Hello Guest it is March 28, 2024, 04:05:10 PM

Author Topic: PID Tuning, one axis perfect, the other causing a headache.  (Read 6884 times)

0 Members and 1 Guest are viewing this topic.

A brief background,
I've started another thread talking about retro fitting a ROMI lathe with mach3 , which had Bridgeport EZ PATH on. I've used a CSMIO A type to drive the existing hardware, brushed DC servo motors with encoder feedback going straight to the CS module.
Last night I very quickly tuned the Z axis which is being driven off the existing servo amplifier 100v DC. Smooth, no hunting, repeats accurately, everything is great.
Today I received a new servo drive for the X axis as the old amp died. It's an Axor MCS 110 6/12amp.
After wiring it in I'm struggling with the tuning config. After auto tuning several times the best I can achieve still leaves very slight ocilation/hunting and next to no rigidity. By that I mean I can twist the ballscrew pulley from 11 o'clock to 2 o'clock with little resistance. The motor screw ratio is 1:1.5, 6mm pitch b/screw, but it is a proven system from the previous control incarnation.

I thought the axis using the 20 year old servo card was going to be the issue, not this one with the modern drive!
Any suggestions greatly received.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #1 on: July 21, 2016, 06:13:20 PM »
Sounds like you need to tune the drive/motor a bit better before attempting to tune the IP-A.
Hood
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #2 on: July 22, 2016, 05:11:09 PM »
Well, I've left the drive motor as it came configured from the factory. But, the drive was meant to handle +-12V rathen than the IPA's +-10v, so I think you might be right. I the event that I can't get the tune right with stock drives my other option is a drive from advance motion control.
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #3 on: July 22, 2016, 05:16:53 PM »
One more thing thats worth mentioning is that when I turn my feedrate overide up, the machine gets even more jerky. like a lot more.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #4 on: July 22, 2016, 05:18:23 PM »
Shame you cant find info on the drives but your other option may be to set the velocity in Mach to be what the drive would expect at 10v.
So say you had 1200rpm at 12v then you would expect 1000rpm at 10v, so if you worked out the velocity for that RPM the drive and IP-A should be equal.
Hood
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #5 on: July 22, 2016, 05:33:41 PM »
I did configure the machine that way. Didn't make much of a difference though. So you're pretty confident its a servo tuning issue?
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #6 on: July 22, 2016, 05:34:46 PM »
oops! I didn't realize I was posting in someone else's thread! Sorry about that.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #7 on: July 22, 2016, 05:42:35 PM »
Well I can't be certain but I have seen enough to know that unless you have a good tune between drive and motor then you are not going to get a good tune between the IP-A and the drive.
In fact I have always been meaning to manually tweak the drives on my Chiron as I reckon I could get better tuning in them and thus would be able to get better in the IP-A, just always seem to have other things to do so never got round to it yet.

Also matching the rpm for 10v in the drive and the equivalent velocity in Mach is also important. For smallish differences it may not harm much but recently a friend was having a hell of a problem trying to tune his lathe, couldn't work out the issue until I discovered he had 3000rpm for 10v and something like the equivalent of 1000rpm for 10v in motor tuning. What was happening was the IP-A thought it needed 1000rpm so put out 10v, drive saw 10v and thought 3000rpm was being called for so it shot off. The IP-A saw the overshoot so hauled things back but the drive thought it wanted 3000rpm the opposite way and so it continued, shaking pretty violently bacvk and forth. Set the IP-A/Mach to the correct velocity for 3000rpm and all was well with the world :D
Hood

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #8 on: July 22, 2016, 05:43:47 PM »
Having said that though, if the drives and motors were on the original machine then they should have had a decent tune so maybe that is not the issue.
Hood
Re: PID Tuning, one axis perfect, the other causing a headache.
« Reply #9 on: July 24, 2016, 05:13:40 PM »
Hood you were bang on, thank you.
The existing servo amp was obviously matched for its motor when the system was built. My new drive Axor was set ex stock and performed satisfactory with no load attached. But on the axis it was fighting it. There are pots for gain and differential, which can be set using there instructions with a scope.
However I have given it a good dose of gain, and then auto tuned again and it's responding lovely! No hunting, humming or heat generation. Response is smooth and crisp, but just a hint of overshoot when postioning at speed.