On Friday 09 September 2016 13:01:28 Sarah Armstrong wrote: > iv'e had ground differences between a machine and a pc > blow a breakout board , it transpired they were across 2 different > supplies, in the building , one 3 phase , one single > Nothing so complex here, in fact it is all plugged into the surge arrestor, and a 2nd one plugged into the first one, so the work light and computer are on the first one thats plugged into a std 127 volt wall socket, and the 2nd one has all the other electricals plugged into it, basically NEC sick bird, but that gives me a master power switch for everything but the computer. Its a flea powered atom, so I leave it on 24/7 so I can write gcode running on the sim here, while using that machines /home/gene/linuxcnc/nc_files as the storage. > On 9 September 2016 at 17:50, Gene Heskett <ghesk...@shentel.net> wrote: > > On Friday 09 September 2016 12:08:27 Chris Albertson wrote: > > > It is hard to even guess the cause of this from your description. > > > You are just going to have to go down the line and test each part > > > independently. > > >
I have a stick of 74ACT245's but no 74HC00's or 74AS04's to completely replace every chip on this breakout board. But while I have a ground jumper from the lathe back to this box full of drivers, the box is transformer isolated, and running on a 2 wire line cord. So I'll go see if I can find a $3 3 wire cord I can put into it. Make another ground loop I guess. > > > I remember reading here in this forum someone posting about > > > "redundant" onto isolators being on both the breakout board and > > > the motor drivers. Maybe this shows that even using two of them in > > > series is not good enough unless you take other precautions in the > > > design. No opto's on this particular BoB. So it what National Semi once referred to as "damned fast" with 10ns delays going both ways. > > > I Think the #1 thing most amateur and many professional engineers > > > forget is that all ground wires and ground traces have some finite > > > amount if resistance and Ohm's Law applies to them Yup. Forget it, and get bit. > > > A good why to demonstrate this is to place an AC voltmeter into an > > > AC socket in a large building. Measure between Ground and Neutral > > > and you might see as much as 1.5 volts. Yes the ground and > > > neutral are tried together at the service entrance but the neutral > > > is currying some current, the ground wire in the normal case is > > > not. I have seen a nearly 20 volt difference. Some ID10T had replaced or patched the line cord on a big but aged chest freezer, and had wired the motor return to the static ground. I was not making favorable comments about his genealogy as I sorted that while 400 lbs of food was warming up. We were being a relief distribution center for the folks put out of work when a large local glass facility shut down without any warnings. > > > I'm guessing the root cause here is mixing up the types of > > > grounds. Logic level returns (called "ground" in many circuits) > > > are not the same as protective ground or AC mains return paths but > > > many times "it is all just "GND" and gets connected haphazardly by > > > those metal shells and screws on cables and inside power supplies > > > and other places. This is almost impossible to get "right" when > > > you are assembling a system from little boxes made by others. That last should be all caps. > > Amen but no other joyous shouts. Its a basic truth. > > > > But theres a bit of a hint there too, I have a hardwired ground from > > the lathes frame to the box all the drivers EXCEPT the spindle are > > in. but I am not 100% sure I tied that thru to the logic ground. If > > not, my bad dog, no biscuit. > > > > Something to triple check. > > > > Everything connected to that machine is on one wall plug, with surge > > absorber extensions furnishing the static ground to everything > > plugged into them, 2 of them in series, one nailed to the wall the > > computer is plugged into, and a second one acting as a master power > > switch for everything but the computer. Perhaps I need to verify > > that it is indeed connected. Something came in thru the grounds and > > used a very small ground trace on the BoB's pc for a fuse. And > > thats all I have for clues so far. I unboxed the motors housing so I > > could shine a strong light into the brushed end of the motor, but > > cannot see any winding damage. The other end of the motor is covered > > permanently by its fan/pulley, which has been drilled and tapped in > > the threaded joint, twice, 180 degrees separated and 4 ea 10-32 set > > screws installed as locks to keep the flywheel from unscrewing > > itself on a violent reversal, which Pico's PWM-servo amp can do. > > All sealed together with green threadlocker. (I hope) As for the > > reversal brutality, I have some hal slowdowns, but they get in the > > way when doing rigid tapping, so its a compromise between > > demagnetizing the motor, and how many turns the spindle overshoots > > the stop point, trying to get deep enough threads yet not bottom the > > tap and either spin it in the chuck, or break it. So I have some > > more hal trickery I can watch with a halmeter that counts the > > encoders overshoot. I grab the calculator and convert that to > > distance, and subtract tghe figure, obtained while tapping air, from > > the Z depth to program the g33.1 with. That was enlightening, above > > 200 rpm is so much overshoot its not worth cranking up the speed. > > 300 rpm is 3.75 turns to get to zero speed. The chuck is heavy, but > > effectively ok, its that pound plus of motor flywheel/fan/pulley > > that counts as the heaviest mass. > > > > But thats not finding the short that did all this damage. Later. Which I still haven't found. I am beginning to think I might have to hipot the motor. Something with some serious power came down a logic grounded wire and isolated the ground for all the inputs by blowing a board trace. > > > On Thu, Sep 8, 2016 at 2:59 PM, Gene Heskett > > > <ghesk...@shentel.net> > > > > wrote: > > > > Greetings Peter; > > > > > > > > Someone made the remark a while back that it was indeed possible > > > > to look at the output of a function, in this case the first > > > > pwmgen of a 5i25 with the prob-rfx2 firmware in it. > > > > > > > > I've had a blowup, the ultimate cause of which I have not found > > > > yet, but from the damages done to the cnc4pc C1G BoB that I have > > > > found so far, I am inclined to think the spindle motor psu, or > > > > possibly something line powered failed. The C1G damages so far > > > > found are 1 blown 74ACT245N, and an almost invisible use of a > > > > ground trace on the pcb as a fuse. All the opto's in my spindle > > > > encoder would appear to have failed. Those I have a bag of, but > > > > that pcb makes the tinfoil of a pack of smokes look like boiler > > > > plate, so I'll have something I won't call fun replaceing them. > > > > > > > > I can run the x/z motors just fine, but the limit switches would > > > > appear to be closed as those leds are at full britness when its > > > > several inches from the switches. > > > > > > > > What happened is that I had started a short program that called > > > > for a 400 rpm spindle but had neglected to engage the headstocks > > > > back gear, so the motor took off and wound out to max speed > > > > before I could hit the big red button. And about 1/2 second > > > > before my hand hit the button, there was a pop that came from > > > > the box over the spindle motor and everything coasted to a stop. > > > > The pop sounded like it could have been the top of one of the > > > > hexfets on the Pico Systems PWM-servo amp I am driving the motor > > > > with. But they look pristine, and all the other components in > > > > that box look good and a DVM says they are properly powered. > > > > Thinking the motor may have thrown a winding, I've looked it > > > > over and checked for shorts and come up empty so far. > > > > > > > > Back to the 5i25 question. So I fire up the halscope and send > > > > it to gpio1.in or in_not. And try to start the spindle from axis > > > > with the cable to the bob unplugged. > > > > > > > > I am seeing a short pulse of 1 volt p-p, which becomes inverted > > > > if I look at the in_not, and trying to increase its speed has no > > > > effect. So its beginning to look as if I may have blown this > > > > 5i25. And I've still no clue what did it. > > > > > > > > Do you concur, Peter? In that event, I'll need three more, with > > > > low profile brackets as I believe the one I was going to run > > > > either my toy mill, or this Sheldon lathe, has a bad pin that I > > > > was using for the probe, G38.2, and it got flaky. That was > > > > because the big dummy I can see in the mirror here forgot to run > > > > a decent ground to the frame of my G0704. And it was pretty > > > > darned noisy when I put a scope on it. Fixed, but too late for > > > > that 5i25. So now I need one for the little monster, one for the > > > > Sheldon, and one for my toy mill that I've not found a round > > > > tuit to do yet. > > > > > > > > Thanks Peter, or John T. > > > > > > > > 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> > > > > > > > > ------------------------------------------------------------ > > > > ------------------ > > > > _______________________________________________ > > > > Emc-users mailing list > > > > Emc-users@lists.sourceforge.net > > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > > > 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> > > > > ------------------------------------------------------------ > > ------------------ > > _______________________________________________ > > Emc-users mailing list > > Emc-users@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/emc-users 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> ------------------------------------------------------------------------------ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
