Problem- Servo stutter on 2 axis Millpwr mill

[pontiacg5]

The Y axis on this machine has started stuttering pretty badly, and I'm not sure where to start to fix it. Here is a quick video of what it's doing

http://www.youtube.com/user/pontiacg.../0/spFBKkSmHTk

The X axis does this too, but not anywhere near as bad as the Y axis. It seems to be random too, sometimes it's not as bad and other times it gets a lot worse.

Are these just out of tune or is it possible that the linear encoders are not working right? The DRO seems to be working perfectly fine, though I have not measured to see if it is still spot on.

I have no idea where to start with this, but it is bad enough that it's useless for any work so I really need to get it fixed!

Thanks!

[EPTech]

This looks like jitter on the encoders. Check whether they are dirty. Cleaning them is tricky business, as is disassembling them. Have you changed the machine parameters recently? If the gain is set too high, the slightest error fed into the error amplifier of the machine may cause unwanted servo movement.

[pontiacg5]

I don't think the servo's actually have encoders on this setup, I believe the system uses the acu-rite linear encoders for feedback. I would be a bit nervous to tear into those things

I haven't changed any parameters, but I'm not sure if anyone else did. From what I've been told it's been doing this for quite a while. The software was just recently reloaded though so all of the parameters should have been reset to the factory settings I would think.

I did find this though,

http://www.acu-rite.com/CMSFiles/Fil...20Settings.pdf

I would like to check the settings on these pots, but this mill is backed up against another mill so close I can't get in the control panel! I can't really see a pot changing it's setting on it's own though, but they say it's been doing this forever so who knows.

This mill is running an old copy of MillPwr, 1.1.0 to be exact. There's no automatic tuning in this copy like the later versions either.

I guess I'll search around for a teardown of these linear encoders, Id really like to leave them alone if at all possible though.

Thanks for the tips!

[pofo]

Correct, they do not have encoders on the motors. I've torn down all three motors on mine. Inside is a Servo Dynamics amplifier and a custom board made by Acu-rite that has a power supply and a small control processor.

[EPTech]

Then I would leave the encoders as a last resort. The parameters are an uncertainty since the factory defaults are not necessarilly the right settings. There are several parameters to look into: the absolute gain, the loop gain, the in position value, (acc/dec time constant). Write down the original parameter, better not to change more that one at a time, keep your hand on the EMG stop because the servo may run away when you play around with these parameters.
Lower gain paramters may fix the problem.
If the in position parameter is too low, ex beyond the resolution of the encoders, the axis will never get in position.
If the acc/dec times are too low (combined with too high a gain), the axis will overshoot.
Sometimes the servo amplifiers themselves also have a gain setting (potmeter) This is mostly the case when the servo's have an analog tacho generator. Too high a gain here would cause vibration (humming) in the servo's but is does not tend to cause stutter as slow as your's.

Good luck.

[pofo]

Pontiac,

This is off-topic, but I would appreciate if you can tell me what the power supply voltage is going to the motors. There are two cables; one is the control and the other is the AC power. I was assuming the AC power is straight 120VAC, but if that is the case, then that would be converted to 170VDC. The motors are rated at 140VDC. So I'm wondering if they step down the AC in the console (they don't do so with the power supply in the motor assembly).
The power is daisy chained and on mine, the last motor on the chain has a screw cover over the AC output. That would be an easy point at which to measure.


I don't have a console. I purchased a bridgeport with an Acu-rite 3 axis retrofit, minus the console (destroyed as it came from a secure government facility). I'm going to retrofit it with a DSPMC controller and Mach 3.

Appreciate your help.

[pontiacg5]

EPTech,
I looked at the machine a little bit more before I went home and I couldn't find any menu for parameters or mention of parameters in the manual either, unless there is a secret set of buttons I have to push I don't think there are any parameters that can be set on the controller. I can squeeze the panel open wide enough to see around 3-4 pots on the servo amp boards, they are marked with something but I can't make them out. I'll try to grab the mill with a forklift on monday and move it so I can get to the panel.

I still don't think its the encoders because the DRO does not stutter at all, it works perfectly fine. If the encoders were the cause then I would think the DRO reading would be sluggish or inconsistent but it is spot on.

Pofo,

I'll bring a multimeter in on Monday and report back with what I find. I don't think my motors have a power supply though, all I could see when I opened them up was a screw terminal block for the 4 wires. I'm pretty sure this machine is really old, I'm finding that most MillPwr stuff doesn't seem to apply to this machine. I'll take some pictures for reference, it would be nice to know how different my machine is from yours.


I was messing around with the machine again before I went home and I could move the axises with quite a bit of force while the servo's were on. I could hold the axis off of its intended location by around 100 thou before I would get a servo fault error. The servo's would fight to get back to the intended location, but not as hard as it I thought it should be. I figured that might mean something was set way off, but I don't know enough about servo's to tell what!

Thanks for the help guys!

[pofo]

If the servo amps are Servo Dynamics like they are on my unit, then you can go to the SD website and it will tell you how to adjust the pots.

The first thing I look for problems like this is simple stuff. It is usually something like loose connections, worn cable, corroded pin or similar.

[pontiacg5]

So I forgot to bring my multimeter to work today so I can't measure any voltages but I already see a pretty big problem, my motors are DC brushed servos not AC. Here is a pic of the label, my flash took a little bit of data out though...

http://i879.photobucket.com/albums/a...g?t=1284401797

I got to looking around inside and I've got three pots, I looked them up on the datasheets that are in the manual and they are labeled Signal Input (gain), Compensation, and Balance. The middle pot (Compensation) is glued down. I assume that compensation = dampening, but what is Balance then? The amps are Servo Dynamics SDFP01525-17-261.

I also noticed something else that does not bode well for this system, when you rapid one axis the second axis will "bounce" as the moving axis decelerates. If they were not tuned properly could moving one axis interfere with another axis? I'm starting to think there might be more to it than tuning. I really don't want to just start turning stuff willy-nilly (especially if one of those is glued, there's probably a good reason it's glued and the others are not.)

I just need to hijack the power feed and do a full 3 axis CNC conversion

[pontiacg5]

Servo Dynamics was nice enough to send me the original manual for the servo amps, I'll post it here, hopefully it can help someone else out too!

http://www.scribd.com/full/37371178?...vjldvavid1p37p

[underthetire]

We have one like it here. Same problem. I will list what I know to have been issues with it.

Carbon build up inside the motor, remove brushes, blow out with air, inspect brushes.

Tach on motor getting loose, sig adjustment can help

Smoothing capacitor failing, thats the large one on the servo amp card, check the ceramic resistor on it as well.

And the number one problem......backlash. I know, sounds crazy. You have scales for feed back, but the motor will just slightly hunt in the backlash range trying to hold a tight position on the scales.

Also, ours has a very hidden air filter in the cabinet. We had some bizarre issues until that was replaced.

[pofo]

Originally Posted by pontiacg5 So I forgot to bring my multimeter to work today so I can't measure any voltages but I already see a pretty big problem, my motors are DC brushed servos not AC. Here is a pic of the label, my flash took a little bit of data out though...

Pontiac, I have the same motors.



Note that the AC feeds the motor assembly and inside the motor assembly is a power supply that converts the AC to DC. That is why I'm wondering if the AC is down-converted; if 120VAC is rectified, then that equals 170VDC, which is above the rating of the motor.

Here is what my unit looks like.

Originally Posted by pontiacg5 ...I just need to hijack the power feed and do a full 3 axis CNC conversion

That's what I'm doing.

Cheers.