Centroid Community CNC Support Forum

Hi all,
I'm dealing with a strange phenomenon that I have no idea where to begin to address.
Sometimes during a cut, the torch will start to jitter. You can hear and see it happen towards the bottom of my build thread:

https://photos.app.goo.gl/4isqNxrL3p5wBykr6

It's definitely being told to do it, the machine is far too rigid to have some harmonic vibration build up.
Do I need to tune my AC servos? They're DMM DYN4's, I haven't played with any settings outside of that in the Wizard.
It's strangely intermittent, sometimes on orthogonal cuts, sometimes on diagonal.

Does THC ON or OFF make a difference with the Jitters?

Can You tell if the Jitter is in the X/Y or Z?

Can you turn your Plot to ON in the utility menu so more plots are captured.

Can you throw the g-code file also where the jitter happens?

Good idea about the THC, I haven't tried with it off, I'll do that tomorrow.

The jitter seems to be worse in the X direction now that I re-watch some cutting footage.

I'll switch on the plot thing you mentioned on my next cuts, I do keep an eye on the plot on the screen as it's cutting, and haven't noticed anything strange going on there.

Something else perhaps unrelated that I noticed it doing today while trying to cut some ½" steel: it ramps up the feedrate through the corner arcs. It doesn't change the % on the screen, but the torch moves noticeably faster than it does along the straight cuts. So much that the arc stopped piercing through in the corners unless I manually titrate the feedrate down below 80% before it gets to a corner. I can't find a setting anywhere that might affect that.

The speed up in ARC's is low resolution with smoothing. For slower feed rate jobs like 1/2 inch. Turn smoothing off.

What steps per rev is the drives set too?
Upping the steps per rev of the drives can help solve the low resolution speed up of smoothing

I'm assuming your drive steps/rev pulses/rec can be set higher. Closer to 6400 work better with smoothing.


I found it hard to see the jitter in the videos. I did notice the jitter in the diamond plate video towards the end I think.

I have seen the THC on some of the higher interference Chinese plasma cutters oscillate up and down. Following the high and low voltage but it might make the cut a little serrated but nothing like that. This would be the most extreme case of Z oscillation I've ever seen. You would definitely notice the torch going up and down rapidly

ok wow, I thought the jitters was bad. I'm having all kinds of problems now. Uploading videos to build thread now, it's going to take a while.

THC is crashing the torch randomly on very flat 1/4" steel. I switched it off and that stopped the crashes, but didn't help with the jitters. It's definitely in the XY direction, not the Z. Mostly the X judging by the sound of the vibrations when I do a dry run.

Maybe I'm misunderstanding how to use the restart feature, but it's completely deviating from the cut path. Both step forward and step back send the torch way off target.

Also the cut is disobeying the program. I set a program to cut 7 circles,
It kept trying to put the circles in the wrong places, overlapping them, it put the lead in on the wrong side of the line on one of them, and another one isn't even round. Between the crashes and the inability to use Restart Mode, everything kind of went to hell.

I've dealt with steppers missing steps before, but these are AC servos, I have no idea what could be going on. Maybe someone who can read gcode can see something obvious?

Steps per rev on the DYN4s I left at the default 4000. I suppose if I increase it, I'll have to recalculate the overall turns ratio, no big deal. I'll see if that has an effect on the jitter.

Wow. I would not run the machine until we figure out what's going on.

Can you creat a fresh report in the F7 Utility menu and post please.

No need to change the step rates at this time. Just run with smoothing turned off in the profile manager.

The drives and the wizard steps per revolution should always match.

Changing the steps per revolution of the Drives and the Wizard should not cause the turns ratio to change or be wrong.

This jigging motion will cause the torch to loose speed causing a a rise in voltage. THC is lowering the torch to lower the voltage.

Hi Joey,
here's a fresh report.
Thanks so much for your help with this
sorry my videos are still struggling to upload

Where was the G code generated?

Can you cut a right hand triangle? So we can see a cut with each axis and together?

Does this jitter accure with diagonal jogging?

You can turn Diagonal jogging on in the Wizard.

The G code came from the post processor in Fusion 360

I did some jogging before I slice up any more steel, and made a slight discovery:
While jogging I can only recreate the jitter in the X axis, and only from the VCP. if I use the MPG, it doesn't do it.
I couldn't find the diagonal jogging button in the wizard, but I'll look again.

I'll put some fresh consumables in and try a triangle now

I did some test cuts at different feed rates on some .063 steel. I saw some oscillation at the lower speeds, but it didn't show up very well in the cuts, the dross was pretty thick at 50ipm. Definitely seems to be worse along the long axis of the table, which sounds more like a hardware issue than software. But I can't find anything loose or worn out. X and Y are basically identical: toothed belt reduction, rack & pinion, ¾" linear rails, same size motors & drives.
And it only jitters intermittently. At 350ipm it moved smooth as butter. Although it didn't cut through the 16ga which I thought it should have