[semenov419]

Good day everyone.

Couple of days ago, we ran in to a problem, when trying to mill a cylindrical part (Aluminum, Ф 278, 8Kg) clamped in 3 jaw chuck on our Chinese TCK45H lathe equipped with power turret, and Mitsubishi M70 contrller. Here is the video
We have not worked with parts of this size before, thus I can not say for sure that it's either existing problem or a consequence of our previous issue..(here is a thread with a previous issue).

Pretty much, situation as follows: when trying to mill a part in C axis mode, machine will rotate it and at zero point it starts to resonate as in video. When applying force by hand it stops, so obviously it has to do with positioning, where motor tries to compensate by moving back and forth, but can not. When in hand wheel mode, after jerking is stopped, we can rotate the part in a very slow setting and even then it tries to resonate. Anything faster than 10X causes this issue. Some time later we have discovered that operator have not told us that when milling smaller parts (about 1-3 kg) do cause the same issue. Anything smaller that that weight is perfectly fine. This issue occurs only in C axis mode when milling, and makes it impossible to process parts we need. In all other instances spindle works great, and there is no positioning issues found. Manufacturer of the machine gave us several suggestions, and so far: we have tightened belts, checked for oil or grease in places where its not supposed to be, and played with the settings (parameter 13005, 13006, 2205). So far, no luck with solving an issue.

Thus, given a dead end situation with now solution, Now I am here with you, and really hope that someone can help solve this problem.

Thank you in advance guys, you were a great help last time, hope one day I will accumulate enough experience to give back.

P.S. And as for our operator of the machine - well he mentioned that when milling smaller parts (in C axis mode with rotation) that do not cause spindle to resonate, there is not much firmness on the spindle and parts do wobble. (little bit)

[spumco]

Somebody smarter than me will hopefully chime in, but this looks like the spindle servo is dithering. Either it's hunting for an encoder position and is falling in between pulses, or the PID settings are causing it to move between a couple of positions. The heavier the part, the more inertia it has when the servo 'corrects' the position - it then overshoots and gets corrected back. Looks like an improperly tuned PID setting.

If you haven't changed settings on the servo drive, then my first thought is that there's a mechanical problem with the spindle servo encoder. Perhaps you can check the encoder to see if it's mounted properly (clamped on motor shaft square, tight, and the anti-rotation tab isn't loose).

If you have changed servo drive parameters, then you'll need to discuss this with someone more expert than myself on what parameters (P, I, or D) to change to reduce dithering or hunting. You might get more responses on the CNCZone "Servos" sub-forum.

If you haven't changed drive parameters, and everything looks tight (including electrical connections), you still may have a servo drive issue. Your operator reported this has happened in the past - if you've only been milling small, light parts since day one, this may be a drive setting problem right from the factory that has only become very noticeable with the larger part.

Again, check the encoder first, then re-post in the servos forum if you want some expert help diagnosing and tuning the servo drive parameters.

[Jim Dawson]

I agree with spumco, looks like a servo tuning problem. I would bump the D parameter down a bit In the PID settings, If that doesn't do it, then bump the P parameter down a bit. Exactly how you do that on you machine I have no idea.

The P parameter is the Proportional gain, reducing this value will make the servo less stiff, and less prone to resonance.

The D parameter is the Derivative Constant, effectively increases the settling time when set lower. That is not a description of the function, but rather the effect, will generally smooth things out a bit when set lower.

EDIT: If the above doesn't work, try increasing the P and D parameters, P first.

[spumco]

Jim,

Does the frequency & amplitude of the resonance give you any clue - generally, and not just in this case - of the tuning needed? I've never tuned a servo, but I've seen them appear to vibrate at high frequency/low amplitude as well as low frequency high amplitude (as above).

Or is it like everything else in this crazy CNC/motion control world and utterly dependent on all the other details (backlash, encoder count, inertia, ad nauseum).

Just curious.

-Spumco

[Jim Dawson]

Normally yes. In fact after thinking about this one a bit more, I think it's overshoot rather than resonance. In that case my advice above is incorrect, ( I added an edit to my post above). You would want to actually increase the P and D values to make the system stiffer and decrease the settling time.

High frequency resonance is generally a sign of the P being too high.
Low frequency resonance is generally a sign of the I being too high, causing overshoot, or the P set too low so the system is too ''mushy''.

The best way to tune a servo system is with the tuning software provided by the drive or motion controller manufacturer. But I find that just changing the PID values and observing the result is pretty effective, and is normally how I do it.

[spumco]

Jim,

Thanks much. Hopefully the OP reads this.

-S