Measuring harmonic via CT - PT 7

[tem1234]

Hi,

If i measure harmonic, the best is to measure it directly to the drive (on the power cable or bus), but if i can't, do i'll lost a lot of data if the power analyser is connected to CT and PT?

I was thinking, if the PT is connected in Delta, i'll probably lost a lot of 3rd harmonic?

So i want to know if it's acceptable to measure it that way or not, and if there's a standard to refer at. Do the standard IEC 61000-4-7 is very usefull and do i must buy it absolutely or not? I must measure untill the 50th harmonic and it's the harmonic current which are important for this case.

Thanks

 

[dpc]

PTs and CTs reproduce quite faithfully well beyond the 50th harmonic in most cases. Response may start to roll off maybe around 15 kHz or so. It's much safer to use the CT and PT secondary signals whenever possible.

There are some classic IEEE papers on CT and PT frequency response, but I can't lay my hands on my hard copies right now. If you search the IEEE database, you should find them.

 

[edison123]

Thank you both. I have had this question unanswered for a long time.

* Women are like the police. They can have all the evidence in the world and yet they still want a confession - Chris Rock *

 

[ScottyUK]

Phase relationships between voltage and current at high harmonic orders will be unreliable, the problem starting as the frequency response starts to roll off. The effect isn't likely to be significant to anything except precision measurements below about 1kHz, but it is something to be aware of as frequencies rise.


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

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

 

[Skogsgurra]

Yes, that's right. And not only that. If there is a piece of cable between VT and where you connect your analyzer, the VT leakage reactance and the cable will interact when frequency gets hih. All sorts of problems, from phase shifts to resonance. Try to measure without any long cable attached.

Gunnar Englund

http://www.gke.org

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

 

[slavag]

Hi.
Scotty, Gunnar you are right.
Like to Edison123, I have this Q a long time, but never ask:
why power meter's companies never write about it and EE installed power qulity meters hungreds meters from CT/VT.
Regards.
Slava

 

[dpc]

FWIW - Paper references:

Current Transformer Accuracy with Asymmetric and High Frequency Fault Currents
Douglass, D.A.
IEEE Transactions on Power Apparatus and Systems
Volume PAS-100, Issue 3, March 1981 Page(s):1006 - 1012

Potential Transformer Accuracy at 60HZ Voltages Above and Below Rating and at Frequencies Above 60 HZ
Douglass, D.A.;
IEEE Transactions on Power Apparatus and Systems
Volume PAS-100, Issue 3, March 1981 Page(s):1370 - 1375

 

[scottf]

dpc-

I agree with your comments in regards to CTs. However, I don't agree completely with VTs being good to the 50th harmonic.

The frequency response of VTs vary greatly by design and voltage class.

In many cases, VTs become quite unreliable above 1 kHz, especially units rated 15 kV and above, which is where I have the most experience with frequency response testing. I've not personally been privy to frequency testing on 600V VTs.

Of course, capacitor voltage transformers (CCVTs) should never be used for harmonic analysis, as they are only designed to be accurate at the fundamental.

 

[tem1234]

Thank you all for your reply.

dpc, I'll try to find these references, seems interresting.

And what about the connection of the PT-CT?
They must be wye-wye to not lose all the neutral harmonic i think!

 

[LionelHutz]

I've seen 600V CT's that were rolling off by 1kHz. I believe they actually had a lot of core loss at the higher frequencies.

15kHz may be possible for a good CT but that doesn't universally apply to all.

 

[scottf]

LionelHutz-

That's interesting. What kind of CT was it. I can't imagine any "normal" CT design that would have problems below 36 kHz, much less 1 kHz.

 

[dpc]

scottf,

Thanks for the clarification - I agree that the VT are more of a concern than the CT. My experience has been mostly with 15 kV class PTs in metal-clad switchgear and 600 V PTs. I've never seen a major issue with ratio error, at least that was apparent at the time.

The Douglass paper I cited above indicates ratio error is highly sensitive to burden as frequency increases.

For a 7200 V VT with a 1000 ohm burden, the measured RCF at 10 kHz was 1.01. For a 4800 V VT, the error was 1.02. With a 100 ohm burden, the 4800 V VT RCF was off the chart at 10 kHz.

Sounds like your experience indicates the error could be much higher above 1 kHz??

Dave

 

[scottf]

Burden certainly plays a big part.

I think it's important to remember that VT designs vary greatly from one manufacturer to another. In the case of frequency response, the winding design and the insulation grading plays a big part, in that they effect the capacitance and stray capacitance of the primary. Of course the capacitance seen by the primary, both across the primary and to ground, forms a very complex RLC circuit with implications on frequency response

A common mistake is to take measurements only at high and low frequency points and assume it's the same in between.

For instance, the response may look good at 1kHz and 10 kHz, but the response in between could be quite erratic.

This brochure:

http://www.pqs-consulting.com/ritzbrochure.pdf

Shows a good example of the frequency response of a 230 kV VT. We had other similar curves for MV VTs when I wrote this brochure, but the topic was in reference to HV applications in this case.

 

[dpc]

Thanks, Scott - that's interesting.

 

[busbar]

(From an older {early 1990s} power-measurement seminar) CTs are better than PTs are at passing harmonic quantities given their lower inductance [fewer turns].

 

[scottf]

Busbar-

I think you have that wrong, in the sense that the inductance of the winding of a CT or VT is not what causes error at higher frequencies. Rather, it's the capacitance of the winding/insulation reacting with the inductance of the winding.

Because VTs typically have many more turns on the primary and secondary, the inter-winding capacitances tend to be much higher, not to mention line-to-ground capacitances of insulation (i.e. across the same voltage being measured).

In theory, a transformer that is perfectly inductive in nature would perform linearly with frequency up to a certain point. After all, accuracy performance tends to be better (up to a point) with increasing frequency.

 

[busbar]

Scottf, I stand corrected. Thanks for the update.
 

 

[edison123]

Scottf

Thanks for the additional info. I have a question.

I am often required to measure THD (up to 50th) of a new / rewound ac generator and always have used the generator PT's for such measurements. I could never be sure whether these PT's faithfully reproduce the acutual generator harmonics.

In such a case, what would be your suggestion to measure the harmonics ?



* Women are like the police. They can have all the evidence in the world and yet they still want a confession - Chris Rock *

 

[scottf]

I assume these are 15 kV class VTs?

I would recommend resistive dividers if you need something remotely accurate. You could also try a capacitive divider if you can find one.

Or you could ask your vt supplier to confirm the frequency response, but I'm pretty sure it won't be good for what you're trying to measure.

I believe it is quite common to take THD measurements with normal VTs...most folks don't realize that they're probably looking at garbage

 

[edison123]

Thanks Scott for the quick reply. Yes, normally they are 3.3, 6.6, 11 and 15 KV/110 V PT's and the PT OEM's are useless in giving info about the frequency range.

Resistive / Capacitive dividers ?

I am kinda dumb on them. Could you educate me more ?

Thanks again.

* Women are like the police. They can have all the evidence in the world and yet they still want a confession - Chris Rock *