MPCNC lost it's way

I finally got to the point where I could surface a spoil board. Kind of a “this is the last thing I have to do before it’s ‘ready’.”

My travel area is 460mmX460mm. I used estlcam to do an engrave on the outer edge of a 460x460 square and a pocket inside that same square. I was using a 5/8" flat router bit which means that the pocket operation did about a 450x450 area first and then would go back and do the perimeter.

Because I’m bad at math, I had it remove 2mm of material. It did two 1mm deep passes. The first pass went great, but 3/4 of the way through the second pass, the X axis kind of went wonky. I actually saw the gantry pause for like 100ms or so, like it was getting hung up but I didnt see any reason for it. When it tracked back to the left, it basically went past x=0. I stopped the job before it tried to do the perimeter. First pass was totally fine.

The really confusing part is that peninsula on the right side. I would think that if it was losing steps and whatnot, that the peninsula you see on the right would have ended up being a bay on the right.

What is going on here?

Can you upload the gcode.

It looks like it was traveling right, got stuck, that’s what caused the peninsula, then when it traveled back left, it skipped again, when it jammed against the left side, and that sort of “fixed” it, so it was able to travel the whole distance on the next right.

Looks pretty cool though. Your bit must be very perpendicular to the work. The bottom looks very smooth, especially for a bit that size.

You might be right about that. It was a couple of days ago so I don’t remember exactly, but I thought I was watching it when it traveled back to the left and it didn’t overtravel, but I am going on 5 hours of sleep so who knows!

I am REALLY happy with how perpendicular it is. You can barely see tracks in person.

 

gcode uploaded

460x460-test-1.gcode (5.73 KB)

Are your stepper drivers properly tuned? Make sure they are not too hot.

1 Like

I bought my electronics from here so I assumed they were already done. I will try to check that today.

They should be but double checking can’t hurt. Could still be the gcode.

One of the motors was bad

1 Like