X always = 25 ????

[murphy625]

I solved "most" of my ground loop problems and now have a new one.

My motor is disconnected from its load and will freespin in open air for testing.
I have 1 servo motor using the X axis. (encoder is on back of motor)

Starting from Machzero X, no matter what distance I tell the servo to travel, the camsoft interface X axis readout ends up at 25.xxxx . If I tell the machine !FEEDRATE 10:MACHGO 10 the servo motor will make rotations to 25.xxx
If I tell the machine !FEEDRATE 10:MACHGO 2 the servo motor will make rotations to 25.xxxx

Now I realize that the setup gear ratio and encoder counts per inch need to be entered correctly but that should have nothing to do with the camsoft interface X axis readout right?
If I tell camsoft to go to x=10 then the readout should say X=10 when it finishes its move right? (regardless of the actual move made).

So why does it seem to be going to the same point every time?? I am also getting an error "Begin with motor off?" after each move. I am assuming that this is caused by massive position error.
The encoder counts I am getting seem smooth now. When the motor is under torque, the encoder readout is very stable +/- 1 count.

Also, with the control wiring disconnected from the drive, the drive sits stable under torque. As soon as I connect the +/- 10v signal wires, the motor oscillates. (ground loop?).. So I disconnected the ground wire from the +/- 10v reference signal from the ICM2900 and the problem went away. Is this normal to have it hooked up with only 1 wire like this? I did a google search for Ground Loops and read about certain conditions when only 1 wire is desired. Any thoughts on this?

And why does this thing always want to go to 25 ??? Something strange here..

Thanks,
Murphy

[murphy625]

After removing the GND line from the ICM2900 to my +/- 10v reference signal at the servo motor, the motor only goes one direction.

My servo drive takes a +/- 10V signal with inputs marked as "inverted" , "NonInverted", "0 volts".
The IMC2900 gives me McmdX and Gnd.

If I hook it up the way I used to have it, I get oscillations in the motor. (Noise?).

I have tried to shield the +/- 10v signal cable at one end, then the other, then both. Nothing makes a difference.

Why oh why did this thing work perfectly 2 months ago and now is giving me so many problems.

Thanks for all the help
Murphy

[Al_The_Man]

What makes you think it is noise? It sounds like it could be because of tuning. Both analogue connections should be made, the oscillations probabally stop because you have disconnected one of the inputs. I go back to what I said previously and tune and trouble shoot using the Galil terminal program, and when you have good results, then go through the Camsoft interface.
Also you have to phase the encoder rotation correctly, this can be reversed at the ICM2900 by reversing the A & /A phases.
Al

[murphy625]

Al,, thanks again, you are the man.

How much is that galil program?? WSDK ???

Phase the encoder?? The servo has a resolver phasing program that runs program and you have to enter the number in another parameter. As far as the High and Low coming from each encoder channel, I am almost positive I have them correct. (A to +Max), (A- to -Max) , (B to +MBX) and so on.. I have checked the connections with a continuity checker (disconnected from the terminals so I dont send multimeter power into sensitive circuits) Everything checks out good... I even ripped out the CAT5 line I had for the encoder simulator port and replaced it with the real encoder cable. (No difference at all in operation)


Do you think it better for me to get the galil program, or a scope that I can use on anything? (Its been 20 years but would you believe it if I told you I used to fix oscilloscopes and spectrum analizers?) LOL.. Ya.. the airforce days...

What is bugging me most about all this is that this servo motor worked absolutly perfectly for over 1/2 hour of trial tests.. Back and forth over 20 feet. It lost less than 1/8 of an inch error after probably 500 feet of travel. (rolls on rubber wheels on a track). Now, I can not get it to move 1 inch correctly. (I cant even get it to move in one direction without it running away).


Thanks for all your help AL.. I dont forget

Murphy

[Al_The_Man]

The SmartTerm program is free, the WDSK is 195.00. The encoder phasing I refered to occurs when what is known as the rotation direction of the pulses is opposite to what the controller expects, in other words the controller sends out a particular analogue signal and expects the motor (encoder) to count either in the up or down direction, if it is opposite to what the controller expects this confuses the control and unpredictable things happen, and this can be reversed by swopping the A & /A chanel pulses.
Al

[murphy625]

I think the encoder phasing is correct.. (not 100% sure) but this unit was working perfectly before..

Anyhow, I have been trying to get this servo on-line and working again for the past 2 or 3 weeks now. I have re-routed grounds to a "star" format..(as close as I could get it).. I have switched servo drives, hooked up and un-hooked shields.. NOTHING IS MAKING A DIFFERENCE.. grrr.. (I'm getting so fed up with this thing I am about to scrap it out)

Look,,, this is not brain surgery.. When I first got this package, (Camsoft, Galil 1832, ICM2900, etc),, it all worked perfectly. I hooked it up to a small (300 watt) table-top test servo and servo drive and I quickly found out what an "Un-tuned" servo was and it took me less than 20 minutes to figure out how to get it tuned or at least working within the parameters I need. (which are very loose by cnc standards). The single phase servo package worked perfect over and over for months while I played around and learned the camsoft software. Then I hooked up to a 3 phase servo. What's so different right? Its just more power. Sure enough, after I was done wiring it, I put it into a test and it worked perfect. (after a quick tune). I hooked it up to my load and again it worked perfect.
Then I decided to hook up the lifting part of my machine. Using a 1.5 hp motor and a 5hp VFD. I programed the drive to respond by way of the OPTO22 relays and the 24volt control strip. Made it all work. Then, as soon as I went to test the servo drive and the lifting device, nothing has been the same since.
First I found my 5volt supply start tripping out, then I find the encoder simulation port on the servo drive sending ghost signals to the galil card.
The problems just keep adding up. Solve one, another pops up.
I have a feeling that VFD screwed up the galil card or something.

I'm at my wits end... I knew this was going south when we started talking about grounds and ground loops.. Something just was not adding up but it almost made sense because my control box is powered by 2 different power supplies. I use 110V from the wall to power all controls and I use a 15hp rotary phase converter to get 3phase to the servo drive.
I thought perhaps there is a ground differential or something happening because of the fact that both the 110volts for the control side, and the 480 volts for the power side were coming from the same 240volt supply. I know it sounds weird but it looked good on paper and when I originally wired it all up, it worked like it should.

So I spend $175 on their software to find something wrong so I can send the card to them and spend more money having them fix it. (why does that not sit right with me?)
This card has already had 2 other malfunctions that were covered under warranty. (but I guess I should make sure its the card before I rant about that)

Thanks for the suggestions AL and everyone else who has helped me in these last few weeks.

Murphy
(Everything that can go wrong, has gone wrong)

[murphy625]

Also,
What would cause the camsoft readout to go to 25.0XX inches no mater what distance I tell it to go.

If I tell it to go 1 inch, it goes 25
if I tell it to go 10 inches, it goes to 25.

(this is something its never done before)..

[HillBilly]

I would check with the drive manufacture for phase converter operation. In most cases there is a single phase control transformer inside the drive that supplies the drive logic circuits. It is desirable to have these two lines hooked to the original L1 and L2 that is 180* out of phase (Not the sythetic L3 which bounces somewhere in between.). I take it that you have your step-up transformer, 240V to 480V, on the output of your phase converter (This could add another dimension to your sythetic L3.).

X always = 25 now that is strange! Does the CamSoft LOG.FIL show you anything?

Darek

[HillBilly]

Oh yea, the same should apply to your VFD if you are running it on three phase. You will also want to check the VFD manufacturer for analog input isolation (Some VFD's are only setup for remote pot operation, no isolation.).

Darek

[Al_The_Man]

Originally Posted by murphy625 Look,,, this is not brain surgery.. So I spend $175 on their software to find something wrong so I can send the card to them and spend more money having them fix it. (why does that not sit right with me?)

Actually brain surgery is probabally easier

You only need the smartTerm program which is free...
What card do you have the 1700 or 1800 maybe?
Download the manual, if you don't already have it Look at step 9, page 34 and armed with about 10 of the two letter commands you can tune, command a move & interrogate to what is happening in a few minutes of operation. Have the VFD powered off and anything else that might affect the outcome.
Also with the drive enable off, If you turn the motor shaft in one turn increments and type a TP the encoder count should increment or decrement by the amount of encoder pulses/rev.
Likewise with the system enabled, do a simple move equivalent to say 10 revolutions (encoder count x 10) and check the motor for 10 revs, and again do an interrogation (TP)
Al

[murphy625]

Originally Posted by HillBilly I would check with the drive manufacture for phase converter operation. In most cases there is a single phase control transformer inside the drive that supplies the drive logic circuits. It is desirable to have these two lines hooked to the original L1 and L2 that is 180* out of phase (Not the sythetic L3 which bounces somewhere in between.). I take it that you have your step-up transformer, 240V to 480V, on the output of your phase converter (This could add another dimension to your sythetic L3.). X always = 25 now that is strange! Does the CamSoft LOG.FIL show you anything? Darek

HillBilly,
Thanks for the suggestion... I'm no expert obviously but I was under the impression that the step-up transformer would improve the signal on the 3rd leg. Are you suggesting it may be doing just the opposite? This is the reason I am using such a large (15hp) rotary phase converter. I figured the larger it is, the more "real" that 3rd leg would appear to a smaller drive. You make a good point about that internal transformer on the drive. I just can not believe that switching a couple of the 3 phase input lines would solve all these problems. (It just seems to easy! LOL) But I will give it a try.... I'd jump off bridge if someone told me it would fix this problem!!


AL,
I have "smart term" already installed from another problem months ago. (The nice folks at Galil walked me threw it and I ended up sending it in under warranty).
I have gotten real good at using that TP command. I scrapped my startup.fil and timer.fil and set the program to do a TP command and response every 100 milliseconds to variable 500 within the camsoft interface. (Only because I know how to use it much better than the galil program for the time being)

Ok.. now the strange part.
Part1: The encoder simulation port is set to send out 256 pulses per revolution. (I can set it for 256, 512, 1024 or 2048) When the motor shaft is turned one 360 deg revolution by hand, I get 1024 encoder counts as shown by the TP command. I dont know if this is normal or not because I am not educated enough to know what the signal looks like going into the galil card and what part of that signal its counting.. Its almost as if its counting the leading edge and the trailing edge of the pulse. Kinda strange that I am set for 256 pulses per revolution and its counting 1024... Each pulse has 2 sides (leading edge and trailing edge) and there are 2 channels. 2 x 2 = 4.. 4 x 256 =1024 ? it almost seems like its working becuase the numbers come out even.
Yes, when I turn the motor shaft the TP response counts smoothly in both directions. (counts up and counts down)

PART 2:
No matter what I type in, no matter what I set my gear ratio to.. the camsoft interface always reads 25.xxx after I make a move. (motor uncoupled from the load) Would incorrect encoder pulses cause this? I dont see how. The encoder pulses only tell the galil card where the motor is at right? So even if the encoder pulses are mucked up, the camsoft interface should still go to the correct position right? I tried to type in !feedrate 1:machgo 1 and the motor turned way more than it should have (not that it matters) and the camsoft interface went to 25.xxx.

I am using the Galil 1822 card and ICM2900 with OPTO isolation. All outputs and inputs go into OPTO22 boards also.

I will dig into it with the smart term program when I get back to the shop this weekend. I am going to check out those encoder pulses very carefully.

Thanks for your help AL and Hillbilly.
Looks like I will probably have to make a donation to this forum eh?
Murphy

[Al_The_Man]

Originally Posted by murphy625 Its almost as if its counting the leading edge and the trailing edge of the pulse. Kinda strange that I am set for 256 pulses per revolution and its counting 1024... Each pulse has 2 sides (leading edge and trailing edge) and there are 2 channels. 2 x 2 = 4.. 4 x 256 =1024 ? it almost seems like its working becuase the numbers come out even. Yes, when I turn the motor shaft the TP response counts smoothly in both directions. (counts up and counts down)

You assume right on the encoder count, it can be set to multiply by 1,2 or 4 by looking at the appropriate edges, It looks like you have confirmed the hardware side is OK, which seems to indicate an anomally in the Camsoft code somewhere.
Al