Help Troubleshooting Problem R991H +/- 10

[jmullett]

I'm posting this to see if someone can help me through this very strange problem I'm having with a retrofit on my 2 axis cnc lathe. I've been beating my head against the wall few the last few days trying to figure this one out. I'm starting to think I need a priest to cast out some demons or somthing because I've tried everything I can think of. The long explanation beloy is for clarity please excuse me for being verbose in my post.

Anyone out there see somthing I've missed.

My Set-up:

(2) SEM DC brushed servos w/tach feedback driven by Westamp drives, BEI 1250 line (5000 pulse using quadrature) differential encoders 8830 line driver
-all the original equipment on the vintage 1987 lathe
-both drives run perfectly when given command signal from 1.5 v AA
battery and axis run smoothly
-both tachs give proper voltage and polarity when run with above signal

R990mb - Motherboard
-mounted with metal stand-offs onto the main control panel
-later revision supposedly not requiring resistors for SPI tuning

R990io - I/O board
-appears to be functioning normally responds to signals from Mach3

(3) R991H - +/- Analog Servo Drive (one is a spare)
-a bit flakey on connecting to SPI software one links up every time the
other 2 are often not detected or report "old version numbers" or
floating registers, all 3 however respond to Mach 3 Step/Dir signals
-I've used the same R991H for all testing to eliminate any possible
issues between different Rutex cards

2 GHz AMD cpu, 1GB RAM, Win2K OS

My Problem:

In SPI tuning software (DOS based bootable floopy from Rutex) I can see all the registers. Kp, Kd, and Ki change appropriately when I turn the pots on the board.

I can connect either encoder and see the pulse count change when I manually (with my bare hands) move the axis. The error count also changes obviously because the command output is not connected to the Westamp drive at this point. Also if I pull in the contactor for the servo amps there is no noise affecting the encoder signals (both are in shielded cable grounded on one end only)

Now I connect the analog signal from Rutex card into the Westamp servo drive. Z axis works perfectly. Can issue jog commands from SPI tuning software change speed go both directions everything is smooth, can even use the tuning routine and determine that a low Kp of about 100 is required to get no overshoot (supposedly). I can measure a command signal of about 7 volts out of the Rutex board when jogging at 75000 pulses per second.

I'm thinking great I'll be turning parts before lunch. So now for X-axis tuning I swap encoder and signal connections and power up the servos. X-axis appears to be correcting and fights me if I try to push on the tool turret. It fights me hard enough that I cannot get but two or three encoder counts out of it at my best according to SPI tuning software. So next I issue a jog command to see how this axis moves. I start out small and issue the default 100 pulses at 100 pps. Seems to move OK but at 25000 encoder counts per inch on my machine it does not impress me much. So now I issue 25000 pulse move at 25000 pps. Expecting it dart over an inch in just one second, I'm surprised to see the axis take 15 to 20 seconds to correct the servo error. I put a dial indicator on the axis and issue the command again and it indeed moved the appropriate distance. The servo error register runs up to 25000 almost immediately (or at least a quickly as I can hit the Regs. button on the SPI software) so I assume that it is sending the pulses in a timely manner.

So I measure the command signal and see less than 0.01 volts coming from the Rutex card. So I try in vain to bump the Kp value up on the Rutex Card to solve the problem. I had the thing jacked up to 30000 taking voltmeater measuremtns along the way still no increase in command signal.

So I think thak maybe somthings up with one of the Westamp servo drives, but a 1.5 volt battery ran and still runs the drives perfectly. I completley swap out the drives; X drive on Z axis and vice versa. Now I try again, The X still will not move but at a snails pace and still no more than 0.01 volt command signal from the Rutex card. Z axis works perfectly with either Westamp drive.

I've tried swaping encoders, I've run the X-Axis motor unattached to the leadscrew (no load) I've even measured identical tach signals from both motors when I am commanding motion from a 1.5 volt battery. I cannot understand this.

I even switch to Mach3 and jog the axii. Z is perfect, X is unusably slow and a command signal over about 0.01 volts is never generated from the Rutex card.

This thing is killing me. Anyone got any suggestions?

[MetLHead]

I'm not familiar with the Rutex boards, but I've seen this sort of thing before. Make sure you have the command signal from the Rutex to the servo amp connected properly. That is, make sure the 0v on the Rutex is connected to the 0v on the drive. If you have them reversed you will essentially be grounding the analog signal from the Rutex. If swapping this around causes the axis to run away, you will need to also swap the A and NOT A encoder signals to get it back in phase. THEN, if the axis direction is backwards, this will have to be changed in the control software.

Hope this helps,
Scott

[ViperTX]

Have you contacted RUTEX??

[jmullett]

Exactly MetLHead, like the signal for some reason shorts to ground and cannot source enough current (and thats a good thing) to keep the voltage up. I was under the impression that the signal coming from the rutex board was differential and not referenced to ground, but I do see a little GND logo by one of the command output terminals on the board.

I tried what you said and viola it seems to have corrected the problem. I'll have to do some more testing this afternoon and see how it works out. However it doesn't make any sense, I had the two command (x and z axii) signals wired exactly the same on the signal input side of the Westamp drives. I'm going to have trace back the original manufacturer's wiring I think they may have referenced the positive side of one of the dc power supplies to ground. I think the lathe was originally built in England, would they do something like this on purpose?

[MetLHead]

Originally Posted by jmullett However it doesn't make any sense, I had the two command (x and z axii) signals wired exactly the same on the signal input side of the Westamp drives. I'm going to have trace back the original manufacturer's wiring I think they may have referenced the positive side of one of the dc power supplies to ground. I think the lathe was originally built in England, would they do something like this on purpose?

Most drives I've worked with do not have a true differential input. One input is usually tied to 0v directly or through a low ohm resistor. This is one thing I always make sure to check as it _can_ damage the DAC output. If the axis runs away, I always swap the encoder signal instead of the analog.

Anyway, glad you got it working.

Regards,
Scott

[rutexus]

Gentlemen,

Thank you for supporting one another while I was on vacation. I'm back now, so I'll try to get caught up with all the emails etc. That two weeks was the longest vacation I have ever taken.

Tom Eldredge
Rutex LLC