On Sunday 15 January 2017 04:11:52 Nicklas Karlsson wrote: > > > I should probably start a new thread here but there is definitely > > > a similarity in the noise problem. > > > > > > In my jet 9 x 20 conversion I was getting lots of noise spikes on > > > my spindle encoder feedback to the 7i78 card. > > > I went through all the regular suspects, grounding, shielding > > > twisted pair etc. with little or no change. > > > On a "real" scope I was watching the noise spikes as vertical > > > lines superimposed on the encoder A and B square waves. The noise > > > was there even if the spindle was NOT turning and when the spindle > > > motor controller was OFF. > > > > > > I started turning off things to see if the noise was from an > > > external source such as the monitor or computer's switching > > > supply. I even disconnected the DB-25 to the computer and supplied > > > the 7i78 from a linear supply.....No change. > > > > > > Then I wondered what would happen if I turned off the G320 servo > > > drives. I pulled the error/reset connections on the G320's and > > > the noise went away. > > > When I hooked the error/reset lines back up to one of the G320's > > > and decreased the horizontal sweep on the scope to 10 microsec per > > > division I got a beautiful rock solid 2 volt burst recurring every > > > 50 microseconds and ringing for about 10 microseconds. > > > > > > My noise source is the G320's PWM generator which switches at 20 > > > khz. > > > > > > I'm still not sure how such a horrible noise signal gets from the > > > 48 volt motor armature switching circuit in the G320 to the input > > > of the 7i78 but at least now I know the source and can try to fix > > > it. I'm pretty sure that I can't just put a cap across the > > > armature outputs of the G320 because that does bad things to the H > > > bridge. > > > > > > Has anyone else experienced the PWM switching noise of their > > > drives getting into their signal level circuits? > > > If so, how did you deal with it? > > > > > > Thanks, > > > Cecil > > > > Ahh, so I have company. I've converted the switching motor psu's to > > analog, but its pretty hard to block when the rediated noise is 12 > > volts p-p at about 100 megahertz before you start building a star > > ground. > > Common mode filter on motor cables? > That is what the clamp-on ferrite cores will hopefully do. When you cover all 4 wires with one core, its not much of an impediment to the currents going to/from the motors, as the opposite currents cancel. But only IF they truly are in time, but most of these drivers are not since they all sequence the switching such that there is a microsecond or so of dead time to prevent shoot-thru between the upper and lower hexfets in the H bridge. This generates a lot of longitudinal noise, that the clamp-ons will oppose, typically by being a resistance in the 150 to 250 ohms range at 100 MHz, blocking the noise by around 30 db under ideal conditions. This kit has cores as large as a 13mm bore, and one of those will be put on the vfd's wires to the motor, and another on its power input wires. Then one of the smaller ones, 5mm bores if they'll close on the stepper motors cables close to the green plugs, 7mm if the 5's won't close.
And of course since the encoder cable is so long, as are the limit/home switch cables, they'll get whats left of the kit. I've not even mounted any of the switches yet as the trip across the top of the carriage is dangerous to the wires, so there will have to be a switch on the carriage that stops it from crushing the 5/16" bore SS tubing its already in by hitting the tailstock. And another in front of the z screws tail bearing in case the tailstock is off. I've rigged a conduit the length of the bed, with a cable exit in the middle, so those cables that need to get to the carriage can exit into a cable chain going up to the rear of the carriage, thence thru the SS tubing I've yet to fasten down. Those will restrict the rightward movement by about 3/4" to keep the tailstock, which fits nicely between the legs of the carriages H frame shape, from hitting the SS pipes acting like miniature conduit. And I just thought of a use for that 4 channel opto-isolator board I bought, and which is currently watching for non-existant x motor drive but I just moved the hal file paths to use stepgen #3 in place of stepgen #0 since #0 and #2 are dead, put it on the outside of the big box in its own box, and use it to isolate the home/limit switches from all this noise. That way the noise those long wires can pick up, stays out of the big box. Once the noise is under control, replace this blown 7i90 with a good one. Thinking out loud, I wonder if that thing is fast enough to isolate the encoder? I know its slow as when I tried to use it to drive the vfd's speed input, it was non-linear as could be even with the pwmgen running at a measly 10 kilohertz. The bottom of the cable chain rides the top of the conduit, but I've yet to see if I can rig a sliding "swarf roof" over the cable chain, but its a ways down the to do list, as is a shield over the chuck whose major purpose will be to contain swarf. Sigh. It never ends... Now back to your regularly scheduled programming. :-) Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page <http://geneslinuxbox.net:6309/gene> ------------------------------------------------------------------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today. http://sdm.link/xeonphi _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
