Harmonics on AC power from switching supplies

[hrc]

I do understand that putting alot of switching power supplies (talking 1000 500W units) can cause issues with harmonics on the AC feed.

First, whats the main cause of this?

Second, using a type K transformer is supposed to help? Or what can be done to help this?

The customer is seeing more issues when running off a UPS vs the main power feed, but his power isnt very clean to begin with.

 

[tinfoil]

Any device that takes a sinusoidal power waveform and extracts a different waveform from it (could be DC for a power supply, could be a different frequency for a VSD) is going to have 'leftover' bits on its supply side. This shows up as frequency harmonics. (A review of Fourier transform theory will show why this is inevitable).

The better the device, the smaller the resultant harmonic. But there will always be SOME distortion. As the number and size of these devices increases, so does the amount of harmonics, which to NOT tend to statistically cancel out.

 

[jraef]

Actually, all the "triplen" harmonics (multiples of 3x the line frequency) will cancel each other out, as will all the even order harmonics. The problems come from odd order non-triplen harmonics, i.e. 5th, 7th, 11th etc., that will tend to circulate in the transformers and overheat neutral conductors. That is the basis of the "K factor" transformer you mentioned. They are simply overdesigned to be capable of handling a larger than normal amount of circulating harmonics without damage. The higher the K factor, the more they can take. It is not really a solution to the total problem, just the part of it that used to cause transformers to melt down. There are going to be other significant problems with having 1000 500W SMPS's in your system, especially if you check out other contributors such as energy saving flourescent ballasts, HID light fixtures, VFDs etc. that may be present. You should find a decent power quality consulting firm and commission a study.

"Venditori de oleum-vipera non vigere excordis populi"

 

[mpparent]

You may consider the possibility of the UPS seeing the power factor approaching a leading condition. As this occurs, the UPS must be derated. I hope this isn't what's happening, as I've heard of some issues with UPS's dumping their critical load once a certain load condition is reached at a leading power factor. Can you be more specific as to the issues you're seeing?

Mike

 

[hrc]

Well...I was more asked my opinion about this installation and since I wasnt sure...asking here of course!

I really don't have much more to go on other than looking at some waveforms of the voltage and current that were emailed to me. I don't know how to attach a file here, but some of the current waveforms are approaching what can be said.....not much of a sine wave component left.

I'm just looking for something I can point these guys to look at, its our gear but not our UPS so I don't need to solve it, but makes us look better if we can.

Sorry that I don;t have much other info, guess I was looking for a magic bullet of sorts....maybe a sine wave filter on the output of the UPS?

 

[davidbeach]

Actually, all the "triplen" harmonics (multiples of 3x the line frequency) will cancel each other out

WHAT????? Were that true there would be no need for oversized neutrals or any of the other means of dealing with increased currents due to triplen harmonics (principally the third). As stated, the even harmonics do all cancel (to the extent that they are not phase shifted relative to each other) so it is only the odd triplens that are of interest here, and the ninth is the next one beyond the third to be of concern and it is usually of a significantly lower magnitude.

What triplen harmonics do though, is circulate in the delta winding of delta-wye (harmonic source on the wye side) transformers.

Taking half the load and phase shifting it 30 degrees will cause the fifth and seventh harmonics to cancel. Additional load splitting and phase shifting will help deal with the eleventh and thirteenth harmonics (from 12 pulse rectifiers).

Because there is a certain amount of small, but random, phase shifting during harmonic production, there is always some canceling, and the more harmonics produced the more of this type of cancelation happens.

 

[jraef]

Woops, can't believe I posted that. Musta been asleep. Triplen harmonics are ADDITIVE, not cancelling. I wish sometimes I could edit posts like in other forums...

"Venditori de oleum-vipera non vigere excordis populi"