I seem to have a talent for bringing out the best, as well as the worst in my fadal lately.
I decided to run a program from internal memory tonight, the part ran great through several tools all the way up to the final toolpath - rigid tapping some 2-56 tapped holes 0.125" deep in aluminum using a form tap. The rigid tapping cycle seems to not work well at all when run from internal memory. The spindle jerks around really bad to the point that it just stops a lot, then tries to resume, then stops, then resumes, and all the while the lights are flickering and the spindle drive cannot be very happy through this violent jerking.
When I run the exact same code over DNC, no trouble at all - works great like the spindle does a little ballet at 500 RPM.
I may have noticed that when run over internal memory that I didn't see the spindle do it's usual 3 or 4 turns before actually stepping down and running the rigid tap cycle. I believe this is used to get the speed readout correlated with the encoder / c-axis controller card.. I don't know.
It's like the spindle over-shoots its speed target then has to try and slow down, which it overshoots too and comes to a halt. All the while during this spindle jerking the processors are trying to keep the z axis sync'd up with all the bullsh*t spindle behavior, which it doesn't do so great. The tap didnt break, but the threads are very loose and messy.
I went back to DNC, and - no problem at all.
Maybe its some kind of gain setting issue? ..no idea.
wtf, again.