High Idle Current

[polaraligned]

I tuned the X, Y and Z and they are performing decent. I learned a big lesson that when you turn up the feedrate override, bad things happen. If I program a cut for 60 IPM, it will run fine. If I program for 40 IPM and turn up the feedrate to 150%, it will bust out into oscillations on multiple axis and the power load indicators will be flashing red. So when you are testing the servo tuning don't use feedrate override, and I won't be using it for a programmed cut either. It just doesn't work, at least on my machine, and probably is what has caused me grief in the past.


My W (the knee) is a PIA. Regardless of what settings I use for the PID, I get a high idle and dithering. I can tighten the loop and make my plots look great, loosen them and make them sloppy, or anywhere in between. It just doesn't matter. It idles high and I spent hours playing with the PID settings trying to tame this beast and have failed.

I have used this machine before my current "rebuild" for 8 years in a hobby environment and the motor has always idled high. Today I got a motor overheat error. I might not have had one in the past because the coefficient may not have been set in my old PLC. The motor runs warm when you put your hand on it compared to the other 3 that feel ice cold in my 58 degree shop.

The knee (W) motor is a Baldor 40 in-lb unit. MTE-4090 to be exact. It is geared about 38:1 and has an air assist in the pedestal. I set the max rate at 8 IPM because it gets a bit loud when moving faster, and 8 IPM is fine for tool length offsets. The motor is belt driving a leadscrew thru bevel gears and has to lift 1500+ pounds. It is very accurate positioning with near zero backlash. When I measured it's accuracy and repeatability, I was very surprised. The knee does ride on huge scraped boxed ways.

Here are my tuning settings. The forward values are pretty much from Autotune, though I did slow the accel on the W, and the PID values on all axis were set by me. Like I said, the W axis values don't matter in regards to the high idle best I can tell. I just keep the axis from oscillating and the result is the same- high idle and dithering.

If you watch the video below, you can see that as I move the knee up with the MPG, the current increases very little from idle. When I move the knee down, current decreases and stays lower. Watch the feed rate on the display. The other 3 axis are rock solid. They aren't dithering. So if the current indicator on the DRO is accurate, I am drawing nearly as much current at idle as I am making a move

Maybe this Baldor doesn't play nice with the DC3IOB? Anyone run into such a problem?

[tblough]

You need to tune your axis for the max feedrate you have set on the motors page. For the knee, I'm assuming "air assist" is a pneumatic cylinder counterbalance. Increase the pressure on that to reduce the idle current.

[polaraligned]

Appreciate your input Tom. I did tune for the max rate, which is a whopping 8 IPM, but in tune or out of tune, it doesn't matter- the idle current is nearly as high as the current when moving the knee.

I have the air counterbalance set for the max pressure recommended by Bridgeport. I rebuilt the unit 10 years ago and don't want to blow out the packing.

[cncsnw]

I am surprised that you are getting the "Idling too high" message when (last I saw) you had Parameter 63 set to a value of 40.0.

If the motor is not actually getting more than warm (say, not reaching 140F) then you can probably decrease Parameter 132 and/or increase Parameter 236, to see if you can get the estimated temperature to track better with the actual temperature. Press Alt-T to see the current estimated temperatures in the status window.

Is your W servo drive configured for 15A output, or 12A?

For the record:
1) Kg is the average percent motor torque required to overcome gravity
2) Kv1 (as originally intended and implemented) is supposed to be the additional motor torque required to overcome friction and move at slow jog
Therefore, if the motor torque exceeds |Kg|+Kv1, the axis ought to be moving.
3) "Idling too high" means that all of the following conditions are met:
a) No program cycle is running
b) The axis is not jogging
c) The average torque for the last 10-20 seconds exceeded (|Kg| + Kv1) * Parameter 63.

Since your W axis Kg is around 11 (absolute value) and Kv1 is around 18, any torque over 29% or so should result in movement. In your screen shot, it is sitting still with 38%. That is not too surprising, since in CNC11 and CNC12 Kv1 was redefined to favor velocity-mode control over torque-mode control, and therefore no longer means what the high idle algorithm assumes it to mean.

But Parameter 63 is supposed to provide a fudge factor to avoid nuisance warnings. With the default value of P63 = 1.5, you should not get high idle warnings until around 44% torque. With P63 = 40.0, you should not get high idle warnings until an impossibly high torque (i.e., never). That is why I am surprised you are getting that warning at all.

Perhaps someone from Centroid can chime in and tell me where my understanding of the algorithm and P63 has gone astray.

[polaraligned]

Good morning Marc.

I did not say that I got a high idle error, but I did say that I got a "Motor Overheat" error.

The motor is idling near the same current that is required for a move, and that is just wrong. The motor does get warm, and I can hold my hand on it comfortably, but the other motors are ice cold in comparison. This was after being on for about 4 hours while I was tuning- and the W axis ran very little during that time.

Thank you for the explanation of Kg and Kv1.

The servo is a DC1 and it is set for 15A.

Thanks for the Alt-t tip. I might take a look, but ultimately I would prefer not to have a motor sitting at idle drawing about 1/3 current. I do have a button set to M93 for times when I am not machining and want to quiet things down a bit.

I did see that someone said that low inductance motors will idle high with these drives, and the solution was to install a series inductor, but I would have to think about that one. The motor in question has an inductance of 2.5 mH, and the other 3 motors are 3.0 mH. Not a big difference. The "A" version of these motors has a notably higher inductance, and the biggest difference is that those have a voltage constant of 60/Krpm and the ones I am using have a constant of 40/Krpm.

So maybe I have a motor that just won't play nice, maybe it is bad or maybe there is a setting that I am overlooking....I don't know.

[cncsnw]

What happens to the idle current if, after moving the knee up, you move it back down maybe 0.002" before leaving it to sit still?

[polaraligned]

OK, I worked on the tuning after work today and by softening the loop a real lot I can get rid of the dithering.
It has some small oscillations in the moves and I had to be careful to keep it from busting out into big oscillations.

The weird thing is that when I try turning the shaft one way, the current drawn decreases to very low levels, and when I turn it the other way it increases. I would have expected an increase regardless of the direction I try to turn it as the servo tries to hold it's position. See video.

As for your question Marc, the video below answers it.

[polaraligned]

OK, when I do a positive move and I put pressure on servo pulley twisting in the positive moving direction, the current drops. Negative move the current drops twisting in the negative moving direction. The current will stay low when I let go too. Twisting in the opposite direction of move increases current as the servo fights to get back to where it wants to be.

The servo is not making it far enough on moves to put it in a happy low idle current spot. We are talking a handful of encoder counts here as I am not even moving the tenths on the DRO. Not sure if this can be made better by tuning or the servo motor has an issue.

The servo temperature increased 15 degrees playing with it for about 15 minutes.

This video should make it clearer what is happening.

[eng199]

It looks like there is some "stickiness" in the mechanism. The Ki term normally drives out the last remaining error, which I would think could get the mechanicals to their "happy place" where current is low. Your Ki term is very low. Did you try increasing Ki? Ki too high can cause low frequency oscillations, which is perhaps why you have set it low.

[cncsnw]

The Ki term normally drives out the last remaining error, which I would think could get the mechanicals to their "happy place" where current is low.

I respecfully disagree. Ki will indeed drive out the remaining error (provided the "Limit" term is high enough), but when the motor gets there, there may still be significant tension in the drive train. Nothing in the PID loop will necessarily relieve that.

With a heavy load, excessive friction, and perhaps some springiness in the drive train, there only way to relieve tension would be to go back the other direction a short distance.

Think about what you would feel in the hand crank if you manually moved the knee up to a certain position, slowly creeping in until you reached the exact target position, and then kept holding the crank handle in that exact position.

If you want to fix this (rather than just work around it and/or live with it) you need to reduce the friction, and perhaps also the springiness.