Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: BClemens on February 21, 2009, 11:07:08 AM
-
I seem to have developed the same problem but only on Y axis - oval circle and the next Z step down is totally misplaced.
I'll start another thread....
This is a new problem because it was doing fine. I have removed the motor, checked backlash and gib tightness, checked the wiring, re-tuned the G320, and all I can think of. A repeating/revolving Y axis MDI will run as if it is OK then will miss an entire group of steps, then begins to work again - intermittent. Also when holding the Y axis mouse for constant jog it will miss for a group of steps and then begin to move - the DRO has counted all the while. These servo motors more than adequate. I'm still searching and I guess next step will be to make sure that the encoder is receiving constant voltage from the G320. Maybe try more of the computer unplug and replug trick...?
Thanks,
BC
-
BC, You didn't say what machine you were using, but I had the same thing happen to my "Y" axis. It ended up being a coolant leak getting into the encoder housing and the lines of encoder were being damaged ! After all they are printed on with either bug spit or berry juice ! I replaced the encoder three times before I found the wiley leak, but all is well now.
-
Thanks Cruiser,
I checked the encoder when the motor was off. I checked the brushes too and broke one of the phenolic brush holders by tightening brush retainer plug too much....(Thanks for JB Weld!) The only thing I saw was a little piece of something on the encoder disc but brushed that away with a small soft paint brush and the motor brushes have very little wear at all. While the motor was off, I ran it for a while one way then the other.
Still searching....
Thanks very much,
Bill C.
Oh yes....this is a knee mill that I converted over a few years with ball screws and etc. It uses servo motors. The machine has been doing great, then begins to miss steps in Y - can't see the solution!
-
Also when holding the Y axis mouse for constant jog it will miss for a group of steps and then begin to move
Hi, I had a similar problem when I first got my servo mill up and running, One of the drives kept moving or not moving according to some unknown problem. It was only when I was moving the mouse cursor across the screen I found the problem. A quick change of mouse and I was away, problem gone.
Dont ask me how or why cause it was a while back, all I know it had me confounded for ages.
Hope this helps
John
-
This is a Microsoft mouse - couldn't possibly be faulty!!! (yea right!) Thanks for that input. Actually I believe this mouse is doing some funny stuff. I will change that too.
An oscilloscope on the encoder input to the G320, the 'A' signal quits at the same time the fault LED lights and the Y axis motor stops - but the DRO count keeps going.... Could this be a bad encoder? It is brand new (almost) and hasn't but just a few hours on it. I've taken the servo motor back off again to take a closer look at the encoder wheel. Cruiser may have hit on something here with what we're seeing on the scope - and intermittent as well. I'll be back....
Thanks again,
Bill C.
In the opening slot where the wheel is - in the encoder, should the disc be closer to one side of the slot? I cannot find any info on this from US Digital (E2)
-
It's getting better! Now the motor continues to run even with the encoder disconnected from the G320 and no signal coming from Mach3. It will change direction by mousing the continuous jog but it appears that the signal from Mach3 disappears after a short run. The other axis are totally normal and will function even with the Y axis motor running away (I took the belt off after it slammed the limit switch). Now I'm totally baffled. I don't know what to replace; the encoder or the G320 or both. Replacing them both may become a 'political' problem. I'm getting a charge out of setting this machine up, but this is getting down right ridiculous. Cut about a foot of material then replace something on the operating system....
I'm still not sure where the encoder optic wheel goes in the slot in the encoder optics. I must have damaged it when attempting to clean up some dust out of it somehow. I was thinking when this mill was first being set-up that the motors would not even run unless the encoder was input into the G320, that is a note that I forgot to take...
Bill C.
-
Hi, Bill
Sounds like your encoder is bad, Swap your Y Gecko to one of the axises that's working, The terminal strips will un-plug.
A Gecko without a encoder will run off in one direction on re-set looking for encoder pulses.
Mach will continue to count if your axis trips unless you've wired the err/trip signal to Mach to halt it.
Chip
-
Thanks Chip, actually that is good news... With the system in reset, plugging in the Y motor and encoder, it just runs away when the reset is engaged without any input to the Y axis. That's why I was concerned that the G320 may have fried, but then it does respond when an input is given in Mach3. So by 'cleaning it' I basically finished it off!
I'll shift a G320 to another axis as you suggest and see what happens. I'm glad I'm not getting paid for this; I would have starved by now! But I work very cheaply for myself.
Thanks,
Bill C.
-
OK, got it from another person who knows: the encoder is the culprit. "The Gecko G320 has no safety built-in to stop the motor if it loses feedback from the encoder." Although that doesn't seem right, it is proved by shifting the motor leads to another axis and G320 - does the same thing - runs away. Tried the Y G320 on a known good axis - all good. Mach3 axis super quick shutdown (reset) via the limit switches saved a massive amount of damage in this situation.
So; why did a US Digital E2 encoder fail? Are these encoders worthy of purchasing a replacement or should these be changed to something better - another brand? Or should I just go ahead and buy enough of them to replace them all as they fail? (that's a silly question right there)
Thanks, keeping the 'stimulus plan' working from this end,
Bill C.
-
Hi, Bill
Glad you got it sorted out, Hard to say on the encoder, Defective, Damaged on install, Just try to keep them shielded and clean.
Gecko's do have some safeguards, If it see's more than 128 inputs pulses with no encoder movement thy err/trip out, But on re-set it's not seeing any input pulses, So it has no way to know, That's why you have that big red E-stop Button.
Chip
-
Hello Chip,
OK, you cleared up another question mark. The shut-down by the G320 was witnessed. The paperwork for these 320's is around here somewhere but they were purchased in '03.
Thanks,
Bill C.
-
This is what Y did.....
-
Replaced failed (I thought) Y axis servo encoder, parallel port cable, tuned and tuned Y axis G320 - then tuned some more. Still missing steps in only Y axis, both other axis do fine. Continuous jog will work fine two or three times then do nothing except count the DRO - motor doesn't move then for no apparent reason begins moving. Happens both Y directions; + and -.
This is not making sense. The oscilloscope shows a signal from the computer during jog and into the Geko320 but this continues to occur. Should I replace the Gecko, the encoder again, the computer, my shop, wife, or just buy a bottle of Wild Turkey? This machine is all about getting it going and not doing anything. Anyone interested in a nice small NC machine real cheap?
Thanks,
Bill C.
-
Hi, Bill
Is Wild Turkey strong enough, Swap your X & Y Gecko's around and test, If it stays with he Y axis, You need to look at the wire-ring carefully.
Post your Mach3Mill.xml file if that's the Profile your using, C:\Mach3\ folder, Make a copy rename it to Bill CMach3Mill.xml.
Is your scope hooked to the actual Gecko terminals, Common, step & dir.
Chip
-
Bill, I had a problem with the connectors on a gecko once. Sounds similar to your problem. The wires were connected fine it was where the connectors made contact with the header pins on the drive. Maybe you can give it the wiggel test? Next time your jogging manually, try wiggling the black connectors on the gecko and see if it stumbles. I got lucky and actually saw it ark a couple of times.
Brett
-
Chip, The scope probes are held on both the step and dir on the breakout board and the encoder feedback to the Gecko. I'll post the XML file.
Brett, Didn't think of that but that's a possibility since those connectors had to be wired separate of the drives in order to see what I was doing. I'll check that.
I also need to check two other things: to make certain that a motor drive voltage is getting to the brushes when this occurs - bad connection through the drive chassis plug (Molex) or a bad armature in the servo.
Also, I did swap the drives between X and Y but not totally. Only swapped the plugs at the connectors on the drive chassis - sort-of inconclusive. Will repeat that once more but totally by physically moving the 320 under the plug connectors. That may tell the most. But, I fear that the servo motor has a shorted winding which results in a dead spot if the motor happens to stop on the bad commutator segment. These were new motors but Y may have suffered an over heat a while back when the gibbs were too tight and a nylock nut allowed the ball screw thrust bearings to tighten to the stall point. That blew a 3 amp fuse (at 57 VDC) a few times. I may have cooked a motor. It's amazing how strong these motors are....and the 320 to drive them. The ball screw was hardly movable by hand when I finally realized the problem. But the motor had been doing fine until recently.....
Thanks for the help!
Bill C.
-
Checked for voltage to the servo motor and it is not there when the motor stalls yet I see the step signal going to the 320 during the stall. Also noted that the signal in one direction when the motor is running is around +7 volts and about -1.5 volts in the other direction - the one prone to stalling the most. Narrowing it down to tuning or the 320 has bad drive mosfets in one direction (or both).
Bill C.
I'll go through the tuning again very carefully this afternoon.....Thanks again.
-
Chip, what did the XML file show? I know it showed that I had been adjusting the settings for the MPG because of a calibration problem that I have been seeing. It was set for 4 per detent and it should have been 400 per rev, but...It actually should be 1 PPR - 100 per Rev. but cannot get the 1 pulse - it is always 4 or 5 and CNC4PC says this isn't right....Arturo is sending another MPG.
Too many problems here at one time - computer is freezing on start-up, crosstalk on new computer port pins, Y axis intermittent, MPG not calibrating correctly, 401K is still plummeting, etc, etc. The President is giving stock market advice... Having too darned much fun over here!
Haven't gotten to the tuning yet.
Bill C.
-
Hi, Bill
Looked at your xml, In Motor-Tuning, Set the Step/Direction "Pulse's" to 5 and see it that help's, Gecko want's them at 3 or above, It gives them a little more time to see the step/dir. Pulse's.
Hang in there, In general I'm felling about the same way.
Chip
-
Thanks for that setting advice Chip.
I have something to be happy about - StarTech says send their PCI card back to them for replacement - "hardware issue". That makes me very happy - strangely happy. That's one down!
Bill C.
-
Happy that you're happy Bill..........Real happy! :)
RC >:D
-
UPDATE: At last I got around to replacing the encoder and then finally exchanging X and Y G320s and now X is missing pulses. AH HAAAA!
I ordered another G320 for an A axis addition a few days ago so will 'press' it into regular old three axis duty. Then maybe I can use this pile of iron instead of testing, setting up and replacing parts on it. Blah!
Thanks,
Bill C.