Update: So the usdigital differential receivers arrived yesterday spent the day installing on all three axis. I also trimmed all the encoder and power wires as short as i dared. I then decided to test my handy work and make sure i didn't cross any wires. Machine jogged in all three axis and all seemed good. I set a dial indicator up and i commanded the x-axis to jog .1" and got .099", i must have jogged the axis fifty times one inch one way then an inch the opposite and every .1" motion it now comes up .001" short. I didn't even get to see if all the work helped in reducing noise as absolutely nothing was running. I was just moving the x-axis around. Has anyone seen this before? The tune seems stable halscope shows a .000017 overshoot at the start of motion and then a worse case lag of -0.00002.
Thanks --- On Sun, 2/7/10, richard harris <compact...@yahoo.com> wrote: From: richard harris <compact...@yahoo.com> Subject: Re: [Emc-users] Directional motion loss To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net> Date: Sunday, February 7, 2010, 2:22 PM The encoders on this machine are differential sealed units from BEI. When i got the machine one was nonfunctional so i sent all three to BEI and had them rebuilt and upgraded from an incandescent bulb to led. The encoder is mounted to the rear of the motor, and shows no sign of any oil or coolant contamination. Currently the encoders are wired as single ended, i plan on ordering differential receiver from US digital on monday. All wires are shielded and grounded at the controller side only. All cables are run in flexible metal conduits. I checked all the grounds along with all the mechanical couplers, nothing looked unacceptable. I wrote the code below to see if i could manage to duplicate the error without scraping parts. G1 X.5 F30 M3 S4000 #1 = 0 O101 while [#1 lt 150] G1 X2.0 Y2.0 z-1.0 F40 G1 X.5 Y0.0 Z-1.716 #1 = [#1+1] O102 if [#1 EQ 30] S3000 (MSG, 30 loops down) O102 endif O102 if [#1 EQ 60] S2000 (MSG, 60 loops down) O102 endif O102 if [#1 EQ 90] S1500 (MSG, 90 loops down) O102 endif O102 if [#1 EQ 120] S1000 (MSG, 140 loops down) O102 endif O101 endwhile M5 G1 X.05 Y0.0 Z-1.716 F30 G1 X0.0 F10 M2 This code has more linear inches of motion and is longer than the program that was exhibiting issues. I ran it three times without the coolant or spindle running, i had two other machines running at the same time as i did the other day. I also had halmeter showing the encoder count, not sure if it would show anything of value but i was curious. The machine repeated within the tolerance of the indicator.. I then ran it three times with the coolant and spindle running and witnessed two thousands of motion lost over the three runs. However Halmeter showed the same number of encoder counts. I am assuming two things, 1 my problem is noise, 2 the error was less during the test because i did not load the spindle. Do those sound like reasonable assumptions? My plan at the moment is get a differential receiver on all the encoders, and run new wires on the x-axis just in case. If these mods don't improve the situation I'm thinking of adding a line reactor to the VFD. Any other straight forward things I can do to improve the noise situations? --- On Fri, 2/5/10, richard harris <compact...@yahoo.com> wrote: From: richard harris <compact...@yahoo.com> Subject: [Emc-users] Directional motion loss To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net> Date: Friday, February 5, 2010, 6:17 PM Hey, I am running emc 3.2.0 off the BDI with Jon Elson's pwm servo boards. The machine is a knee mill, of similar size and mass to a hurco. This machine gets little use as its bigger brother does all the heavy lifting so the problem i describe below might have existed for some time and not been noticed based upon part configuration and tolerance. I notice during the last few days, while machining a thin wall box that one sidewall was thicker than the other. At first i assumed that i had a chip in the vise when i loaded the part, however the next part had a similar condition. I re-zeroed the machine and noted that it showed i was .010" off. Ran the next part and it was fine, the second part showed some issues, and the third clearly had a problem. Part is taking 17 minutes, and the properties pull down shows 400" of motion. I checked the machine for a loose belt, loose ball screw to thrust washer, and loose ball nut, all of these looked fine. I continued to run the code, cutting air and double checking the machine position to the zero point on the vise. I am loosing .002-.005" to the in the minus x direction. I have checked for backlash within the working envelope with a dial indicator and have not found anything that would alarm me. Any ideas on what would causing a loss of position in one direction only it is repeatably losing in the minus direction only. Mechanical, electrical, anything at this point. As always thanks. Richard ------------------------------------------------------------------------------ The Planet: dedicated and managed hosting, cloud storage, colocation Stay online with enterprise data centers and the best network in the business Choose flexible plans and management services without long-term contracts Personal 24x7 support from experience hosting pros just a phone call away. http://p.sf.net/sfu/theplanet-com _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users ------------------------------------------------------------------------------ The Planet: dedicated and managed hosting, cloud storage, colocation Stay online with enterprise data centers and the best network in the business Choose flexible plans and management services without long-term contracts Personal 24x7 support from experience hosting pros just a phone call away. http://p.sf.net/sfu/theplanet-com _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users ------------------------------------------------------------------------------ SOLARIS 10 is the OS for Data Centers - provides features such as DTrace, Predictive Self Healing and Award Winning ZFS. Get Solaris 10 NOW http://p.sf.net/sfu/solaris-dev2dev _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users