Hello Guest it is November 14, 2019, 09:06:00 PM

Author Topic: 2.5 carving - Z stepper creeping upward  (Read 1715 times)

0 Members and 1 Guest are viewing this topic.

2.5 carving - Z stepper creeping upward
« on: October 01, 2017, 08:10:08 AM »
So my father and I designed and built our own 2' x 4' cnc router:

http://www.openbuilds.com/builds/warnke-cnc-router.4383

All seemed to be well until we began to try our hand at a complex relief carving:

http://www.openbuilds.com/threads/suggestions-to-streamline-my-router-motors.10184/#post-58655

We are seeing an issue in the z motor where it gradually loses position as it works its way through the carving.  It is creeping upward (not downward) and it is a gradual consistent position loss - about .005" loss of position in z (shift upward) per 16" by 1" of carved area.

Some pertinent specs on all of this:

We are generating g code using an stl in artcam.  G code looks fine.
We are using Mach 4 as our router software.  
We have the PMDX-424 for our motion controller

http://www.pmdx.com/PMDX-424

Here is what we purchased for motors drivers and power supplies (these exact ones from this site, tho we added 1 more set to our order as we have a dual driven y axis and we intend to add a rotary axis later; so 5 motors total):

https://www.omc-stepperonline.com/cnc-kit/4-axis-85ncm1204ozin-nema-34-stepper-motor-and-driver-and-power-supply-4-ma860h-34hs46.html

We have tried a lot of different parameter changes, switching voltage settings, slowing everything way down, etc and nothing seems to make a difference.  It's just not holding position in z when it cuts this relief carving.

We did a test cycle where we just moved the z motor straight  up and down in repetition a bunch for a prolonged period of time and it held position.  It only seems to lose position when carving stuff like this.

The coupling is not lose at all (we double set screwed each side to each shaft - plus that wouldn't account for consistent z position loss upward anyway).

We are running a windows 10 PC USB'd to our PMDX board.

At this point we pretty much figure the problem is either Mach 4, the pmdx motion controller or the motor driver...I guess it could be the computer or USB? I don't know.

To add to the frustration, after a week of running great (beside the position loss) now when you tell our VFD air cooled motor to turn off in Mach 4, it doesn't stop right away...it just keeps going and gradually slows down whenever it wants bit by bit (sometimes not at all and you have to turn power off)..

Neither of us are electricians or controls guys.  We are machinists and mechanical designers, but the controls stuff was a first try rookie effort so it's a little frustrating stumbling around this all.  We just want everything to work like we need it to.

Any ideas on what this could be?  

Does anyone else run Mach 4 with the PMDX?  Anyone else see these issues?  Position loss in z or spindle not turning off when you push the spindle off button in Mach 4? (Spindle on button is working fine).


Thanks for any help.  We keep trying things to no avail and we are starting to get discouraged  here!

Online ger21

*
  • *
  •  6,288 6,288
    • View Profile
    • The CNC Woodworker
Re: 2.5 carving - Z stepper creeping upward
« Reply #1 on: October 01, 2017, 09:02:33 AM »
Quote
We are seeing an issue in the z motor where it gradually loses position as it works its way through the carving.  It is creeping upward...........

Try reducing the Z acceleration by 50%.
Gerry

2010 Screenset
http://www.thecncwoodworker.com/2010.html

JointCAM Dovetail and Box Joint software
http://www.g-forcecnc.com/jointcam.html
Re: 2.5 carving - Z stepper creeping upward
« Reply #2 on: October 01, 2017, 11:46:04 AM »
Hi,
I think ger21 is right, its an acceleration problem with the z axis.

The spindle control sounds a bit weird tho. I'm guessing/assuming that you are using a relay and an analogue voltage to control the spindle?
If that the case I think noise is getting onto you analogue control wire. I would be tempted to put a resistor of 5-10kOhm from the wire to 0V
and see if it makes any difference.

If there is a very noisy circuit within your machine its not impossible that the noise could be coupling into the Z axis and causing the drift you
mention. The fact that its limited to just the z axis and in one direction only points to a problem other than noise, but often pays to keep an open
mind and not rule it out absolutely but rather put it down the list of possibilities.

Craig
My wife left with my best friend...
     and I miss him!

Online ger21

*
  • *
  •  6,288 6,288
    • View Profile
    • The CNC Woodworker
Re: 2.5 carving - Z stepper creeping upward
« Reply #3 on: October 01, 2017, 12:52:56 PM »
You should have a relay controlling the VFD's run terminal, which would stop the spindle immediately.
Gerry

2010 Screenset
http://www.thecncwoodworker.com/2010.html

JointCAM Dovetail and Box Joint software
http://www.g-forcecnc.com/jointcam.html
Re: 2.5 carving - Z stepper creeping upward
« Reply #4 on: October 01, 2017, 01:07:24 PM »
This could be a Step signal polarity problem. Most drives will seemingly work with
either Active High or Active Low for Step signal polarity, but may actually sample
the direction signal incorrectly sometimes if the Step signal polarity does not match
the drive's requirements.

The effect may not be noticeable for jobs where the Z axis only changes at
the ends of long cutting passes, but be greatly magnified when doing a relief
carving because Z direction is constantly changing.

With the PMDX-407 spindle control (or with just the relay on the PMDX-424)
spindle control is done using the on/off control inputs of your VFD. You should
not be trying to use only the analog control voltage to control the VFD. If
you do have the on/off control connected, then I would suggest that you
try connecting a manual switch directly to these inputs of your VFD and see
if the problem still occurs. If so, then your VFD has a problem or is severely
miss-configured.
Steve Stallings
www.PMDX.com
For PMDX product support, please use PMDX forum or direct email for quickest response. We do not use this forum as our primary product support site.

Online ger21

*
  • *
  •  6,288 6,288
    • View Profile
    • The CNC Woodworker
Re: 2.5 carving - Z stepper creeping upward
« Reply #5 on: October 01, 2017, 01:25:30 PM »
I think Steve is right on the Z axis. I wasn't really paying attention to the fact that the error is so small.
Gerry

2010 Screenset
http://www.thecncwoodworker.com/2010.html

JointCAM Dovetail and Box Joint software
http://www.g-forcecnc.com/jointcam.html
Re: 2.5 carving - Z stepper creeping upward
« Reply #6 on: October 29, 2017, 12:58:45 AM »
Well we fixed the VFD issue (I believe the only change my dad made was the installation of a relay for vfd run terminal?).  Thanks very much for help there.  Something someone said tipped my pa' off though.  Spindle is no longer an issue.

The z creeping issue remains though.  He did mess with active high/low signal polarity and made sure requirements were matched.  

He did  reduce acceleration to see if that would help; and I think it may have helped a small amount?  The issue was still there though..

We tried everything with the basic steppers we had.  We had been considering  changing our our steppers with clear path motors:

https://www.teknic.com/products/clearpath-brushless-dc-servo-motors/clearpath-sd-stepper-replacement/

So we went ahead and did that today- I love these new motors (they are much faster and much more quiet, and they eliminated the need for the other stepper drivers we had mounted in our controls box), but to our dismay the exact same problem is still there, even with these servos.  We are still creeping upward in Z when relief carving.

It did not matter if we ran it at 75ipm or 175ipm.  We slowed down acceleration and it did not remove the problem....just...nothing worked.

At least we know it's not the motors or drivers now.  

Im a big layman here, so I apologize for sounding so crude in describing all this; it seems to me at this point that it's either the PMDX, Mach 4, our PC or something I'm doing wrong in outputting g code from artcam.  Either some Mach 4 setting we have wrong (but I'm told my father has tried seemingly everything), or maybe a computer/USB that isn't capable of feeding voltage signals to PMDX?  or .....something... I don't know what.  

We were thinking about trying to switch PCs out (our current one is a cheapee Lenovo touch screen) but I don't believe Mach 4 will  let you install on more than one comp; so we have to wait until Monday to contact them.

Wish I understood all of this a little better.  

Thanks again for the help.
Re: 2.5 carving - Z stepper creeping upward
« Reply #7 on: October 29, 2017, 02:47:15 AM »
Hi,
you can up to five licences active at one time with a maximum of seven licence re-issues in a two year period. Mach4 licencing is pretty generous, not open season
like Mach3, but still generous.

I think your clutching at straws though. Mach4 has thousands of users and probably hundreds with the exact combination you have, if there were a problem
with Mach4 or the PMDX-424 it would be known already. I understand that all the things you've tried have not helped and so you will try something else.

You've tried moving the Z axis up and down repeatedly without apparent fault. Yet when carving it does fault. Evidently it is something about the millions
of very small incremental moves that causes problems. I suspect therefore that if you were to repeat the up/down experiment over tens or hundreds of
thousands of moves the fault would show.

I rather think that some error along the lines that Steve has proposed still applies. I'm wondering if you could make some setting that made the problem worse,
way worse would be best. It would hopefully make plain the problem. Steve has suggested that if you had an active low setting for instance on the step
pin but the output of the PMDX is logic low at idle it could mean that the driver miss one microstep at each axis reversal. Over a period of time with the vast
number of axis reversals when carving the fault occurs. My proposal is to reduce the number of microsteps so that the problem is accentuated. I'm wondering
if the electronic gearing that you now have available courtesy of the clearpath servos could be used. The idea is to reduce the 'steps per' to some low number
but have the servo drives counteract that so the effective 'steps per' is actually higher. So an error occurs where the drive misses one step pulse from the
PMDX but gets amplified by the drive to 1000 steps, the error would stick out  like dogs balls!

Of course this idea may just be the ramblings of a nutcase...up to you to decide.

Craig
My wife left with my best friend...
     and I miss him!

Online ger21

*
  • *
  •  6,288 6,288
    • View Profile
    • The CNC Woodworker
Re: 2.5 carving - Z stepper creeping upward
« Reply #8 on: October 29, 2017, 07:48:31 AM »
Quote
I rather think that some error along the lines that Steve has proposed still applies.

+1.

Do like Craig says. Write a small g-code program that moves up and down, and have it run 1000 times.
Gerry

2010 Screenset
http://www.thecncwoodworker.com/2010.html

JointCAM Dovetail and Box Joint software
http://www.g-forcecnc.com/jointcam.html
Re: 2.5 carving - Z stepper creeping upward
« Reply #9 on: October 29, 2017, 12:46:33 PM »
Looks like we may have just solved it.  We switched motor settings back to active low instead of active high.  (Simple enough)

A little history:  with the first budget steppers we had we had this issue either way - in fact, if I remember right, active high ran better than active low, but it still creeped up throughout the course of the cycle.

With the installation of these new clear paths, we didn't think to change it back until this morning.  For whatever reason that seems to have solved the issue.  I set up a few different contour/relief carve cycles and it's holding position perfectly now.

...and there was much rejoicing...