Please explain signal processing steps

[Beraht]

Could someone explain the "what and where" of signal processing for a ShopBot? Start with the gcode toolpath file being processed in the computer software to the processing that takes place in the control card, control board and the drivers that ends up being a step for a motor?

As a side note, this question is driven by weird behavior being seen on my 5x10 Alpha. The behavior is best observed when cutting a straight line at constant Y value. I have a simple demo code that should cut a straight line from X7,Y17 to X19,Y17. The cut is 0.25" deep. There are ten 0.15" high x 0.5" long 3D tabs. The gcode for this has zero change in the Y value, but when the cut is made, every Z movement is accompanied by a ~1/8" negative Y move. Going up and down a 3D tab is two Z moves and the cut location shifts ~1/4" to the negative Y direction at each tab.

This only happens on a toolpath with an X move. A similar line oriented 90 degrees so that only Y & Z are moving will cut perfectly.

Now I know some of you are saying that can't be, but here is the code that results in a 12" travel in the X direction and an ~2.5" travel in the negative Y direction:
M3,7.000000,17.000000,-0.250000
M3,7.280000,17.000000,-0.250000
M3,7.600000,17.000000,-0.100000
M3,7.920000,17.000000,-0.250000
M3,8.480000,17.000000,-0.250000
M3,8.800000,17.000000,-0.100000
M3,9.120000,17.000000,-0.250000
M3,9.680000,17.000000,-0.250000
M3,10.000000,17.000000,-0.100000
M3,10.320000,17.000000,-0.250000
M3,10.880000,17.000000,-0.250000
M3,11.200000,17.000000,-0.100000
M3,11.520000,17.000000,-0.250000
M3,12.080000,17.000000,-0.250000
M3,12.400000,17.000000,-0.100000
M3,12.720000,17.000000,-0.250000
M3,13.280000,17.000000,-0.250000
M3,13.600000,17.000000,-0.100000
M3,13.920000,17.000000,-0.250000
M3,14.480000,17.000000,-0.250000
M3,14.800000,17.000000,-0.100000
M3,15.120000,17.000000,-0.250000
M3,15.680000,17.000000,-0.250000
M3,16.000000,17.000000,-0.100000
M3,16.320000,17.000000,-0.250000
M3,16.879999,17.000000,-0.250000
M3,17.200001,17.000000,-0.100000
M3,17.520000,17.000000,-0.250000
M3,18.080000,17.000000,-0.250000
M3,18.400000,17.000000,-0.100000
M3,18.719999,17.000000,-0.250000
M3,19.000000,17.000000,-0.250000

And no, I am not making this up. I have run a number of tests swapping drives, cables, etc. as suggested by ShopBot Support. Current plan being suggested: try a new control card first and if that doesn't fix it, try a new control board. Would be great to learn someone else has seen such a thing and found the fix!

[steve_g]

Actually, to me, your issue sounds like a mechanical problem… perhaps a loose pinion or something that shifts in response to the change in forces by the Z axis hopping over the tab.

I’m hesitant to suggest other scenarios, but start with the standard “grab and shake” method, looking for any looseness in any axes.

SG

[Beraht]

Actually, to me, your issue sounds like a mechanical problem… perhaps a loose pinion or something that shifts in response to the change in forces by the Z axis hopping over the tab.

SG

Oh I so wish that was it, but no ... there is a signal going to the Y stepper. The movement is consistent and in one test I swapped the Y & Z motor cable. Since those motors work in reverse, I had to reverse movement direction as well. Once done, it cut a mirror image of the exact odd movement with the spurious Y movements going in the positive Y direction.

EDIT: Also, the motors are tight. When powered, I cannot budge any axis manually.

[Beraht]

One other point I should add ... when first started for the day, it would cut the first 6 or so tabs correctly and then start the stairstep movement. After the system warmed up, it would start going off path with every tab. That suggests "electronic" to me as in some component or board connection moving with thermal expansion. Since I don't know what does what in the communications path from computer to stepper, I don't know which part to suspect.

[bill.young]

Does it lose position ...cut the part the wrong size and not return to the correct home position afterwards? Or is it that the tabs are shifted but the parts are correct other than that?

[Beraht]

Does it lose position ...cut the part the wrong size and not return to the correct home position afterwards? Or is it that the tabs are shifted but the parts are correct other than that?

It loses position. Tabs are actually formed on the part of the line that is moving at an angle. Each wrong move by Y moves Y0 the same amount in the same direction.

[bill.young]

That's definitely a weird one!

After warming up does every Z move have a corresponding Y move? When it lifts straight up at the end of the file to its safe Z location does the Y also move, or is it only during a two axis XZ 3d ramp move like during the tabs?

[Red F]

Have you tried running the program while doing an air cut and is it the same result?
I wonder if it is a static/grounding issue?

[Beraht]

Red,
Yes, same thing on air cuts. It is not static or power cable interference. Have run every combination I can think of including dust collector off & hose removed, spindle power off, etc.
Becoming more convinced this is a fault on a circuit board.

[Beraht]

That's definitely a weird one!

After warming up does every Z move have a corresponding Y move? When it lifts straight up at the end of the file to its safe Z location does the Y also move, or is it only during a two axis XZ 3d ramp move like during the tabs?

When warm, does on every Z move, but I haven't thought to observe the move to safe Z. Can't do that now. Have boards pulled looking for some telltale sign of a crack or short.

[srwtlc]

What are the temps in your shop? Just for fun, leave the control box door open and place a fan blowing directly into the box. Possibly a bad solder spot that opens up when warm or other component.

[Beraht]

It is late June in Texas, so it on the hot side. Too late for that test. Already pulled the boards. While I understand the suggestion, I have a hard time wrapping my head around a break being the culprit. The behavior seems more like a short that allows Z signals to get to the Y motor. But I'll never claim to be an electronics guy.

[Beraht]

Follow Up: Apparently there is a problem on the control board. Installed a new control board and the problem disappeared.

[Red F]

Thanks for the update. Glad you got it sorted out