Losing Y steps - but odd

[twehr]

Do any of the engraving options (single direction left to right, scanning up and down on the Y axis, etc.) change the symptoms?

Regards,
Leon

I haven't tried them yet. going to single direction should not affect it. I thought about trying the y scanning. Won't want to live in the that world as it y scanning is typically tougher on the mechanics that x - too much mass to scan with, but it would be interesting to know. Just not enough time this pass weekend to test all possibilities.

There are a couple of other builders that will be up on the DSP soon. We can confirm (or not) the symptoms and go from there.

[twehr]

I spent some time this weekend working on the laser - cleaning optics, checking alignment, etc. I also decided to see if making some setup changes would affect the shortening issue.

I reduced the microsteps by 50% and doubled the pulse width (DSP calibration setting) to compensate. That reduced the speed at which the shortening effect would show up. If I carried this theme to an extreme, I could get it down to about 12 mm/sec before the issue showed up. At those settings, however, there was a lot of hesitation between scan lines, so I brought it back up to a happy compromise. I now only get the shortening effect at scanning speed < about 28 mm/sec. I never have to do raster work that slowly anyway, so I am happy with that.

Still don't know why the slower speeds are affected, when intuitively we would expect lost steps to come at higher speeds. But it is working, and that is what matters.

One thing to note - I am using .9 degree (400 step/rev) motors. Most of you, including Bill Meade who is also running the DSP, are using 1.8 degree (200 step/rev) motors. Since Bill does not have the same issue, I can only assume it has something to do with 400 step vs 200 step. I might pick up a 200 step motor for the Y axis (X has no problems) and see if that changes things.

[JBGrav]

Hi,

some progress in the problem? I have DSP with .9degree motors and the same problem too. Hovever all speeds (fast > 114 < slow) generate same Y-lenght loosing steps (microstep 1/8). The strange thing is problem happen only in X_swing raster mode. Y_unilateralism raster perfect Vector work perfect too. I have no idea what's next to fix.

Does anyone have any experience with DSP and loosing steps?

Best regards,
JBGrav

[bdring]

Does anyone have a scope to check the step pulse length?

[twehr]

Hi,

some progress in the problem? I have DSP with .9degree motors and the same problem too. Hovever all speeds (fast > 114 < slow) generate same Y-lenght loosing steps (microstep 1/8). The strange thing is problem happen only in X_swing raster mode. Y_unilateralism raster perfect Vector work perfect too. I have no idea what's next to fix.

Does anyone have any experience with DSP and loosing steps?

Best regards,
JBGrav

Look at your Pulse Width setting in Manufacturers Parameters. If you have it at about 12.xxx then you need to double the pulse width and halve your stepper drive microsteps. For my .9 degree (400 step) motors, I finally had to go to pulse width of 25.416064 with 800 microsteps. That made it so the shortening only takes place at < 30mm/sec. I never do raster work at that slow a speed anyway.

[twehr]

Does anyone have a scope to check the step pulse length?

I have a scope. As soon as I get out from under a terribly busy schedule (soon I hope) I will measure the actual pulse width. If he sets it up like I suggested in the previous post, he should be OK. I think the DSP just does not like too many microsteps and short pulse widths.

[bdring]

Tim,

What stepper driver are you using? Can you change the resolution via that?

[r691175002]

I get the same problem using the Retina controller. It happens only on certain speed/DPI combinations so I just avoid them. Increasing voltage to the steppers made the problem occur at fewer speed combinations but did not entirely eliminate it.

I was never able to come up with a plausible explanation for the problem, I'm almost 100% sure that the drivers are getting the right pulses.

I am using keling drivers (4020), I feel like the blame lies with the microstepping allowing the motor to stick at very slow speeds but that doesn't explain entire steps being lost.

The more information I get on the issue the more confusing it becomes. Each piece points to a different problem. I find it incredibly unlikely that this is a software issue because of the situations which cause it. I would try swapping in polulu drivers but I'm at school right now so I can't use the machine for a few months.

[bdring]

Some motors will loose step due to resonance. It only happens at slower speeds.