Ed Nisley <[EMAIL PROTECTED]> wrote: > On Wednesday 29 November 2006 21:54, > [EMAIL PROTECTED] wrote: > > So we can review it, can you paste your ini file and > > you hal file(s) on www.pastebin.ca, and post a link > > http://pastebin.ca/262379 > http://pastebin.ca/262384
Thanks. 16000 steps per inch max_vel of 0.3 = 4800 steps/sec max_acc of 0.6 = 9600 steps/sec^2 25uS base period means absolute max step freq is 20KHz Everything there looks fine, and having the numbers is helpful. > > Please make sure your following error limits are set to > > well under a half a turn of the motor. Usually you can set > > the limit to maybe a 50th of a turn. > > Bingo! > > The current limits are 1.0 & 0.010, which doesn't match > either of the two original stepper ini files. Looks like an > obvious brain cramp on my part. I don't know what, if > anything, I was thinking. I'll tighten those up before > doing anything further. > > The Sherline is 0.050 per turn, so max and min should be, > say, 0.001 and 0.0002? Seems awfully tight, given what's in > the original stepper_xyza.ini file: 0.050 and 0.010. Yeah, 0.001 and 0.0002 is a bit tight. But 0.050 is a full turn of the motor... maybe 0.0010 or 0.005, and 0.002 or so. > I evidently have two unrelated problems: the direction > glitch and a hole in my foot. Dang that footgun! The FE limits are unrelated to the problem. I just want them tighter as a diagnostic measure. When the machine goes nuts and moves down instead of up, there are several possibilities: 1) The EMC motion controller is incorrectly telling stepgen to move down, and stepgen is correctly doing what its told. This is unlikely, because you are reporting abrupt direction changes. Both EMC and stepgen individually limit acceleration, so abrupt changes would require both to be broken the same way. So I'm ruling that out for now. 2) The motion controller is telling stepgen to move up, but for some reason stepgen is generating down steps. In that case, the position feedback from stepgen would show the downward movement. If the command from the motion controller is going up, and the feedback is going down, as soon as the difference between the two reaches the FE limit the controller will trip out and indicate an error. By setting the limits tight, the controller will either trip, pointing at stepgen internals, or not trip, pointing at possbility #3. 3) Stepgen is correctly generating up pulses, but something between stepgen and the motors is messing up the direction signal. > > Does it seem to happen more on one axis or another? > > Nope, it seems equally distributed. The cam is symmetric > across the Y axis and the largest drift is toward -Y, which > is what I'd expect from lots of fiddly Y-axis motions with > an occasionally low direction signal. The other axes also > drift in the "low direction signal" direction. > > The most glaring error, however, happens when the final G0 Z > upward jog that's supposed to clear the clamps goes the > wrong way and gnaws into the fixture at full speed... Until we get to the bottom of this, I suggest removing the workpiece, fixture, and tool from the machine. If you have a jog that is moving an appreciable distance in the wrong direction, that is something completely new, and also completely unrelated to an occaisional pulse the wrong way. How repeatable is this "last move the wrong way" ? > That could just be my bogus FERROR settings at work, so > unless / until I can present some solid evidence, > concentrating on the leading edge makes perfect sense. The "bogus" FE setting won't cause this problem. FE is just a way to detect improper motion and shut down - in this case, a diagnostic tool. We did quite a bit of investigation last night, and understand the cause of the "wrong pulse at the start of the first motion" issue. However, we are sure it is NOT the cause of the "move a long distance the wrong way and hit the clamps" problem. More later, gotta go. Regards, John Kasunich ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
