: 2400.000:2400.000:420.000:420.000:420.000:420.000:420.000:420.000:420.000

M560 is not set in config.g

MaxReps: 115137, StepErrors: 0, LaErrors: 0, FreeDm: 240, MinFreeDm 216, MaxWait: 788103523ms, Underruns: 0, 18

3DPMicro

Wowzers. That's some serious steps/mm

janbbeck

@3dpmicro Yeah. Linear motors with 1 micron or better feedback on the X and Y. They can also move wickedly fast....
https://www.youtube.com/watch?v=lq-wMWWwWzQ

dc42

Yes, your MaxReps figure is way too high, you should try to keep it below 100. Your Y steps/mm in particular will cause serious problems. Can you configure your linear motors to use a lower steps/mm? What M569 T parameters do they need?

janbbeck

@dc42 Note sure what the timing parameters are they need, but I do know that I can put scaling parameters in for the steps. Adjusting those so that I get 100 steps/mm should be easy.

Will report back here. Thanks for the help!

dc42

200 steps/mm should be no problem if you don't want to go as low as 100.

janbbeck

@dc42 I am happy to report that this seems to have fixed my problems. Gonna do a few more test prints, but so far I think it's solved.

Thanks for the help!

janbbeck

@dc42 No more crashes. But I still have the staircase behavior. It has only happened on travel moves, but I don't know if that is a coincidence with them being faster and longer moves. Is there any easy way of finding out whether the Duet is sending the pulses out as bunches?

dc42

Can you check your MaxReps fixture again after a few travel moves?

janbbeck

@dc42 maxreps is typically this now:
MaxReps: 4, StepErrors: 0, LaErrors: 0, FreeDm: 231, MinFreeDm 120, MaxWait: 721129289ms, Underruns: 0, 0

dc42

That sounds OK. It looks like at worst, when some interrupt (probably a tick interrupt) occurs 3 pulses become overdue, so it sends a bunch of 4 pulses when the interrupt completes. Hence MaxReps = 4. This will probably happen one per millisecond.

janbbeck

@dc42 Ok, I take that to mean that I should look at the servo drive causing the behavior, right?

dc42

@janbbeck said in Travel moves are not straight. Is this by design?:

@dc42 Ok, I take that to mean that I should look at the servo drive causing the behavior, right?

Yes. If you use an oscilloscope to look at the step pulses, I am sure you will find they are regular. If they were not, then normal stepper motors would show the same problem. Even when step pulses for more than one motor become overdue, RRF continues to serve the motors evenly, except in extremely old firmware versions.

janbbeck

@dc42 I will replace the servo drive with one from a different manufacturer. Thank you for your time and help!