Odd voltage output of CT

[theit8514 (Jack)]

Hello,

I am trying to use some smaller solid-core CT sensors that I will be installing into a rack mount PDU. I wanted to test these CTs before I move forward. My tests seem to work after I calibrate them to a specific amperage, but then fall apart at different amperage.

The two CTs I have on hand are the PE-51719NL and the CST306-3A. The former has 3 pins and is listed as turn count 200CT (what I assume is 200 turns to the center tap and then 200 turns to the end). The latter just has 2 pins and 200 turns. I have tried both of these at 200 turns with both 18 and 33 Ohm resistors.

I did some sampling with my “Arduino oscilloscope” to get an idea of what the sine waves look like after connecting the CST and I encountered this strange output at 8 to 10 amps which is about half of what the maximum is. I was expecting a sine wave with higher fluctuations. Can anyone tell me what’s wrong with my test circuit? I have followed this diagram in setting it up and I am using the 18 Ohm resistor for both outputs above.

The fun maths I did:
Primary peak-current = 20A * 1.414 = 28.28 (I will only be measuring a 15 A max circuit, 12 A sustained, figured 20A was fine here)
Secondary peak-current = 28.28 / 200 = 0.1414
Ideal burden resistance = 2.5 ÷ 0.1414 = 17.680339463 (using an Arduino Mega running at 5v)

Last question: At .5 Amp, I am seeing a fluctuation of ±0.09. The math seems to suggest it should be ±0.045 (.5 A primary / 200 turns = 2.5 mA secondary * 18 Ohm = 0.045 V). Is this correct?

Thanks in advance for any advice on this. I am fairly new to electronics on this level.

[dBC]

Welcome Jack,

@theit8514

The two CTs I have on hand are the PE-51719NL and the CST306-3A.

If you check out their specs you’ll see they operate in the 20kHz to 200kHz range (not 50Hz) and are typically used in switch mode power supply designs to measure the switching current of the switch transistor(s). You need something that works in the 50Hz to 400Hz range, something like this maybe.

[Robert.Wall]

Jack, please post the pictures here. If that 3rd party site goes away, nobody can see the picture any more, and any value your post might have had to someone else will be gone.

I’ve raised your trust level so even if you couldn’t attach the picture, you should be able to now.

[borpin (Brian Orpin)]

@Robert.Wall

Jack, please post the pictures here.

@theit8514, posting a picture is as simple as pasting it into the post.

[theit8514 (Jack)]

Sorry about that. I have to go to work right now but here are the images and the code I used to generate it:

Half amp:

Eight-Ten amps:

#include "Streaming.h"
int data[1662];
void loop() {
    for(int j = 0; j < 1662; j++) {
      data[j] = analogRead(i);
    }

    for(int j = 0; j < 1662; j++) {
      Serial << data[j] << endl;
    }
    Serial << endl;
    delay(1000);
}

I took the values and multiplied by the 0.0049 ADC sensitivity constant and got the actual voltage.

Thanks!

[Robert.Wall]

That’s very much as I’d expect. As the frequency falls, the amount of iron (or ferrite most likely in your case) needed to carry the same power increases. That explains why it’s more or less OK at a very low current, but saturates horribly at what would otherwise be well inside the working range at 20 kHz.

So I’m afraid that, for 60 Hz, you’ll need to have a physically larger c.t. There’s no alternative - it’s pure physics.

[theit8514 (Jack)]

Thanks @Robert.Wall and @dBC for the input. I was not aware there was a minimum frequency on the parts I purchased (AFAIK, this is not listed on the SCT-013-000 datasheet, at least in such plain terms). I ordered one of these 60Hz parts and will test it.

[dBC]

No problem. If you’re interested there’s an old thread here on how frequency impacts on CT saturation. it’s even visible down at 50Hz Vs 60Hz if you’re asking the CT to drive a bigger burden R than it was spec’d for. You 60Hz folk have a slight advantage over us 50Hz folk… and airborne 400Hz folk have an even bigger advantage.

[theit8514 (Jack)]

Received the new part today and everything is looking good. Voltage on the Arduino output is looking like a proper sine wave at 7 amps:

After some calibration, matching my commercial amp meter. 0.01-0.1 Amp is probably well within my acceptable tolerance, but I can probably make it better by dropping the allowed amperage and swapping the resistor.

I noticed some weird wobbling of the sine wave when reaching 9 Amps. Is this just bad resistor tolerance for the 2.5V baseline or is it something to do with the voltage drop on the primary (120V to 115V when drawing 9 Amps)

Thanks for your help!

[Robert.Wall]

It might help if you drew your graph as “lines only”.

What you’re seeing in the second graph is a low frequency - about 5 Hz - superimposed on the desired waveform. What do the power supply rails look like? I wouldn’t be surprised if you can see the l.f. on there too.

@theit8514

is it something to do with the voltage drop on the primary (120V to 115V when drawing 9 Amps)

I don’t understand what you’re getting at there. A c.t. knows nothing about the voltage on the conductor that passes through it - except for the very slight chance that there might be a tiny bit of capacitive coupling. But given the impedances associated with the c.t. input, I’d be very surprised if you could see or measure that.