VSD outout is erratic in no load test only.

[ahendawy85]

Hi all, I have a problem with a VSD at work thats been confusing.
I am a field service engineer for an oil service company, my line of work is in electric submersible pumps (ESP). The ESP motor is a 3 phase induction motor. Power here is 380V 50 HZ.We were using a VSD for motor control. The control scheme used by this VSD is constant V/HZ.
When performing a no load test the 2nd phase outputted a lower voltage than the other two.
For example when operating the VSD in no load at 50 HZ output frequency, measuring the output terminal phase to phase yields these readings:
U-V 270VAC
V-W 270VAC
U-W 380VAC
Obviously the V phase has a problem so I disconnect power and start doing some static checks on the VSD. I check the converter's SCR's and the inverter's IGBT's they all check out according to the VSD manual. Measured the DC bus and it does have the correct DC voltage value.
The confusing part is when I connected a small motor to the VSD and started testing it it checked out fine, readings are all balanced and the motor runs smoothly.
U-V 370VAC
V-W 380VAC
U-W 379VAC
Tried a no load test after and same thing happened, the V phase is giving erratic readings. And one last observation; the V phase has DC volts in it, when measuring with my AVO meter for DC volts at the output I get those readings:
U-V 80 VDC
V-W 80 VDC
U-W OL

I'm sorry if it took too long to explain, I just cant seem to find the problem. According to my static test results all components are intact. Any thoughts?

 

[LionelHutz]

You likely need to put a scope on the outputs to determine what the issue is. That way, you can confirm the switching patterns look OK.

 

[Skogsgurra]

There is probably no problem at all. Extra measurement components, possible snubbers, lots of second and third order effects. Doing measurements on a non-connected VFD is quite meaningless.

Also, if you use a meter (RMS, d'Arsonval with rectifier, moving iron or whatever) you will not get any meaningful readings. You either need to use a filtered meter or a spectrum analyzer to see the fundamental voltage. And it is only the fundamental that counts. All the other components (harminics and HF) do not contribute to motor torque.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[ahendawy85]

I am using a filtered meter,(Fluke 87), and I also measure from the output of a Sine wave filter so there shouldn't be any distortion there. What's boggling is that any measurement with the V phase in it is erratic. Once load is connected all is well. These equipment take hours to assemble and are pretty expensive, we can't run the risk of running a motor in a well 8000 Ft deep underground just to have the VSD malfunction on us. I don't know, other VSD's don't seem to do this.

 

[Skogsgurra]

I am having a Fluke 87 in front on me on the desk. It is not filtered.

What kind of sine filter are you measuring with? A motor filter or a 1 Hz filter? Motor filters do not give the right value.

Run another "non-well" motor for a couple of days and see if it is OK. That would be the simplest test. As I said before, measuring a VFD without load is not a good idea.



Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[biggeorgetechnical]

ahendawy85

You will likely need a Fluke 289 or equivalent, this has the low pass filter. I agree with Skogskurra. To make any judgement about the output transistors, my advice is to use a suitable oscilloscope with appropriately rated probe(s)

George

http://www.edeinst.com

 

[ozmosis]

ahendawy85
what type of VFD is it? If it is a type with a low DC link capacitance then hunting can occur at very low loads. It can be a combination of the PWM system switching of the IGBT's and the fluctuation of the smoothing capacitor voltage due to the lagging current of the motor.
I have seen this happen first hand. It is generally nothing to worry about. As soon as load is applied, it stops.
Have you carried out a motor measurement function from the drive? If yes, has the sinewave filter been in circuit? if not, you need to add this and also the fact that you will be connecting 8000ft of cable and the impedance this will bring.

Also, you can get a phenomena called beat when operating the motor at the same frequency as the supply frequency. Try running at lower frequency or slightly higher and see if it cahnges.
Submersible pumps will be a low impedance winding (typically) and this will have a very different make-up to a "standard" motor that I guess the VFD has already built into it's algorithms to measure the motor circuit. This could also have an impact.

One other, whilst I think of it. Is the carrier frequency (switching frequency) the correct one for the Sinewave filter? Some manufacturers expect a higher CF than normal when using a sine filter as this improves the performance. Check with the drive supplier if this is the case. If you do and then set this and try again.

 

[ahendawy85]

Ozmosis

Thank you for the reply. Its very true what u say, once load is applied the imbalance disappears. However this is not only at supply frequency, it's all along the frequency range. I did measure the motor and yes I am using a sine wave filter with motor load of course.
I am interested in your first point though, about it being a combination of PWM system switching and capacitor voltage fluctuations. Why is it happening in one phase though? And does that mean that I need to replace the V phase IGBT's and DC Bus capacitors?

Thanks for your help.

 

[Skogsgurra]

If you are still worried, then bring in a guy with the right experience and equipment to do a thorough investigation. I think we have reached the limit for what EngTips can offer. If you need more, you have to find a suitable EE to do the job for you.

My tip stands: Nothing to worry about. There are so many second and third order effects that you just have to accept what you have.

BTW, what did the waveforms look like on the scope? Pure sinewaves? No? I thought so. Very few sine filters, if any, output a sinewave.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[ozmosis]

My first point was more focused on VFD's that are low DC link capacitance, there are a few out there.
Your IGBT's typically either work or they don't. They are just a switch. I do not think there would be anything wrong with your switching devices if the VFD is not providing any error codes. Also, if there was a problem with DC bus capacitors, your VFD would indicate an error as well.
As mentioned by others, it is likely that there is nothing wrong other than no load.