Need Help!- MR-J2S-40A, HC-KFS43, art project, servo tuning

[krazatchu]

Hi ppl..

I'm hoping to get some insight into tuning a Mitsubishi BLDC servo....
The servo is 400 watts, part number HC-KFS43... The drive is MR-J2S-40A...

The machine is a very large rotating arm that extrudes ABS/HDPE plastic..
It's part of an art project that is currently on it's way to Austria...
We ran out of time to get it tuned before it was shipped yesterday...
My project partner will be in Austria to setup the machine sometime later this week...
So I need to sort out the tuning before he gets there....

It has brushed dc servos on the x and z axis's... And the BLDC on the rotational axis...
The dc servos on the x and z are allright, I used the modified elmchan drive for them and there are no problems there...

At the moment, when I power up the system, it stays still, however with a bit of a movement or push it starts into a slow oscillation that keeps getting larger until the drive senses an over torque or an over position error and shuts down...

The auto tune function doesn't seem to want to correct the problem...
As well, with manual tuning I can seem to get it to behave...
It always enters the increasing velocity oscillation, it swings one way, stops, swings back even faster/further, etc...

As it's key to the success of the project we haven't ruled out paying a professional to tune the system when it arrives in Austria...

Any ideas? Attached are some pics of the machine...

Thanks!
Michael

[krazatchu]

Here is another picture for a more accurate idea of the size of the machine...
I really didn't want to post this image as I wasn't looking so good after more than 24 hours without sleep...

I'm roughly 173cm tall... The machine is a bit taller than me...
Also, I can dig up the inertia calculations from the cad program if that would help...

Thanks!
Michael

[krazatchu]

Hi...

Maybe I should be more specific....

Can someone tell me how GD2 is calculated?
(Ratio of load inertia moment to servo motor inertia moment)

Thanks!

[123CNC]

Probably won't be of much help, but as you get more desparate, it may point you in the right direction.

Looks like a new install, have you tried contacting the vendor or manufacturing on some drive specific tuning help?

Is it a standard PID control loop? Read up on the Ziegler-Nichols Method. It may give you a manual methodology and a little more background which might help you pinpoint which parameter is not autotuning all that well and you might be able to focus on the troubling variable(s) and manually adjust.

Good luck, looks like an interesting project.

[krazatchu]

Hey...

It's not a new install... We bought the drive and motor used....
So no software, even we had to find the proper connectors ...

It appears to be a standard PID function...
However, there are quite a few other parameters as well...
The GD2 parameter is prolly the tricky part...
(Ratio of load inertia moment to servo motor inertia moment)

When I set GD2 to a high level, the motor oscillates loudly (small movement, very rapid)...
When it's set lower the machine oscillates in ever increasing movements back and forth...
Until it gives an over torque error...

I'm thinking the belt drive is much too small and flexible....
And is therefore acting like a spring causing the oscillation....

I knew of the Ziegler-Nichols Method previously from playing with dc brushed servos...
Just didn't know what it was called...

Thanks~~
Michael

[macona]

You are going to have to tune manually. And its not going to be fun...

You have a horrible inertial mismatch here. These motors are intended for low inertial loads. I would be looking at PG2 more than GD2. You might also set ATU's fourth digit to 1.

Have you read the manuals instructions on how to manually tune the motors?

[krazatchu]

Hey...

The auto-tuning function using the pc software worked allright...
It gave me the clue that GD2 is inverse as to how I was treating it...
Originally I was trying high values but once it was set to 0 it tuned allright...

Once the machine gets back from the exhibit, prolly something will be done in the way of minimal backlash gearing ...

Thanks!
Michael