Jon Elson wrote: > John Kasunich wrote: >> >> You might want to use halscope to look at the index enable signal and >> the position feedback from the encoder to EMC. Trigger the scope on the >> falling edge of index enable, and verify that position either is or >> becomes zero at that moment. If it doesn't, its probably a driver bug. >> > I got all this debugged around Jan 8, 2007. I have sent Stuart > a set of updated EPROMS for his encoder boards. > > At the end of the debugging process, I observed some behavior > that may have been the same as Stuart is describing. But, I was > a lot faster to hit e-stop, so I did not wait to see whether EMC > was going to stop somewhere within the machine's travel limits > or not. The final problem turned out to be the polarity of the > home switches was backwards. This was evidenced by the machine > moving in a direction opposite to the sign of HOME_SEARCH_VEL.
Stuart has posted his ini params, SEARCH_VEL is position. He has also stated that his machine moves in the positive direction when he tells it to home. So it would seem that he has his switches right. > Well, instead of understanding why it moved the wrong way, I > just swapped the sign of that parameter. It seemed to home kind > of like I expected, but would occasionally (like 50% of the > time) take off at a fast clip. Reversing the sensing of the > switch's polarity fixed the problem, and made the home approach > consistent with the HOME_SEARCH_VEL parameter. > > I have homed a number of times since then, and it has never > taken off rapidly since. > > Does anyone know if there have been any changes to any code that > might affect the homing process since 1/8/07? I don't believe > my Bridgeport computer has been updated since then, and it is > the only one with home switches or encoder index signals. The homing code is in src/emc/motion/control.c. I certainly hope that you as a board member have seen and know how to use the CVS browser at http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/ IMO, the revision graphs are the greatest thing since sliced bread: http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/emc/motion/control.c?graph=1 You can see when the 2.1 branch was made, and when every change was made. along with part of the commit message. Click on a box and you see that version of the file. Click on the link between two boxes and you see the diff between those two versions. You can also see any version of any file completely annotated to show which lines of code were changed at what revision: http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/emc/motion/control.c?annotate=1.107 The main homing state machine runs from line 1282 to 1736. The most recent change in that range is line 1735, but that was just a comment change. The next most recent is line 1724, which _would_ affect index homing. Clicking on the "1.85" link there tells me that change was made on January 3. 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
