On May 26, 2012, at 9:49 AM, Jon Elson wrote: > Jeshua Lacock wrote: >> Greetings, >> >> I am now setting up my new X axis servo with an AMT 102 encoder. >> >> The motor is normally dithering between 0.0003 and 0.0006 inches. >> >> But every now and then - for no obvious reason it jumps up to 0.02+ inches >> and doesn't attempt to return to the correct location. It just jumps >> randomly. >> >> I think it must be interference. >> >> > Does this position change show up on the LinuxCNC DRO display, or are > you reading > it with a mechanical indicator? One possible cause is that interference > is being read > differently by the USC board and the G320, leading to a difference in > position between > the two. But, the PID algorithm should cause LinuxCNC to command a move to > null out the error. If LinuxCNC got false counts due to interference, > then you > would get a discrepancy between displayed and actual machine position.
Yes, it shows up on the DRO and the machine physically moves. So I guess they agree. >> The cable is double shielded and the shields are grounded to the encoder >> ground the length of the cable outside my enclosure. I am suspecting the >> interference is inside the enclosure where there are unshielded encoder and >> power wires. >> >> > Yes, keep these separated. it is best practice to ground the shield > only at the end where the signal > is being received, ie. the control box. It is. >> I was thinking of carefully wrapping aluminum foil around the power wires >> and grounding them, then shrink-wrapping or covering the aluminum with >> electrical tape. >> > Yes, this may help. but, a short length of exposed wire may not be the > cause, unless it > is really close to noisy power wires. Noise on the cable shield has a > MUCH higher > ability to couple noise, as it contacts the wire intimately for the > entire length. That makes sense. I just checked and I have continuity of the encoder DC ground from the beginning of the shield all the way to the encoder. The problem is the power wires are really close given their position on the Gecko terminals. The drive in question has all 3 Gecko power lines running right under it. I think this might be the problem. Here is a pic of inside my enclosure: <http://3DTOPO.com/CNC-enclosure.png> Don't judge me too harshly, I was making sure everything works before I finish tiding things up! Any suggestions would be greatly appreciated! >> Would I want to ground the power wire shield to the power supply ground or >> the encoder ground? Do you think this might help? >> > Shields for the power wires should NOT be grounded to the signal-sensing > components, > but to a different shield location. OK, so that should be the PSU DC ground then? Or the earthen ground? >> Anyone have any other suggestions? >> >> I wish I could see magnetic fields! > Well, with an oscilloscope, you really CAN sense them. With a DVM, you > may also be > able to. Find something scrap with a coil of fine wire. A few hundred > turns of tightly > coiled wire can be connected to your DVM on AC Volts mode, and will pick > up these > fields. Then, you can move it around and see where you get a high > reading. Near > the motor output wires of the Gecko 320 drives should give several > Volts, at least! Good idea. Best, Jeshua Lacock Founder/Engineer 3DTOPO Incorporated <http://3DTOPO.com> Phone: 208.462.4171 ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
