Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: replicaprops on May 07, 2007, 06:43:24 PM

Title: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 07, 2007, 06:43:24 PM
Ive been running tests all day and im at a loss to figure out whats wrong.
I just installed v2.0 and am having trouble figuring this out
Here are the symptoms.
1. The z axis when traveling down 1" only goes down .995  . When I cycle it up and down, it always comes back to perfect 0 but never goes down below .995  .
     This exact .005 error is happening with each Axis. It always arrives .005 short of the mark, but always returns to zero.
2. No matter how much adjusting I do with backlash compensation all my circles are all slightly round boxes. When I de energize the servos and roll them back and fourth by hand, there is zero play.
3. When I run a 350 line gcode file, that has a final depth cut of .225, the actual depth is .350  .

If anyone has any ideas, please let me know.
Rob
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 07, 2007, 10:39:52 PM
sounds to me like you have some PID loop tuning that you need to do :(
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 07, 2007, 11:51:30 PM
what is that?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 08, 2007, 10:11:12 AM
what are you running for servo drives?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 08, 2007, 12:04:54 PM
geckos
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 08, 2007, 12:37:24 PM
Read your docs about setting your Gain and Dampen pots on the back side of the drives...

Also can you turn the shaft by hand with the servo on?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: fdos on May 08, 2007, 12:59:45 PM
If you're DRO's in mach are saying -1" and the job is ending up 5 thou short of 1" then I agree with Brian your servo tuning is off.  Or you are losing some steps somehow.   

If your DRO's read .995" then there's something else wrong.

Wayne....
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 08, 2007, 02:41:49 PM
So even though the drives have been running fine for over a year, you thing I need to tune the drives again?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: fdos on May 08, 2007, 02:56:21 PM
Well who knows!    If you tell mach to move say 1"  what do the DRO's in mach say?

If it's been working ok for a year and not now then something has changed surely?

In one of the cases you gave above the distance fell short, in the other it gained?

Wayne....
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 08, 2007, 04:25:43 PM
I was running a file and I noticed the z was droppng, so I sat and watched the servo and the servo was slowly lowering the Z. The DRO in mach 3 was remaining constant at a set level even though the servo was moving. I had an extra servo, so I swapped them out and it did the same thing. So I think the servos and the encoders are ok. With the servo turning and mach 3 not compensating, what does that mean?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: fdos on May 08, 2007, 04:33:06 PM
OK thats good information

Sounds to me like you have a noise issue, either getting into the Step input on the drive or the encoder cables.

Brushed spindle motor?   Could be brushes arcing a lot.

These noise issues are the reason I like to use differntial signals for both encoders and Step/dir signals.

In retrospect with this new information it's unlikely to be drive tuning.

Wayne.....
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 08, 2007, 04:34:11 PM
You have an encoder noise problem :( Do you have a scope to test with?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 08, 2007, 04:38:56 PM
Ok so if it is noise, how do I verify that. The only thing I have is a multimeter which I don't know if that will work. I don't know what a scope is.
Should I try moving the gecko box to a different location, or maybe re route the wires?
I made the wires my self and they are just wrapped with electrical tape.
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: fdos on May 08, 2007, 07:45:45 PM
first just try to determine where the noise is coming from.

Try running your program without the spindle running.

Is this a router?

Wayne....
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 08, 2007, 08:24:41 PM
The servos are actually loosing steps when I run them with out the spindle powered up. I wrapped the cables in foil and I can see a difference while running a program with the spindle powered and spinning. The servos dont move as much. Still though there is something causeing the servos to move even when the spindle is powered down.
On a side note, the computer looses time when Im running Mach 3. If I run a 20 min program the clock is 20 min slow.

Up untill 3 days ago everything ran perfectly for over a year.
If this were a new build I can understand tweaking required.
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 09, 2007, 10:36:36 AM
is there a way to track down the noise source?
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: fdos on May 09, 2007, 12:58:54 PM
Easiest way to find source of noise is to eliminate it by switching off electrical items which may cause interference.

Is there anything that may have changed recently?  Have you checked the condition of your encoder cables?   Look them over very carefully, ensure you still have good grounds without ground loops.  If in doubt try another cable.

Wayne....
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 09, 2007, 05:37:17 PM
I know it is coming from the computer, because I called Gecko and they told me to unplug the printer cable and see if the problem went away.
So I did that and the movement stopped right away. They told me to reformat the hard drive and install windows again. So I did that and when I turned the servos on and ran the spindle there was no movement so I figured the problem was solved.
When I installed Mach 3 the problem was back and just as bad as before.
There has got to be something to do with Mach 3 since it only happens when Mach 3 is on.
Over 20 lines of code the machine looses .25" on all axises.
I ran the motor tuning utility and it is not accurate. I told it to move 1 inch and it moved .995. Then it adjusted the motor steps to compensate and when I jogged the axis to zero it only went to -.086.
This is all 3 axises and the likely hood that all 3 cables, servos or geckos to fail at the same time is so remote.
I know there is no noise because the servos sit still when the printer cable is unplugged. When the printer cable is plugged in, the servos start to move, when the spindle is turned on they move faster. When the printer cable is unplugged, and the spindle is turned on, the servos dont move.

So what is next?
Rob
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 09, 2007, 05:55:11 PM
I just ran a experiment to verify the source of the problem
If the spindle is running and the printer cable is plugged in the servos move at approx .001" every 10 seconds
If the spindle is running and the printer cable is un-plugged the servos do not move

This experiment was done with nothing running on the computer. So some how the computer is sending a signal to the printer port even though mach 3 is not on. However I ran this experiment before I installed mach 3 and the servos did not move, so by installing mach3 there was some change done.

When the same experiment is done with mach 3 running, the servos move 3-4 times faster.
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 09, 2007, 07:19:03 PM
could the parallel port is bad?
I tried switching the bios to all the different parallel port configurations, but none fixed the problem.
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 10, 2007, 07:59:10 AM
I think you have a ground loop.. you are gettng noise into the step lines.

Could you tell us exactly what you have for hardware and pictures would also help. 
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 10, 2007, 12:32:24 PM
holy cow. I decided to video tape the whole problem and in the middle of taping it got worse!

Normally the servos stop moving when the parallel cable is take out. Now the servos move with out the parallel cable.

It only moves when the mill spindle motor is running at half to full speed. The movement increases in speed the longer the spindle is moving.
I started to play with the trim pots on the gain dampen and limit on the G340's and I have no clue what each of those do, so I let the mill run and only hooked up the wire to the z axis. So the only wires connected to the servo driver box is the z axis, the power cord and the 5 volts coming from the USB port.

So now there is a problem with in the servo box.
I have dc servos, g340 geckos, camtronics servo power supply
Here are photos, I dont know what you want photos off so here is the servo box and the placement of the equipment
(http://www.acmeartifacts.com/images/amill2.jpg)
(http://www.acmeartifacts.com/images/amill1.jpg)
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Lawley on May 10, 2007, 03:11:58 PM
I had a similar problem when I first started and eventualy traced it to the DC power supply that I was using to power the logic to my servo's (Panasonic) this despite the fact that I was using diferential signals for step and direction.

Brian
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: replicaprops on May 10, 2007, 04:54:27 PM
Right now im on my second cut test, but it seems as if the problem was in the terminal screws on the geckos.

In other words one of the screws was loose.

I don't know for sure if this was the problem, since I simply went in and tightened every screw just to make sure none were loose. Right now the mill is cutting perfectly. Knock wood.
I just wish I knew for sure what the problem was in the first place. I hate when something pretty much fixes it self and I don't know why. It leaves the possibility that the problem is not fixed.
Ill let you know if it continues to function properly.
Rob
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Hood on May 10, 2007, 05:08:39 PM
I have had (and so have a few others I know of ) bad connections with the plugs on the Geckos, maybe just pressing it down has made the connection again.
Hood
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 11, 2007, 09:03:11 AM
What do you have for a breakout board in the box?

I think you may need to have a Breakout board to get your voltage up  (Just an idea at this point)
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: fdos on May 11, 2007, 09:22:33 AM
A few suggestions for what it's worth.

Although you have a simple breakout board it's unbuffered and your PC seems to be along way from your machine.

I don't see any screened cables in that box?   Particulary your encoder cables ought to be screened even better use twisted pairs.

You also have many potential antenne in that box ;)   Well routed cabling goes a long way to eleviate problems.

Wayne....
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: klmark on May 11, 2007, 12:01:42 PM
Brain is right on the Breakout board I had a very bad noise problem Brain tooled me to add a Breakout board that took care of the problem I would suggest every one to use it
  Glad someone was able to help me Thank you
Title: Re: servos not traveling proper distance, new install of 2.0
Post by: Brian Barker on May 11, 2007, 09:30:48 PM
I think you can put this in the setip that you have now and I think it will fix the problem...
http://www.pmdx.com/PMDX-103/index.html


I would put this one in if it was me (there are many other great boards that you can use as well)
http://www.pmdx.com/PMDX-122/index.html

Thanks
Brian