Need Help!- Noise problem, induced by VSD-E?

[lfleiva]

Hi everyone!
Does someone recognize this noise?

It appears as soon as I enable any one of three VSD-E drives. If the drive is disabled (blinking blue) it disappears.

It is induced by the VSD-E drives to everywhere in my cabinet lines. If tried disconnecting everything except one drives and it is still there. I can't why is it periodic (1 to 5 seconds). Besides this everything else works fine.

If tried with a different DC power supply and different +12V source also.

The nois burst seems to start very low frequency and resonate and then disappear until coming back again some seconds later.

Here are the pics.

Has someone experienced something similar?

This is my controllers +5V output:


This I took creating a loop between gnd and oscilloscope tip. (Purely induced)


Thanks for any help!
Luis

[scrambled]

Luis,-

I had the same type of problem, and I tried EVERYTHING. Here are some of the items I tried.

1. Installed a Capacitor on power connector
2. Installed a capacitor on the encoders
3. Installed a ferite core on the USB programming cable
4. Rerouted the power cables

In the end, it was a bad capacitor on the Z axis. I am sure Tero will chime in later tonight.

Are you using the VSDEPI? If so, you have to disconnect the CMD connector inorder to communicate with the drive.

Is everything else working ok?

Can you get movement out of the axis?


Steve

[Xerxes]

Yes, its probably generated by drive but its mostly emitted by wires and motors. This is what you will see in oscilloscope when using any PWM drive. I have ensured that VSD-E is not emitting more noise than typical PWM drives do. Actually most drives I compared output more noise.

The PWM signal is square wave voltage going to motor where signal edges are very steep (slew rate up to few kV per µs) and it easily couples capacitively to other conductors including oscilloscope leads. So its not necessary noise in measured circuit but rather in oscilloscope leads or in oscilloscope itself. It's totally harmless if there is no problems with any devices.

However, as the most noise is actually escaped after the drive, the best ways to reduce it are:
-Use shielded motor power cables and connect cable shield to drive Frame Ground (FG) (recommended)
-Connect motor frame metal to FG (recommended)
-Use shielded cable also for HV supply from PSU to drive (usually not necessary)
-If still want more shielding, add an edge filter after drive. It will reduce the PWM signal slew rate and greatly reduce emitted noise. (usually not necessary)

[lfleiva]

Thanks for your replies scrambled and xerxes (Tero?).

This noise burst is not the 40kHz (aprox) noise from the pwm because it is not constant and has much much lower frequency.

Scrambled wrote:
>In the end, it was a bad capacitor on the Z axis.

Which one? I did have an incident which ended up with a damaged drive in which I finally had to remove one side of the mosfet bridge. It is possible that other the drives got affected also.

>Is everything else working ok?
>Can you get movement out of the axis?

Yes, actually the machine axis works ok. But I/O are being affected by the noise severly. Also, if I have one drive enabled and other disabled the disabled drives will jump/move a bit synchronized with the noise burst. (I assume by noise into the encoder line and drive trying to compensate)


Xerxes wrote:
-Use shielded motor power cables and connect cable shield to drive Frame Ground (FG) (recommended) -> it has
-Connect motor frame metal to FG (recommended) -> it is connected
-Use shielded cable also for HV supply from PSU to drive (usually not necessary) -> don't have, but distance is short (<30cm)
-If still want more shielding, add an edge filter after drive. It will reduce the PWM signal slew rate and greatly reduce emitted noise. (usually not necessary)

>Noise is generated by drive but its mostly emitted by wires and motors

Agree.

Right now my guess is a ground loop may be amplifying the noise produced by the drive as xerxes suggests and somehow creating some a positive feedback resonance, so suddenly the noise increases out of control.

I found a large ground loop due to 12V PS chassis and also found to have frame ground connected to motors PS ground at one point which seems not correct. I am working on checking for other ground loops and correcting this couple thinks and will reconnect to see what happens.

Will feedback what happens.

Regards from Costa Rica!

Luis

[Xerxes]

Oh, I missed the time scale.

My best guess then is that on every noise peak drive is correcting single encoder count position error which induces a peak in current draw. Do motors make any sound when peaks occur?

"But I/O are being affected by the noise severly. Also, if I have one drive enabled and other disabled the disabled drives will jump/move a bit synchronized with the noise burst."

Hmm, do you mean that VSD-E I/O's are affected by that noise? Do you have the flat CMD cable plugged on drive? As you told some drives are disabled, did you do that by unpluggin CMD cable from BOB? If yes, then a freely floating cable may act as antenna and pick enough noise to affect drive I/O.

[scrambled]

Luis-

You need to talk to Tero about the capacitor. He is the one who found it.

Are you using shielded cables? I had to use shielded cables both the encoder and the motor leads. Make sure the shield is connected at only one end.


Steve

[Xerxes]

Luis, are these I/O's you're talking about some other devices than VSD drives? If yes, are those referenced to same GND than drives?

If yes, then the problem really may be grounding. As drive draws high current from power lines, it may cause quite high voltage drop in ground cabling, which may be seen as invalid logic status in I/O of other devices. This kind of problems are typically cured with star grounding scheme. And if nothing else helps, then optoisolating signals should help as last resort.

[lfleiva]

I reduced the noise enough to not affect the I/O lines. I think the main reason was having the transformers secondary center tap connected to frame ground being also motors power supply ground. But I also changed other grounding cabling, so I am not really sure if that was the main problem.

I still see the noise burst but it is much much smaller now and not affecting any I/O.

This is probably the best explanation:
>My best guess then is that on every noise peak drive is correcting single >encoder count position error which induces a peak in current draw.

Thanks for your help!

Luis