I used to have problems with one particular LED board I designed that would reset or stop talking on IIC bus when I turned on my workbench circular fluorescent light. The type with the magnifying glass. Turned out to be a PC board layout problem and pull up resistors in the wrong place and the wrong value.
Not an easy problem to solve when on a deadline. Way back in the early nineties a stepper motor that adjusted the tool height of a punch press by turning a ball screw would put noise onto the control network. It was worse during each punch cycle since the force on the press would try and rotate the ball screw a tiny bit and the drive circuits became noisier. The solution as simple but not applicable to CNC unfortunately. The engineers designed in a mechanical brake. Once the tool was inserted and the press closed to the calibrated position, the brake was enabled and the stepper driver was disabled. Noise gone. Network better. Moral of the story, always better to try and remove the noise from the source rather than protect every device from it. John Dammeyer > -----Original Message----- > From: Stephen Dubovsky [mailto:[email protected]] > Sent: December-20-15 6:59 PM > To: Enhanced Machine Controller (EMC) > Subject: Re: [Emc-users] Grounding Issues > > > Peter, this is the root cause in the change of behavior between <=2.6 and > 2.7 w/ the 5i25+7i77 that a couple of us have been having (was discussed > here about a week ago)? I have noticed an occasional warning in the prior > version I was running (just at startup iirc.) I need to put some more > noise filters in, but my error is from the motor contactors pulling in when > I hit enable. The original dc servo drives have big contactors that > disconnect the motor from the drive and put a braking resistor across it > when you disable them (or hit a limit switch, etc.) Hitting the enable > (F2) causes those 3 big guys to pull in and reconnect the drives to the > motors. That causes the sserial glitch and trips the contactors right back > off. I can't get past that step. I haven't had time this week to play w/ > the inc/dec parameters like Andy Pugh suggested. > > Best, > Stephen > > On Sun, Dec 20, 2015 at 1:34 PM, Peter C. Wallace <[email protected]> > wrote: > > > > > I looked into the 2.7 source and see that there is a problem with sserial > > error reporting that makes some single errors fatal (because the error > > reporting does not distinguish between per cycle error bits and sticky > > error > > status bits) > > > > That is errors that caused just a random popup on 2.6 will cause 2.7 > > to shut down the sserial port (you still need a pretty bad noise spike to > > cause these errors but they should not be fatal) > > > > > ---------------------------------------------------------------------------- -- > _______________________________________________ > Emc-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-users ------------------------------------------------------------------------------ _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
