[arsenix]

I put an oscilloscope on the power input to the drive cards. Quite a lot of noise on there, which appears to be around 60Hz. This seems to be hinting that I need to replace the big input filter cap. These aren't cheap but not too bad ($100). I'll do a little more poking around before I pull the trigger on it. I could not find a listing for the capacitors on the drive cards (3x 180uF 250V). As you say they seem to be a special type. Maybe a custom part?

[mactec54]

I put an oscilloscope on the power input to the drive cards. Quite a lot of noise on there, which appears to be around 60Hz. This seems to be hinting that I need to replace the big input filter cap. These aren't cheap but not too bad ($100). I'll do a little more poking around before I pull the trigger on it. I could not find a listing for the capacitors on the drive cards (3x 180uF 250V). As you say they seem to be a special type. Maybe a custom part?

There should be no problem to find the replacement Caps, can you post a photo of them, 180uf 250v is quite common cap, depending on type you can look here or other suppliers, even Ebay may have something.

https://www.newark.com/search?st=180uf%20250v

[arsenix]

Been some months and I've poked around at this but still not made any progress on it. The vibration remains. Otherwise machine runs great. Only manifests as slight surface finish issue.

Things I've now tried:
1. Replaced main filter cap, as well as the big caps on the Z axis drive. No change. I really thought this would be it but didn't do anything. The old caps all tested good anyway after I removed them.
2. Vibration occurs on all three axes, or at least Z and X (did not try Y). If axis is stationary but under load (ie you put torque on the motor pulley by hand) the response it not smooth it vibrates. Z is always under load so always vibrates, which manifests as ~.001" shake @ ~60hz. I also made sure it still happened on the X axis even when Z was unloaded (removed the belt after homing).
3. Looked at 5v going to encoders. It is ~4.6v. This is a bit low. I tried hooking up a separate 5v supply to the encoders. This made no difference.
4. Looked at the PWM signals with a scope from AXSBOB to the cards. The PWM signal is changing with the vibration (probably expected).
5. Looked at encoder signals from Z on a scope. I do have the Allen Bradley (Silver) encoder. They look extremely clean. Moved motor around by hand looking at encoder signals. I couldn't see anything amiss here. Slight movements cause it to give pulses. Don't see any double triggers or anything. Not sure what else I would see to indicate the encoder is a problem.
6. Also thinking about encoder issues previously I put a dial indicator on the Z axis while it was running (vibrating) for a few hours to make sure it wasn't drifting. I also did a bunch of moves to make sure it wasn't losing encoder counts.
7. Checked brushes on the Parvex Z motor. They look fine.
8. We went through all the parameters and all that seem relevant to this match the printout of the parameters that came with the machine, which I believe are from the install. I have not tried reducing the position loop gain parameter. There are some digital filter gains but they are set to the default in the manual.

I'm pretty stumped. Given that it impacts all axes the DC power bus seemed like an obvious place to find the problem, but swapping the big cap did nothing and I can't find anything else amiss. The feeder diodes for the cap seem fine. Not sure where else there would be a voltage drop in the system.

If folks have other ideas I'm all ears. I assume this isn't a normal state of operation for these machines right?

[arsenix]

At least temporarily I gave up on trying to troubleshoot this more. The machine works fine, the Z just rumbles a bit. I haven't noticed any adverse effects when milling. Surface finishes seem fine.

I do notice that when I was running some parts last night the machine gets "jerky" during adaptive tool paths. Maybe the tool path has too many lines/s than the machine can handle? I was cutting at only 40ipm. Not sure if this is a problem with the tool path or some limitation of the machine (maybe related to this issue).

[arsenix]

I did some testing. The machine is able to move smoothly at 100 and 200IPM. The "jerking" seems to be it slowing down significantly for the "turn around" at the end of the adaptive clearing passes. The lead in/outs generated by my software have a LOT of tiny moves, which may also be slowing it down a bit but these are fairly inconsequential.

It seems during these "turn arounds" it may just be hitting the machine's accel limits?

There are a few parameters that seem relevant:
Attachment 484806
Attachment 484808

My accelerations and decelerations are currently set to 4.5. The "profile selection" is 0, and "Residual Feedrate" is .998031.

No units are specified on any of these so unsure what they mean. 4.5IPM/s acceleration would be uber slow. It says "the amount of acceleration per iteration of the interpolator". Not sure what speed the interpolator runs at.

I attempted to increase the accels to see what would happen and appear to have broken the machine I assumed the safety's would avoid any machine damage. I immediately got a Z axis overcurrent fault when I tried running my program (initial rapid downward). Even after power cycling the machine the Z axis appears dead. When I turn the drives on the Z axis drops slightly and I get a following error. Disconnecting the motor from the Z axis drive doesn't change it. Fuse on the Z axis is still fine. Seems like I blew out the Z axis drive!

[arsenix]

After digging around I realized the motor was at high impedance and the driver seemed fine. Cable to motor ohmed out fine. Pulled the motor. Noticed that when I jiggled the motor leads the resistance was changing which was a giveaway there was a bad connection. This required disassembling the whole motor, which required fabbing a grooved locknut socket. Once I got it all apart the issue was very clear. A FACTORY DEFECT. One of the motor leads was not soldered to the brush holder and was just lightly pressed in. My hard driving of the motor must have heated that up enough to make it oxidize and go high impedance. Someone at Parvex owes me a beer. Long since retired I'm sure! I soldered it up and reversed all steps. Machine works fine again!

The smoking gun:
Attachment 484828

Soldered up:
Attachment 484830

The Z axis rumble is still present, although I would like to dream it is better after all this effort. It is still vibrating about .0001-.0002 which is very clear on an indicator. Maybe this is just normal for the machine. It seems to cut parts just fine but is irritating when I try to check tram on the head!
https://youtu.be/SAuk21fwMl4

[arsenix]

EDIT: I realized that I originally was seeing .002" vibration, and now I'm seeing .0002"! So in fact this seems to have dramatically reduced or even fixed the problem. There is still some rumble, but +/-.0001" vibration may be completely normal for this machine.

All of the solder joints in that motor were pretty low quality. Any owners of Bridgeport's with these motors may want to inspect this area if they have motor issues. I reflowed the ones that looked ugly in addition to the one that was completely disconnected.

[arsenix]

The machine still seems to do this periodically. It doesn't seem to impact surface finish. The Y axis also shudders a bit. I wonder if the P gains are just tuned too high? I have been working on a 4th axis addition to this machine as well, and one thing I noticed when looking at the PWM command to the A axis was that even with a static error (axis stationary, commanded position stationary) the PWM signal has periodic digital noise spikes on it. This does not make a lot of sense to me. This machine uses a special "PWMINT" card on the control card, but it isn't clear what the interface to this card looks like. Maybe it is analog and there is noise getting in here? Unclear. This issue still irritates me but the machine seems to run parts fine with it. It mostly concerns me it might get worse.

[servtech]

I made a program to move the Z axis up and down and listened next to the Parvex motor. No odd noises. No odd noises coming from the thrust bearing either. When stationary the belt is vibrating. I can feel it. The vibration goes away when I apply hand torque to the motor in one direction (my guess is less load?) and gets worse when I torque the other (more load?).

Since the encoder is on the motor I'm not sure this is likely to be the bearing? I suppose I could try removing the belt and jogging it around unloaded?

Check belt tensions using a gates frequency meter or other. Belt too tight or loose can cause axis vibration. Also, check axis isn't too tight, ballscrew alignment and slideway / rails etc. Do this before playing with drive gain parameters, don't mess about with drives until mechanicals are proven ok.

[mactec54]

The machine still seems to do this periodically. It doesn't seem to impact surface finish. The Y axis also shudders a bit. I wonder if the P gains are just tuned too high? I have been working on a 4th axis addition to this machine as well, and one thing I noticed when looking at the PWM command to the A axis was that even with a static error (axis stationary, commanded position stationary) the PWM signal has periodic digital noise spikes on it. This does not make a lot of sense to me. This machine uses a special "PWMINT" card on the control card, but it isn't clear what the interface to this card looks like. Maybe it is analog and there is noise getting in here? Unclear. This issue still irritates me but the machine seems to run parts fine with it. It mostly concerns me it might get worse.

Yes, the main Gain was my first thought, but you have done a lot to find the bad brush connections was a worthwhile find, noise can be a problem add some ferrite rings on those wires may help with the noise or twist the pairs together will help also, if it is gains then that should not change as with noise it will be the same