choosing the right CT /VT for THD measurements

[garimella]

I have to develop a power meter that will record harmonics on incoming 3 phase currents and display the power quality such as THD to the user. Now I am not sure how to select CT for my application as primary CT on the power lines are sensitive to 50-60Hz . Since I need to measure upto 9th harmonics, should I choose CT with higher frequency response? Will the primary CT not attenuate higher order harmonics?

Just to clarify: Primary CT is 1: 100/50Hz-60Hz, Nominal is 500A . The output Nominal current is 5A. This current will form the input to my meter.

 

[Easy peasy]

a standard mains CT, either 500:1 or 1000:1 will give you accurate harmonics up to the 20th & beyond, make sure the burden resistor is correct or ideally lower than max allowed.

 

[KlausST]

Hi,

So the requirement is rather clear:
You want to measure up to 9th hamonics, then your bandwidth needs to include these frequencies .. with the expected precision.

example: 50Hz, up to 9th harmonics, max 1% precision error
Then the CT needs to be able to "transform" the frequency band: 50Hz ... 450Hz with max 1% precision error

Will the primary CT not attenuate higher order harmonics?

It depends on:
* source impedance (at the given frequencies)
* load impedance (at...)
* CT impedance (at...)

I`d calaculate with (worst case):
CT_impedance < ((load_impedance + source_impedance) * expected_precision_error)

I expect the CT impedance to be low enough, then it shoud not harm. But to be sure you need to check all 3 values.

Klaus

 

[garimella]

How does the frequency specifications of CT play a role. This is not clear. Suppose I choose a CT whose frquency response is from say 40Hz to 1Khz, would it be better or worse than a 50/60Hz CT?

 

[KlausST]

Hi,

counter question:

If it is not specified for the frequency range you need...how can you be sure it works like expected?

For sure you may use the "trial and error" method. But buy 10 pieces and you can´t be sure that the 10th one behaves exactly like the first one.
The risk is on you.

Klaus

 

[garimella]

Trial and error is fine and But I wanted some technical basis for my choice.

 

[KlausST]

What exactly do you mean?

 

[garimella]

I wanted to know whether frequency response of CT actually matters or not. Most of the metering CTs have 60Hz CT. I was wondering why. I found another CT CT07-1000 which is commonly available. But its frequency range is from 47Hz to 2KHz. Would this be OK.

 

[KlausST]

Hi,

did you read my post#3?
What additional information do you need?

Klaus

 

[KlausST]

Hi,

I just had the time to read the datasheet.

The first I found (icecomponents CT07-1000"), dated "9/11" talked about 40Hz ... 200,000Hz. I was surprised.
Then I went to icecomponents.com and downloaded their datasheet, dated 11/17. Now they talk about 47Hz ... 2500Hz.
Without any change notice. The datasheet now is reduced from 2 pages to a single page.

None of the datasheets talks how they measure the frequency range. None talks about accuracy, precision, distortion....
Let´s say they used the -3db limit. But -3dB still means 30% error.

This is a classical transformer style CT. Depending on the core they use you may have additional measurement errors.
* typical amplitude loss
* but also additional distortions due to hysteresis and saturation effects.

If you want to use them, a simple "try if 450 Hz goes through somehow" may not be sufficient for your (good quality) THD measurement tool.

You need to measure all the THD effects that are caused by the CT itself. And - worst of all - I guess it is not linear.

If I had to build a THD measuremnt tool, then I´d ensure that the measurement does not add significant THD to the input signal.
Neither with low currents nor with high currents.
Thus my way: either a resistive current measurement, or an active CT with field compensation. (like many LEM current sensors)
You may also check whether hall sensors are good enough.

For sure you don´t need to go my way.

Klaus