Current Transformer in VFD application

[place1234]

Can a traditional Current Transformer be used in a VFD application? Is it capable to read correct current? For example , 10 HZ, 20HZ .
A traditional Current Transformer I 've used normally works in a 50HZ application.

 

[Skogsgurra]

It will probably not work at the lower frequencies. An iron core can only handle a certain volt-second "area" and when frequency goes down, the area gets wider and the core will saturate. It could work, if you have a very low burden, but it is usually not done that way. Use an LEM instead. Or a similar DC current transducer. DC current transducers work reliably all the way down to 0 Hz (DC).

There is also the problem with HF interference from the PWM voltage. It is not so much of a problem if all you want to do is sending the secondary current to an analogue instrument, but using it in a recording device or a controller usually needs quite a lot of filtering, shielding and grounding.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[OperaHouse]

You can always measure the current going into the VFD for a relative indication. The other problem is what do you measure the DC current transformer with, the complex waveform can cause errors.

 

[DickDV]

Bottom line---NO. Don't use a CT on the motor leads. Instead, why not use an analog output from the drive. This is almost always available and is far more accurate than anything you could hope to get from a CT.

And, with no extra hardware! I think that's called a "bargain", even in the king's English!

 

[ozmosis]

I would agree with DickDV. Use the current measured in the VFD. We just saved a customer about $15k (AUD) by doing just that. They were looking to purchase special PWM suitable CT's with adjustable current threshold to activate a relay when a defined current is reached. Now all done in the drive and the threshold is determed all internal to activate one of the on-board relays..

 

[Skogsgurra]

If I may speculate, the OP is probably aware of this possibility. There are applications where several motors are connected to one inverter and where it is desirable to measure current in individual motors. One particular case, that I remember, is a furnace where several pump motors are fed from one inverter and inverter is used to set total flow rate and it was desirable to measure current of each individual motor in order to see if nozzles were clogged etc.

Such cases exist and sometimes an ordinary CT works quite well if the operating range doesn't go down to frequencies where the CT core saturates.

As to the complexity of the waveform, I would say that it is the least problem. VFDs deliver an almost pure sinusoidal current to the motor windings with very low residual PWM traces.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[DickDV]

Skogsgurra, I know it's a dangerous thing to disagree with you but, as far as CT's on VFD output leads are concerned, I have seen many, have tested for accuracy many, and have found reliable and accurate none, even in the 50-60Hz range.

As mentioned above, I have found them barely acceptable where purely comparative measurements are to be taken but that is all. On multiple motor applications like you mention, comparative might be good enough depending on the situation but I don't see any mention of multiple motors in the original post.

 

[Skogsgurra]

Neither do I. I was just trying to understand why the OP didn't use the obvious solution.

I may expand a little on this VFD/CT theme. I was asked to go to a city named V---s and help a company named A-B measure output current from a VFD. I was surprised - after all, that company is well-known for excellent engineering and support.

They had done everything right. But they didn't get any consistent measurements from their CTs. They had even tried shunts with isolation amplifiers, they had tried LEM type transducers. And now, they were back to standard CTs.

Burden was low enough and no saturation. But still problems. It was only when we wrapped the cables with brass foil and soldered a short braid between foil and cabinet ground rail that we got any meaningful measurements.

So, most problems with using CTs on PWM outputs are, as I have experienced it, caused by EMI that hits the secondaries and (this was an application with a rectifier between CT and burden resistor) is rectified. The current waveform as such was not a problem in this case. But EMI was. And I think that is why there are so many problems with CTs on VFDs. And, of course, saturation when frequency gets low.

It is interesting to hear that you tested CTs and got poor results. I do not know what the final accuracy was in the case I cited. Are we talking a couple percent or tens of percent in your case?


Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[2dye4]

Shouldn't somebody make a CT with a shielded secondary. Shielded from electric field that is. That shield if available should be connected to the secondary lead away wires and then grounded at the CT secondary destination.
Also are there some VFD that do not use sinusoidal PWM. If so then the results might seem a little crazy.

 

[place1234]

sed2developer? is the current reading from VFD accurate? By the way, is the current a calculated value of the VFD or a measurement value?

 

[Skogsgurra]

Good idea to-die-for.

But I would keep shield connection separate and probably ground it as close to CT as possible. Long wires are no good HF conductors.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ozmosis]

place1234
I guess it depends on your accuracy requirements but yes, the measured value in the VFD is typically good enough for itself so would be good enough to provide as an output.
It depends on your VFD type but most these days can 'connect' the measured current to an analogue output or use the current as a monitoring function and define internally thresholds to activate a reaction (via a relay or other switching device).

 

[ScottyUK]

skogs,

Some, if not all, the LEM transducers suffer from E-field coupling when they are installed on an unscreened cable at a non-earthy potential. The easy solution is much as you described.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[DickDV]

skogs, the errors we see are often in excess of 10% but you bring up a very interesting issue of EMI. We had not considered that at all and it seems very plausible.

Next time I run into one of these CT's, I'm going to try your shielding technique and see what happens.

The testing we did was on an ABB ACS600 so the PWM algorithm was not necessarily sinusoidal. That may have been part of the problem too.

 

[Skogsgurra]

If you look at the DTC, it produces a very close to sine current shape. It does so by switching so that the flux vector moves between two concentric circular tolerance limits. At high motor speeds, the current may get distorted (not seen that, but been told).

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...