Servo is cheeping very loud...

[Spartan117]

Hey guys,

i don't know if such things can be helped with via forum, but i try

I have 2 Leadshine L7P-400 / EL7-RS400P drivers with 2 ELM2H-0200LA60F Servo-Motors mounted on my lathe:

viewtopic.php?t=11657

400W, 23bit opical encoder, 230V, pretty high-quality

My problem is:
No matter what parameters i use, default, softer than default, one-click-tuning, simple-tuning, or after a first manual tuning... my Z-axis is cheeping like crazy! It really hurts in the ears, and is bothering extremely.

Is here anyone, who is really good in such things and who can tell me, what else i could try to quieten this axis?
It seems like the sound comes from the motor and ball screw, maybe the direct-driven and relatively long ball-screw makes it even worse.
The cheep is exactly at 8kHz, and only comes up when standing still. As soon as i try to spin the ballscrew in any direction, the cheeping stops.
If i move the axis a few mm via WMPG or keyboard, the cheeping gets louder or quieter sometimes, but not every time.

Does anyone know what i can do to at least quieten this a bit?
Here is a short video with audio

Fiepen.mp4

(1.19 MiB) Downloaded 14 times

Thank you,

Marc

[Chaz]

Its normal. My Panasonics do the same. Reducing the rigidity setting (P in PID) might reduce the sound however then there is less holding torque.

[Spartan117]

I already checked other parameters, but that doesn't help at all. As i said, even on (very soft) default-settings this occurs.
I tried some more things, and i found out, the motor is nearly unhearable as long as it's not mounted!
Only when the motor is mounted to the motor-block/machine, that cheeping becomes that loud.
It seems like the machine is the reason, that amplifies that sound so extremely...

I built a little damping-case around the motor for testing, and it's a little bit better with it, but it's still way too loud.
The machines bed emits the sound, too, I need to dampen the surrounding-structure somehow

[suntravel]

JMC is also cheeping sometimes, A6 is silent.

But if the machine is running JMC is also silent.

With playing loud Metal on my sound system in the workshop, I do not care

Uwe

[WesM]

I think leadshine uses motion studio, so grab the motion studio manual and start reading up on resonance tunning. There should be an automated tool in motion studio for this, but you might have to tweak manually.

You are specifically looking to use the notch filter/s to reduce the high frequency feedback you are getting after using the resonance frequency analysis. You can typically setup multiple notch filters for various operational situations.

Yaskawa has some good videos on youtube for adjusting resonance/notch filter. Check out their channel, different software, but the theory is the same.

[Spartan117]

Thank you @WesM
I already tried the automatically activated Notch-filters. But as the frequency is at 8kHz, and the Notch-filters only cover up to 4kHz, it didn't help. MotionStudio2 didn't place any automatic-filter at all.
I also had a look into the "mechanical analysis" of MotionStudio2, but I only found a little spike at 110Hz, that doesn't match to the 8kHz-cheep, and that spike even didn't cross the 0dB-line at all.

I did see the yaskawa videos once, but maybe I just have to dive deeper here...
Thank you for your idea!

[WesM]

Does it make the 8khz cheep if you decouple the ball screw, but leave the servo mounted on the lathe?

Have you tried swapping the servos to confirm they both cheep when in the same situation?

[Spartan117]

Yes, decoupling from ballscrew doesn't change the cheep, that has been my first try.
Also swapping it doesn't change anything.
Only removing the motor from the bracket/mount does reduce the volume drastically. It's nearly unhearable when not mounted.

And yes, the other motor has the same cheep when mounted in the same place.
It seems like the x-axis with it's belt-drive and different mounting-method doesn't amplify the sound like the mounting-point on the Z-axis does...

I am trying to examine what parts exactly do amplify this frequency that much, to be able to dampen them somehow.
But that's not that easy as I hoped it would be, because this special 8kHz-cheep is barely locateable at all...
And decoupling it from the structure also isn't easy, as everything down there has to be very stiff, too.