THC Start Marks

[Rodw]

Tom, yeh, I know the video was shot a while ago. I think "the problem" with LinuxCNC is there are so many ways to skin the cat so maybe it wasn't an issue. In the video, there is nothing in the spiral cut to slow the trajectory planner's requested velocity so I doubt the OP's issue would have bitten in that example. He was using a similar HAL config to the OP. It would be interesting to see what would happen if that machine cut a star shape on corrugated as the OP's issue bit in those conditions.

On up and down repetitive testing against a dial indicator on my hardware, I pushed it as far as I could and came off a bit. On real cuts, I lost steps on uptravel pushing against the mass of the axis hardware and I saw some pretty spectacular crashes. I did observe that loss of steps was much more closely linked to acceleration settings than velocity. In the end I settled on about 260 IPM velocity and 98 IPsec/sec acceleration which as you say is good enough at a lower cut velocity