3rd Harmonic on VFD line side

[LMAO]

Aside from input voltage imbalance, what can cause 3rd harmonic on a typical 6 pulse, non-active front end VFD? I have checked the source and it is balanced. There is also a passive harmonic filter upstream to VFD which I think creating - or at least amplifying - 3rd harmonics. There is about 3% 3rd harmonic which is pretty high considering that 5th and 7th are about 3%. The passive harmonic filter is a line reactor (upstream and in series with VFD), trap reactor and trap capacitors (both upstream and in parallel with VFD). When I take the trap circuit off-line 3rd harmonics drop to almost 0 but needless to say TDDi goes up to about 20% as expected.
Our client requires us to meet IEEE519 by having TDDi<5% (ISC/IL=20) so it is already hard to meet the specs without the stupid 3rd popping up out of nowhere. Harmonic filter manufacturer says something about VFD PWM frequency "drifting" and thus causing the 3rd but I am not buying it. I even measured the voltage drop across trap reactor and there was about 10% difference between a phase and other 2.

 

[Jberg73]

The only way to check pulse width accurately is to put a scope on it, between that and the filter either one could be bad.

Jason

 

[Jberg73]

At least that's the only way I know of checking it, if anyone else knows something I'd love to learn.

Thanks

Jason

 

[Besoeker]

Aside from input voltage imbalance, what can cause 3rd harmonic on a typical 6 pulse, non-active front end VFD? I have checked the source and it is balanced. There is also a passive harmonic filter upstream to VFD which I think creating - or at least amplifying - 3rd harmonics. There is about 3% 3rd harmonic which is pretty high considering that 5th and 7th are about 3%. The passive harmonic filter is a line reactor (upstream and in series with VFD), trap reactor and trap capacitors (both upstream and in parallel with VFD). When I take the trap circuit off-line 3rd harmonics drop to almost 0 but needless to say TDDi goes up to about 20% as expected.
Our client requires us to meet IEEE519 by having TDDi<5% (ISC/IL=20) so it is already hard to meet the specs without the stupid 3rd popping up out of nowhere. Harmonic filter manufacturer says something about VFD PWM frequency "drifting" and thus causing the 3rd but I am not buying it. I even measured the voltage drop across trap reactor and there was about 10% difference between a phase and other 2.

I agree. It shouldn't be from yout 6-pulse VFD.
Are there non-linear single phase loads upstream of the filter?
Not necessarily large loads, just lots of them? That's an issue I've come across.

 

[Jraef]

I agree. It shouldn't be from yout 6-pulse VFD.
Are there non-linear single phase loads upstream of the filter?
Not necessarily large loads, just lots of them? That's an issue I've come across.

That's the first thing I was thinking too. You are assuming the VFD is the cause, but in reality it's other single phase non-linear loads. Anything with a power supply is a non-linear load, and almost everything we use now has a power supply, including 90% of the lighting now. CFLs, LED, HID, Fluorescent, they all have a form of electronic power supply. In single phase non-linear loads, the 3rd order harmonics are not going to cancel out. They can be effectively mitigated by paying close attention to balancing the TYPES of single phase loads on a 3 phase system, not just the raw current values. It's a lot more difficult though.

Many Active Harmonic Filters are designed to only react and correct 5th, 7th and 11th order harmonics, no triplens unless you specifically request that (and pay extra for it in some).

 

[Besoeker]

That's the first thing I was thinking too. You are assuming the VFD is the cause, but in reality it's other single phase non-linear loads. Anything with a power supply is a non-linear load, and almost everything we use now has a power supply, including 90% of the lighting now. CFLs, LED, HID, Fluorescent, they all have a form of electronic power supply. In single phase non-linear loads, the 3rd order harmonics are not going to cancel out.

Actually add in the neutral of a three phase system.

 

[LMAO]

There is no other load, linear or non-linear. Because this is a 50Hz VFD (harmonic filter is tuned for 50Hz) and we did not have a 50Hz source VFD was initially run by another VFD which had sine wave filter at its output; 3rd order harmonics were present so I assumed the problem was the source VFD so we rented a 50Hz generator to test the VFD. In both cases there was about 4% 3rd on 2 phases only (other one was negligible). I replaced everything: VFD, harmonic filter components and still getting the same 3rd harmonic.
There is very little 3rd when I am taking the trap circuit off-line; also, voltage drop across trap reactors are not consistent; one phase is off by about 10%.

 

[Besoeker]

There is no other load, linear or non-linear. Because this is a 50Hz VFD (harmonic filter is tuned for 50Hz) and we did not have a 50Hz source VFD was initially run by another VFD which had sine wave filter at its output; 3rd order harmonics were present so I assumed the problem was the source VFD so we rented a 50Hz generator to test the VFD. In both cases there was about 4% 3rd on 2 phases only (other one was negligible). I replaced everything: VFD, harmonic filter components and still getting the same 3rd harmonic.
There is very little 3rd when I am taking the trap circuit off-line; also, voltage drop across trap reactors are not consistent; one phase is off by about 10%.

Something odd with that. You wouldn't usually tune a harmonic filter for fundamental frequency.

 

[LMAO]

Something odd with that. You wouldn't usually tune a harmonic filter for fundamental frequency.

What I meant was that harmonic filter is tuned for multiples of 50Hz.

 

[Besoeker]

What I meant was that harmonic filter is tuned for multiples of 50Hz.

Is it active or passive?

 

[LMAO]

Is it active or passive?

Passive. Line reactor in series with VFD, trap reactor and caps in parallel with VFD down stream to line reactor.

I don't think active filters are tunes for the line frequency; the L-C filter on line side of an active filter is to isolate the active filter PWM. but mine is passive anyways.

 

[Besoeker]

Passive. Line reactor in series with VFD, trap reactor and caps in parallel with VFD down stream to line reactor.

I don't think active filters are tunes for the line frequency; the L-C filter on line side of an active filter is to isolate the active filter PWM. but mine is passive anyways.

I used to tune mine to 177Hz. Bear in mind that we are 50Hz here. Many used 225Hz but that was too close to 5th for my liking.
One of the most difficult was on an installation of four 760 kW drives at the end of a long line. It was like threading a needle to avoid parallel resonance with the source.
Maybe that's an issue for you. Or maybe your particular filter has some other resonance problems.
I know it isn't an answer.

I'll see if I can dig up some spreadsheets I used - but they were on a laptop that died.

 

[ron]

Our client requires us to meet IEEE519 by having TDDi<5% (ISC/IL=20) so it is already hard to meet the specs without the stupid 3rd popping up out of nowhere. Harmonic filter manufacturer says something about VFD PWM frequency "drifting" and thus causing the 3rd but I am not buying it. I even measured the voltage drop across trap reactor and there was about 10% difference between a phase and other 2.

Keep in mind that IEEE 519 makes the measurement at the point of common coupling (service entrance) which often will dilute from other harmonics and not be an issue at the PCC unless you have a very high percentage of VFDs to overall building load.