Thanks for the link, Hu, it's info like that, that I was looking for.
Hi Scott_bob,
It's a VR510 Mach One router, 16 hp, ISO 30, 18k rpm. I wouldn't mind trying your test program, but as I said in the previous thread, it runs fine(fast, anyway) with G08 turned off. I did get Komo to fax me with our specs, and the applicable items are:
Look-ahead control (G08)
Feedrate clamp by circular radius
Bell-shaped ACC/DEC after cutting feed
Digital servos A06B series
I think I need to research the ACC/DEC params set by Komo, and tune accordingly.
As to my CAM, I use VisualMill 5.0 and RouterCim 2004. Rcim is used for panel processing. The VM programs I've run range from 7 - 30 mb. I post them with no seq. numbers and set all options to modal to reduce the file size. I am outputting linear moves 99.9 percent of the time, because our 210i only supports G17,18,19 for circular interpolation. I've tried chordal dev. from .001 to .0001. Of course, a coarser setting reduces the problem, but with the advanced ACC/DEC turned off, the control moves through the tight programs just fine. Again, I think the problem is in the params for ACC/DEC, particularly in the Automatic Corner Deceleration Function. I think the control is seeing any corner at all as one that must be clamped, and with such small line segments, it's always in ACC/DEC. I just need to study and get familiar with Fanuc param editing (kinda scary!).
Normally, with panel-processing, we DNC (the Fanuc DOMP) directly from the CAD station over a wireless LAN. That's the beauty of the i series OpenCNC control; I only keep about 3 programs resident in machine memory. In diagnosing this particular problem, though, I have been running DNC directly off of the Fanuc hard drive. Either way, I see no difference. The Ethernet is much faster than necessary for DNC purposes. I don't have enough memory to load anything close to that big directly into the Fanuc memory. I suppose I could try a small test program, just to see, but I would be willing to bet that the DNC is not the problem.
Either way, thanks for your input, and as I said, I would be interested in seeing your BPT test program.
Good Luck,
Jeremy Hill
jerhill1@netzero.com