PDA

View Full Version : VFD problems



Peter_1
04-02-2006, 05:58 PM
Speaking of noise and troubles with VFD's. Those who followed my thread about the market truck with the VFD's: we had this week serious roubles with them and it's even not solved yet. Leak current devices are tripping all the time.
But yesterday, we think we found the cause of it and it's really a difficult theoretical case.

Now it comes, trying to explain this in English.

The earth leak current protectors (LCP) of 30 mA trips constantly due to an 'imaginary' leak, a leak it thinks there is but it isn't. It's a theoretical leak.

So I need some feedback from you guys in this forum regarding the fact that a VFD has zenerdiodes in it's electronic circuit to generate the sine wave.

A zenerdiode measures the 0-passage of the sine wave.

There is somewhere a theory that this device has a voltage drop of 0.6 V before it let through the current. This drop results in a DC signal.

This DC signal is present/transposed in the total sine wave signal.

Now, those who understand the function of a LCP, this device has coils in it who measure the flux of the voltage going in and flux of voltage going out.
As long as both fluxes are equal, everything is OK because what goes in must go out your motor (or any connected to the LCP device)
So, if they’re equal, there's no leak.

Now,... what we have and in fact with every VFD is that all what's going toward the VFD is a pure sine wave (with the common garbage on the net) in his pure AC-form but what returns is 99,5% (just a number) and a little bit of DC, but the algebraic sum together remains the same.

But the DC returning can't generate a flux in the coils of the CLP. Remember school: you need an AC current to generate flux in a coil.
So the CLP thinks that there is an unbalance and shuts of the VFD but there is no real leak. It's an imaginary leak.

Some questions now to look if someone is able to help me and has some experience in this special case.

Most questions are already answered by my son - it was he who found together with the experienced VFD tech I spoke with in the beginning of this thread the solutions - but I think many can learn from these problems which cost me al ot of my hear and some days of my life .

We measure at standstill no earth-leak at all (megger), nowhere...how can we measure then that there's a leak? (we have +/-50mA/VFD now, measured this morning)
Remember, the LCP is jumping as soon the VFD's are starting.

Do the members know how the measure this sort of earth leak (we don't measure any fault with a megger and in some cases like a VFD, it is even forbidden to megger)

We installed special EMI filters from Schaffner - the biggest European company located in Germany - and this didn't solve anything.
It was even worse.

What can be done to prevent this? We have no 2 working solutions, rather expensive; perhaps there are more we don't know yet. Eager to learn these.

Does someone know a good source for this theory, Fourrier analysis, the theory behind the conversion of some of the AC in DC which is then transposed on the main sine wave, the theory that a sine wave we all know consists in fact of many sine waves in a mathematical row (saw this long ago in school but forgot most of it), why distortion on the net has something to do with the uneven ‘members’ or factors in this mathematical row (if this is correct of course)

What's the difference between an EMI and an EMC filter and the theory behind it?

So, I want to learn about all this a little bit more and I think some of us can also benefit from this.

End of the day: earth leak is now 1mA/VFD :D :D

US Iceman
05-02-2006, 12:06 AM
Peter,

I tried to find some websites that may provide some help to your problems. This is not something I am very familiar with but the problem you describe sounds like a harmonics problem due to the VFD's.

The links below are those I found quickly which appear to discuss some of the problem you have seen.

http://www.mtecorp.com/load.html

http://www.ewh.ieee.org/r3/nashville/Meetings_Announcements/2005/Presentation_Files/Harmonics_MCPQG_IEEE_2005.pdf

http://www.processwest.ca/Current_Issue.htm?ID=330

http://www.lappusa.com/PDF/VFD-White%20Paper.pdf

Sorry for all of the English links, I could not find any in Dutch. :o

Andy
05-02-2006, 09:57 AM
Hi Peter:)
your problem has me thinking. Most airconditioning comming out now is inverter, all be it DC motors. Toshiba units use a single phase input.
Would this be less trouble some than a three phase inverter, maybe you are using single phase anyway.

We did a trailer for fruit ripening, this had a generator to power an inverter driven chiller.

Kind Regards. Andy:)

Peter_1
05-02-2006, 12:01 PM
Not allowed in a market truck in general and especially not in a fish vending truck. As soon water is involved, a 30 mA is compulsory.
We solved it temporarely this week with a 300 mA.

Problem is since yesterday solved, leakage now 1 mA.

What interest me now is the theory of the DC voltage transposed on/into the AC sine wave which is the cause of this imaginary leak.

Siemens has a special leak current protector for this(the 5SZ serie, cost 3 x the price of a standard type) - learned all this the last week - with the standard AC coil to generate the necesarry flux in it and additionally a device (can't be a coil because it's a DC signal) to take in account, or better said... to add the flux of this DC signal and add it to the AC flux.

Peter_1
05-02-2006, 12:12 PM
The VFD manufacturer proposed the Schaffner filters which didn't solve anything.
They came via UPS from Germany :eek:

frank
06-02-2006, 04:09 PM
Hi Peter

We fitted some VFD's in the summer and I was reading through the documentation for these Siemens Inverter drives. I thought you might find this of interest. Read section 6 for EMC & EMI definitions. http://cache.automation.siemens.com/dnl/zc4OTgxAAAA_1160430_HB/applications_e.pdf

There is what's called a DC Link between the incoming supply and the output circuits (page 8) and it seems that high frequency interferance on the output side can feed back to the incoming supply, as you are experiencing, if you don't have the correct filters on the supply.

One thing I noticed from reading the document was the need for good earthing and separation of cables (page 43). Does the mobile fish van put down an earthing rod when in use?

From your latest posts it appears that you have found a solution but it's always nice to read up on the fundamentals. :)

Peter_1
06-02-2006, 05:22 PM
Frank, Marc and Mike,

Thanks already for the replies.
The link of Siemens is very good, although the theory is sometimes very difficult to understand.

The Lappusa pdf was also very helpfully, written for someone with not all that electronic experience.
So.. like me.

Installing the special CLP's was not an option: they cost +/- 350 €/piece and we had to install 9 of it.

We solved the problem with a safety transformer instead of an autotransformer (don't know if this is the correct English word for it). They cost +/- 450 €/piece and we needed 2 of it (each compressor)

We also bought a leak current clamp tester (+/- 540 €) to measure direct the leak while the units are running.
In our case, we're measuring an imaginary leak but this is also the leak the CLP sees.
I even didn't know that this device existed.

So you see, never to old to learn more.

Any further help about the transposed DC signal is always welcome

Lancer
09-02-2006, 11:36 AM
Hi Peter'
was talking to the Hagar rep today about this problem, apparently normal RCD's are susceptible to DC pulsed from VSD, and any switch mode power supply. The solution to the problem is to use an Type "A" RCD which has filtering to remove the noise, these unit do cost a bit more.
Lance

Peter_1
10-02-2006, 07:50 PM
Filters haven't solved anything, we tried it.
It wasn't noise we had but a DC component together with an AC source.
Learned today that there's in the Siemens LCP some sort of an inverter which takes the DC current on the 3 phases and the neutral and genrates back an AC voltage so that this source can be used to create an additional flux.

Another problem that happened with the truck.
We have in Belgium 3 different ways of earth protection: TT - TN and IT. Does this excists also in other countries?

The generato was wired IT and the main net was TT which can't work together.

http://www.altanet.it/elettro/teoria/sistemi.htm

Never learned so much in a week and never needed so much mine an my sons old schoolbooks.

frank
13-02-2006, 09:34 PM
We have in Belgium 3 different ways of earth protection: TT - TN and IT.
As far as I know Peter these are universal. TT, TNCS, and IT are the normal types of earthing arrangements.

Most of the installations here in the UK are TNCS (Terra neutral combined systems) where the earth and neutral are combined at the installation.

This is one of the reasons I asked if your trailer had earth rod connections on site