Has anyone else ever had the problem with motion stopping in the middle of a G01 or G02/G03 line? My Mill does it occasionally, and my Wire EDM is doing it nearly constantly. If I hit ESC and select Continue on Path, it will resume. It makes unattended burning out of the question.

It seems like the slower the commanded speed (the wire machine is running less than 0.1 IPM), the more often it occurs.

The Wire and the Mill are being run from completely different computers, so I don't believe it is a hardware problem. I have a Galil 17xx in one and a 18xx in the other.

Any ideas?

I had a similar problem thats almost the same when I first started. At very low feedrate the servo would cog. I was told that servos run better at faster rpm and my friend put a inline gear box on his motor. Now that problem went away. Your problem seems as if the feedrate is so slow that any resistance or drag slows down to a stop. A little push helps.

Carlo

I wonder if a higher resolution encoder would help? You have to consider that to maintain a low velocity, you'd have to maintain at least one encoder count per second (just a guess). Maybe that threshold is way too low. More encoder counts might help.

This may be related but I think Carlo could of just increased the P value and decrease the I value in the servo tuning settings. This stopped the motor jerk at very low speeds for us and may help run servos better at low rpm and slow feedrates speeds too.

Billy

Thanks everyone - as soon as Carlo responded, I realized what I should try - which is just what BillyBoy said. The servos and encoders are extremely high resolution and are OEM, but the drives are not. I believe it may be a tuning issue and I will give increasing the P a shot.

OK, I don't think servo tuning was the problem. I increased P from 125 to 500, and I and D as well. There is no change in the pausing. I still need to do some servo tuning, but that is not the current problem.

I did some more troubleshooting. I checked the voltage out to the motors and it drops off as the motion stops. The Motor Command signals at the interface board also drop off to near zero when the motion stops. Just to further check, I leaned heavily on the table and the Motor Command signals jumped up to keep the table in its current location - this is while in the middle of o program that had just paused. I hit ESC and Continue on Path and it starts off and runs again.

So far, it only seems to happen while burning is in progress. I seem to be able to dry run OK. Could this be some type of interference issue? Wire EDM generates quite a bit of electrical noise when it runs. I have everything shielded and have used all of the manufacturer's shielding as well.

Any other ideas?

If you switch the position screen to 'Distance to go' What is shown?
Al.

We have to say that we are impressed. Some of you are getting pretty sharp.

All of the suggestions made are valid and do come into play at varying degrees.

In essence what appears to be happening is that the slight resistance of the wire against the metal while burning at such a low feedrate simply make the machine run out of steam so to say. There's not enough velocity to counter the resistance. Voltage amounts sent to the amps are determined by gain settings which apply to a direct correlation between servo tuning and amp pot adjustments. So one solution is to get Galil's WSDK kit for tuning, If you haven't already done so.

In drymode there's no resistance.

In a perfect world you could afford gear reducers as already mentioned since servos do have the least torque at low rpm's. This is the opposite for steppers were there torque is best at low rpm's and drops significantly at higher cutting feedrate rates. Servos need a spinning action to maintain smooth motion. The higher the better. Applications such as Wire EDM that are highly accurate and slow burning need their servo's to maintain a higher rpm for the best cut quality.

Tech Support
CamSoft Corp.
(951) 674-8100
support@camsoftcorp.com
www.cnccontrols.com

I will also check the Distance to Go - that is a good thought. I have run with diagnostics on, and there are no clues there. It just shows the "GO" statement and nothing after the motion stops.

I do plan on doing some servo tuning this weekend using the WSDK software.

I'm still not sure I am following the explanation. While the machine is running the part, burning at about .1 IPM, it just stops. Not at a program step, but anywhere in the middle of the programmed motion. At this point, the motion command voltage is near zero, wheras while it was running it was fluctuating around .3 volts. I see this as the control telling the machine to stop. Is this not correct?

Now while I am still at this point with the machine not moving, but in the middle of my program, I lean on the table IN THE DIRECTION of the cut. Not only does the motion does not start again, but the controller motion voltage signal increases to RESIST the motion.

It seems to me that if it were a servo tuning issue, nudging the table in the direction of the cut would make cutting resume (as someone earlier suggested) instead of this motion being resisted to keep the current location.

I would think if the distance to go is registering that there is move left it should be putting out a command to the drive, and usually if the following error gets too great then it should error out. I can see it where the move is practically at the end and the command is too low for the amplifier gain to overcome the last amount due to inertia, but not in the middle of a move?
What amplifiers do you have? if they have a gain pot, Most have to be adjusted until the motor starts to oscillate and then back off 1/2 turn, this before any Galil tuning is done.
Al.

squarewave,

You're on the right track and basically the previous suggestions that all the good people gave, including servo tuning are still valid. On the other hand, we do not believe it is being told to stop on it's own. You would have to go out of your way to program such logic and would certainly be aware if this were the case. Instead it still seems as if it just ran out of steam, not enough force, rpm or velocity to counter the resistance. The gain settings have the greatest influence on how hard it trys to move. The value I represents how hard it trys to prevent drift. P represents gain or power force.

One word of caution. We suggested the WSDK. Depending on your CamSoft software version it would have to match the same version of drivers on the WSDK. Galil frequently updates these drivers to keep up with Windows operating system changes and released service packs. If the CamSoft and WSDK are not of the same vintage one or other would have to be updated.

Tech Support
CamSoft Corp.
(951) 674-8100
support@camsoftcorp.com
www.cnccontrols.com

I did some more troubleshooting. I checked the voltage out to the motors and it drops off as the motion stops. The Motor Command signals at the interface board also drop off to near zero when the motion stops. Just to further check, I leaned heavily on the table and the Motor Command signals jumped up to keep the table in its current location - this is while in the middle of o program that had just paused. I hit ESC and Continue on Path and it starts off and runs again.


I re-read this and it almost look like a program stop or feed hold situation if the controller is trying to hold the motor stationary.
Al.

It could also be a tool change pause.....you gotta look at the actual code.

Based on what has been described, we don't feel that it is out of position or tolerance. We also believe he would know it if a feedhold or new feedrate was applied. The new version shows real velocity on the watch window. The old version shows commanded velocity. Like Al said, the following error would kick the system into e-stop automatically. There are always exception such as the POSERROR is turned OFF or the move itself is already going slow perhaps on a approach of decel which would make it go even slower than commanded. It's not out of position, there's just not enough gain to force it forward.

Tech Support
CamSoft Corp.
(951) 674-8100
support@camsoftcorp.com
www.cnccontrols.com

. It's not out of position, there's just not enough gain to force it forward.
[/url]

The only thing that puzzles me about that is why would the motor buck any physical assistance?
Al.

We wouldn't say buck, but rather a better word is overcome.

Tech Support
CamSoft Corp.
(951) 674-8100
support@camsoftcorp.com
www.cnccontrols.com

I agree with Al's "buck" terminology. When I try to help the machine by applying a force in the direction it should want to go, the controller is RESISTING my force with an equal force in the OPPOSITE direction to keep the table still. If the motor had stalled, the controller should not be resisting my assistance, but should accept it and start moving.

A wire machine really only has miniscule cutting forces, since the wire never actually contacts the part. The high frequency arc is what melts/disintegrates the material and the water flushes it away. When the motion stops, the arc is immediately lost, so even that small force is goes away.

The amps I am using are the Advanced Motion Control units that plug to the back of the ICM2900 module. I believe it has a trimpot on it, but the manufacturer indicated that it was preset for "most" applications, and I have never tried adjusting it.

The amps I am using are the Advanced Motion Control units that plug to the back of the ICM2900 module. I believe it has a trimpot on it, but the manufacturer indicated that it was preset for "most" applications, and I have never tried adjusting it.

I use AMC amps of various models, is there a AMC model number on them?
Al.

At this point it's best to first determine if some thing is commanding it to stop, or it's plain old out of gas.

If you have a feedrate over ride slider bar, knob, function key or USB Joystick and can slightly increase the feedrate and it moves or you can program F1 instead of F.1 and it cuts through then you can determine if you should go looking for something stopping this or not.

Tech Support
CamSoft Corp.
(951) 674-8100
support@camsoftcorp.com
www.cnccontrols.com

Sorry it has taken me so long to get back here. I had some things I had to take care of, but I finally spent more time on it this weekend.

First, THANKS to all who replied - and answering some of the previous questions/comments:

The amps are AMC 10A8, being driven with 80VDC. I thought maybe an amp was shutting down due to excessive voltage (they are rated for 80VDC), so I lowered it to about 60V. No change in behavior so I put it back up to 80V.

Changing the feedrate slider will not restart the motion.

When the motion pauses, the cycle start/top icon does NOT change. Pushing the cycle start icon once or twice does not resume the motion. The distance-to-go is not zero when it pauses.

Now on with some current information:

My touchscreen has been acting flakey, working sometimes, not others. I felt like the USB connection was as secure as it should be, and maybe the connector was bad. I disconnected the USB cable (and my screen) completely and it didn't make any difference.

I loaded WSDK with the intention of doing some fine tuning on the servos. When running WSDK, I frequently get the "Time-out wating for a response from the controller" error, both when starting the program and DURING a tuning cycle. I also get this error occasionally when starting Camsoft, but have never seen it while Camsoft was actually running - so I didn't think it was an issue. Now I believe it may be. When it occurs in WSDK, it pretty well locks it all up and you have to kill WSDK with Task Manager.

My computer motherboard is a Soyo 845PEISA, which is one of the few P4 motherboards with ISA slots. My Galil card is a DMC 1760. I downloaded the latest BIOS from the Soyo website and installed it (mine was three versions old), but it has had no effect.

I am beginning to think it may be a problem with communication between the motion card and the computer. Windows XP reports that there are no conflicts and the device is working properly, so I believe it may be a problem with the Galil card. I'll try to talk with them tomorrow - and I have posted on their Online help forum as well.

Does anyone have any other ideas?

My computer motherboard is a Soyo 845PEISA, which is one of the few P4 motherboards with ISA slots. My Galil card is a DMC 1760. I downloaded the latest BIOS from the Soyo website and installed it (mine was three versions old), but it has had no effect.

I am beginning to think it may be a problem with communication between the motion card and the computer.

From my experience with Galil cards, I think maybe you are right, I have had good success with ASUS MB's, I believe you can still pick up some deals on Ebay.
Is it an overclocked MB or anything?
Al.

No, no overclocking or anything fancy. Just a plain 2.4 GHz P4. Does ASUS make a P4 with ISA slots?

So you think that it may not be the Galil card?

I have never used the Galil ISA version on a MB with that processor speed, I wonder if there is a timing issue, possibly some technology on the Galil card that has too slow a response time with that processor speed?
Al.

That could very well be, but I thought that it would be limited by the ISA bus speed and the processor speed wouldn't be an issue. I want the fast computer for the computations that I will be doing in the TIMER.FIL for modifying the feedrate based on spark gap voltage (which I am not doing as of yet).

I guess I can try the card in my old 300 MHz that has ISA slots. Maybe that would tell me if it a card issue or a computer issue.

Just a thought (not sure if this has been mentioned yet)

Check to see if you have a screen saver or windows hybernation enabled. These two modes may drain computer voltage to below 4.7V that the motion card needs to operate. These two modes simply consume too much power on the system.

intrusion (group)