Got my Maslow finally assembled yesterday, and in watching it operate and seeing the results, I think there’s a flaw in one of my motors (perhaps in the gearbox?).
The left motor is sometimes smooth, and sometimes very jerky - I made videos showing both behaviors here:
It seems to be dependent on chain length, but I’m not entirely certain of that. When it happens, the sled moves jerkily (naturally), and causes a slightly scalloped cut.
Judging by the results, this is not normal behavior…
@grajohnt For faster debugging, it’s advisable to post the Firmware and GroundControl version you are running. We had similar jerky motion with earlier versions, but it’s supposed to be fixed with the latest Version 0.82. Does the >Test Motors / Encoders< from GC run smooth and give you >Pass< for all motors?
Upgraded to 0.82 on both GroundControl and Firmware
Still have the same issue.
Did some jogging around on the axes, and came to the following conclusions(-ish):
The shudder happens much more often when the left chain is being let out, but does sometimes happen when it is being pulled in.
There is no obvious correlation to chain length.
Another few general observations:
The left motor has always sounded different than the right motor - is this common?
My impression is that the shudder always happens a bit, and is perhaps more pronounced when the motor is turning slowly, but is just ‘masked’ when it is moving quickly?
That is strange, did the calibration and moves across the sheet with 0.82 and thought is was much smoother then with past releases. Will try tomorrow with slow feeds to see if I can replicate the shudder.
This sounds like the PID tuning is the issue, it’s far too aggressive at ramping
up the speed.
to check if it’s a bad motor, switch the left/right motors and see if the
problem moves.
As I understand it, with the change that was made to the final motor controller
board, the two motors are using different classes of PWM outputs (or different
timers, something like that), one at a far higher speed than the other. That can
cause the different sounds of the motors (and potentially cause this sort of
problem)
Just swapped the motors and recalibrated the chain lengths…
…and the problem still happens on the left - so good call there.
This seems like an ‘obvious’ problem - this really decreases my cutting accuracy. Here’s a photo of a cut, showing how sometimes the vibration causes scalloping, and sometimes it cuts smoothly. Is there anything I can do to mitigate this? Or just wait for a software fix?
(On reflection - that picture actually illustrates the problem really well. The ‘scalloping’ only happens on the vertical cuts, when the left chain draw is slow. On the curved cut, the chain is being drawn in faster, and the shuddering/jerkiness does not occur)
What feed rate are you using? That might make a difference. @bar mentioned that he would be looking at the PID values Monday, your excellent video documentation will help that. A sample file that reliably causes the issue would help as well .
Cutting this cross-shaped pocket cut now, which duplicates the behavior fine for me: test_pocket_cross.nc (13.2 KB)
Feed rate is 30 in / min (just going with what was recommended initially - have not perturbated this yet)
This short video shows the transition between a vertical cut (jerky motor motion) and a horizontal cut (smooth motion). This shows the jerky motion of the sled, and then pans to the motor to show what’s happening there. I guess you’re right, I could try to increase the feed rate so that it never goes too slowly, but perhaps that might make some of the cuts a bit too fast? Any suggestions for a better feedrate?
I really appreciate the help here - just want to get to making some projects!
As another data point, I have noted a few times during this cut a similar behavior on the Z-axis motor. Most of the time, the motion is smooth, but on occasion, it has a similar jerkiness to the left motor (I initially wondered if my clamp was too tight, but it seems to come and go?).
Having done a little bit of messing around with Arduinos and motor shields before, oftentimes unpredictable motor behavior has to do with power supply problems. This could definitely still be directly a PID problem (similar behavior influencing the different motors), but perhaps a beefier power supply could be tried as well?
Throwing in a blind shot in the dark. This is the new motor shield. I don’t have one, so can we find someone with similar observation skills on the old shield and new shield, to check if this is version specific?
Just swapped the motors and recalibrated the chain lengths…
…and the problem still happens on the left - so good call there.
just double checking to make sure I understand what you did.
you left the wires in place and physically moved the motors and the problem
still happens on the left side of the machine.
This means that it could be a problem with the driver board, or with the wires
on the left side of the machine.
Can you try switching wire bundles so that you are useing a different set of
wires to the left motor?
That would narrow the problem down to the driver board or the software.
I think there is a way to change the config so that you can swap the motors in
software. This would let you switch the cables on the board and see if the
problem is software or hardware.
Thanks for including the .nc file, that makes it easier to try to reproduce . The first time I ran the file, I saw some jerking during the first cut in the center and the first downward cut of the upper limb. Some of the radii in the corners were jerky as well. The jerking did seem to be the left motor, based on the noise it made. As the cut progressed, the motion smoothed out.
This feels like a PID issue to me, but I’m not up to trying to tune yet. For now, you could try using a lower feed rate - I generally use F15. Try text editing a copy of the .nc file and see if that doesn’t move more smoothly.
Correct. I did not do a full re-calibration after this, as I felt that the mounting holes for the motors don’t allow for too much slop.
Happy to try switching the cables and motors…
@blurfl - yes, you can hear it, for sure, and at least in mine, it shows up a a slight ‘scalloping’ on the cut (as shown in the picture previously), while some cuts (like the horizontal ones) are perfectly smooth. Will try a lower feed rate as well…
Fortunately the weather has cooled off a bit - working in a non-air-conditioned garage at 100F in Houston humidity has not been conducive to staying in a good mood!
… I’ll be watching mine today, thanks for the heads-up!