Harmonics on single phase VFD 2

[eeprom]

Hello,
In monitoring current waveform distortions on 3 phase drives, I have never seen any significant 3rd harmonic content. Ideally all triplen harmonics cancel.

Today I checked the current on a single phase drive, and there was significant 180 Hz content, higher than the 5th. I also saw a 9th content. This drive has a single phase input (240VAC) and a three phase output.

Can someone explain why the triplen harmonics are not canceled?

Thanks,
EE

 

[jraef]

First, I assume you are referring to the INPUT of the drive.

Assuming that, think about why TRIPLEN harmonics canel each other on a 3 PHASE system.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[eeprom]

Yes, of course, I am referring to the drive input. I had considered the single phase/three phase issue, and why triplens cancel on a three phase system. I was wondering how much of the harmonics come from the response of a square wave, and how much comes from just the fact that the drive is simulating a 3 phase output.

 

[itsmoked]

The drive output has nothing what-so-ever to do with the input. So scratch any thinking in that direction.

There are no "square waves" associated with the input either. This leaves how many wires to feed the drive?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[eeprom]

We might be getting off course. I'm not sure why you would say that the output has nothing to do with the input. It has everything to do with the input.

The output controls the load through the firing of the transistors. The transistors are on a DC bus, switching on and off, thus producing square waves to the motor. Those square waves make the load appear non linear to the input. The square wave is an odd function, which causes harmonics in the first place.

 

[ScottyUK]

The inpit and output are effectively decoupled from each other by the DC link. The current you see is predominantly the DC link charging current, which is a fairly short pulse near the voltage peak. This waveform is rich source of odd harmonics.

Triplens don't cancel in a three-phase group, they add. They can't cross a delta/star winding, which is where they are essentially contained and prevented from spreading further into the system.


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

If we learn from our mistakes I'm getting a great education!

 

[eeprom]

Yes, they are decoupled in that the AC line does not see the motor. But the AC line source sees a capacitor bank as its load. That load is affected by the pulsing of the IGBTs. The only time the line sees a load is when the DC bus drops below AC peak to peak voltage. I'm pretty sure we are in agreement on this.

It's the triplens that I don't understand. I think I understand why the triplens are showing up in my single phase test. But I could use some help.

The triplens are in phase with one another, so this becomes zero sequence current. In a delta system, the triplens get stuck in the delta. But many systems for industrial applications are 3 wire Y systems on the load side. So in a three wire Y, where do the triplens go? Can they flow? Why aren't they normally present in frequency spectrum analysis?

 

[LionelHutz]

The "square waves" make the load appear non-linear to the DC bus.

The DC bus and the rectifier make the VFD appear non-linear to the power system. You could have almost any type of load on the DC bus and still see the same input.

You should likely Google harmonics and rectifiers to get more info on what causes the harmonics, because your belief that the output switching is causing the harmonics is incorrect.

 

[Skogsgurra]

It is about 3x120=360. Have a look at the trefas.exe program. It is available on

http://www.gke.org/pub/

(there are lots of files, this is about half-way down in the list). Run it and experiment with phase configurations, number of harmonics, activate the "Show sum" and so on. Most guys learn a lot from this.

Gunnar Englund

http://www.gke.org

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

 

[eeprom]

You don't need a computer program to figure out how harmonics on a three phase system add up. This is pretty simple algebra.

Thanks for all your help.

EE

 

[Skogsgurra]

Excuse me. I only tried to be helpful. Did you run the program?

Gunnar Englund

http://www.gke.org

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

 

[jraef]

"Assuming that, think about why TRIPLEN harmonics canel each other on a 3 PHASE system."

Who posted that and what drugs was he taking at the time?

Jeebus sometimes I should read what I'm posting BEFORE hitting the Submit Post button...

 

[2dye4]

The third harmonic is due to input rectifiers feeding dc bus capacitors. Think about the charging, current only flows when the AC line exceeds the cap voltage leading to pulses at the peak of the sine wave.

 

[sibeen]

eeprom, the input to the drive is made up of a fairly simple rectifier, the design of which is constrained by the number of input phases. The type of rectifier can be called the number of pulses that the rectifier outputs per input AC waveform.

So, for instance, with a three phase system using the standard 6 diodes/thyristors (Graetz Bridge) the DC output waveform has a total of six pulses per incoming AC waveform. This can be expanded out by the use of phase shifting transformers so that you ca achieve 12, 18 24 pulse systems.

For a single phase input you either have a rectifier which produces a single DC pulse (simple diode) or a system which produces two DC pulses per incoming AC sinewave using a full bridge.


A rule of thumb for these rectifier configurations is that the harmonic spectrum of the input current depends upon the pulse number of the rectifier by the equation

H = pn +- 1.

Where p = the pulse number of the rectifier
and n = the intergers (1, 2, 3 , 4 etc).


So for a 6 pulse sytem we expect to see a predominance of the 5th (6-1), the seventh (6+1), eleventh (2 x 6 -1), thirteenth (2 x 6 + 1) etc, etc.

The amount of harmonic current produced by each individual harmonic is approximately (again a rule of thumb) the inverse of the harmonic order, so that the fifth harmonic will produce 1/5 the harmonic current of the fundamental current, the seventh 1/7, the eleventh 1/11 and so on.


Note that for the 6 pulse system there are no triplens available to be produced under the above equation.

For a single phase full bridge the pulse number is 2. We therefore have the full range of odd order harmonics being produced with the third harmonic being of the largest order, approximately 1/3 of the fundamental current.

 

[eeprom]

Sibeen,
Excellent information. Thank you.

I hope I haven't been misleading with my question. Let me clarify: The single phase VFD is actually a three phase VFD wired for single phase. The load is three phase, the source in single phase. This is very common, especially on small VFDs. The VFD does have the standard 6 thyristors. But of course it has a single phase input, so the DC bus capacitors see much more ripple.

Wouldn't the firing order of the thyristors be the same, as based on the three phase load? And therefore, wouldn't the harmonics also be based on the equation H = pn +- 1?

 

[LionelHutz]

I believe everyone knows the configuration you are describing.

The rectifier section is 4 diodes feeding a capacitor bank. It is a single phase full-wave rectifier with a capacitor filter. The operation and harmonics are due to it being a single phase rectifier with a capacitor filter.

http://lmgtfy.com/?q=single+phase+rectifier+harmonics

The VFD may have 6 diodes inside but 2 of them are not connected. It should be fairly obvious you can't use all 3 of the phase inputs when you only have 2 wires. The unconnected phase input leaves 2 diodes unconnected.

 

[eeprom]

LionelHutz
There are two ways of connecting a three phase drive to a single phase system. One method is to tie inputs A and B of the drive to L1 and L2 of the supply, and then jumper inputs B and C. This means that all the diodes are connected.

The other method is to only use two inputs, as you suggested. In order to do this you have to disable the "input phase loss" detection in the drive.

If I recall, this is not an option on all drives.

In my example, we have jumpered phases B and C.

 

[Skogsgurra]

It doesn't matter what way you connect it. It will still be a Graetz configuration with four diodes and the mains current looks the same in both cases.

Gunnar Englund

http://www.gke.org

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

 

[jraef]

In my experience the VFDs that require you to connect B and C input terminals (or any combination), typically do so because they derive the VFD's control power from the AC input side with a separate SMPS board, as opposed to using a DC/DC converter off of the DC bus. So because they can't know which two terminals you might connect to, they require you to jumper two of them together if fed with single phase. Whenever I have come across that I have just moved the terminal to achieve the same result. This has no effect on the rectification as skogsgurra explained, so there is no difference in the harmonics created.

That separate SMPS configuration by the way has pretty much fallen by the wayside with the smaller VFDs (those that do not need de-rating for single phase input) because of the need to make them smaller and smaller; a separate SMPS takes up real estate on the board and adds to the watts loss in the VFD istelf. But you do still find it in larger units on occasion.

I'm unaware of anyone still using SCRs on the rectifier section of newer design PWM drives. That was done in the past to (hopefully) reduce or eliminate the pre-charge circuit on the DC bus, but (in my humble opinion) it turned out to cause more problems than it solved.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[LionelHutz]

Jumping 2 terminals together so the drive will operate is irrelevent to the operation of the rectifier. That is done just so the drive sees voltage on all 3 inputs. It puts the 2 diodes in parallel with 2 other diodes which doesn't change the fact that it's still a single phase full wave rectifier.