Current transformer burden resistor

[eslavko]

Hello...

I have 100A/40mA current transformer, and need to measure currents up to 25A.

Now I can't find solution in which case transformer will saturate faster.

1. Single primary loop and 200 Ohm burden
2. Two primary loop and 100 Ohm burden
3. Three primary loops and 66 Ohm burden
4. Four primary loops and 50 Ohm burden

All of them have same output voltage on burden, and I assume that saturation is same on all of them too.
But when I look waveform on scope when the primary current is low the peaks are distorted, and when current is high the peaks are clean. I expect opposite.

[moffy]

Try using a much lower value resistor, say 5R to 10R and see how that effects your output.

[eslavko]

I know that with lower burden I will get better output. I want to have as high as possible burden to use full scale ADC. Otherwise I need to make some amplifier. The question is what theoretical should be better?

[Circlotron]

Current transformers have a VA rating. I expect the I^2R of the burden resistor shouldn't exceed this VA figure.

[eslavko]

So 4 primary turns and 50Ohm Burden should be better than 1 turn primary and 200 Ohm burden?

[Zero999]

The power dissipated in a current transformer is mostly due to the I²R losses. A single turn primary will always give the lowest secondary current and therefore the leat power dissipation.

The secondary voltage is proportional to the flux, which is counteracted by the current. To avoid saturation, the burden resistor must be kept low.

[eslavko]

Still can't answer my question.

If I want same output voltage (let's say 1V) with same primary current, what scenario will work better?
1. One turn primary and 20 Ohm burden
2. Two  turn primary and 10 Ohm burden
3. Doesn't matter.

[Zero999]

What do you mean by better?

Increasing the number of primary turns, reduces the secondary current, giving a lower secondary voltage, given the same burden resistor.

#1 will give ¹/₄ of the secondary voltage of #2, with ¹/₄ of the power dissipation.

[Vovk_Z]

So 4 primary turns and 50Ohm Burden should be better than 1 turn primary and 200 Ohm burden?

In general - possibly better, but It depends on the actual CT VA-curve.

[eslavko]

What do you mean by better?

Increasing the number of primary turns, reduces the secondary current, giving a lower secondary voltage, given the same burden resistor.

#1 will give ¹/₄ of the secondary voltage of #2, with ¹/₄ of the power dissipation.

That's seems strange.

If I hook two turns in CT instead one with same burden then I got twice the reading.

[eslavko]

What do you mean by better?

Increasing the number of primary turns, reduces the secondary current, giving a lower secondary voltage, given the same burden resistor.

#1 will give ¹/₄ of the secondary voltage of #2, with ¹/₄ of the power dissipation.

In my setup I got near same voltage in #1 and #2.
But at least for naked eye the more Turns with lower burden resistance seems to perform better.

[Kleinstein]

The current transformers have a limited secondary voltage. If the voltage gets to high the core magnetization gets to high and than need more magnetization current that causes an error.
The options shown are all giving the same voltage and thus same level where you get saturation.

If one only needs 1/4 the nominal range of the CT one could increase the burden resistor about 4 x, so keep the original maximum voltage (if one ignores the winding resistance, which is often a good approximation). How much maximum voltage is OK with the transformer depends on the design.

To get more output signal voltage one would need more turns on the secondary. More primary windings don't help - the same effect can be had with a larger burden resistor.

[Zero999]

What do you mean by better?

Increasing the number of primary turns, reduces the secondary current, giving a lower secondary voltage, given the same burden resistor.

#1 will give ¹/₄ of the secondary voltage of #2, with ¹/₄ of the power dissipation.

In my setup I got near same voltage in #1 and #2.
But at least for naked eye the more Turns with lower burden resistance seems to perform better.

You're right. I was having a brain fart, when I wrote that.

Increasing the primary turns, also increases the secondary current and if you reduce the resistance in the same proportion, then of course the output voltage will be the same.

#2 will have double the power dissipation of #1 because the current is double, yet the voltage the same. Got with #1 because it will have lower losses/

[eslavko]

The current transformers have a limited secondary voltage. If the voltage gets to high the core magnetization gets to high and than need more magnetization current that causes an error.
The options shown are all giving the same voltage and thus same level where you get saturation.

If one only needs 1/4 the nominal range of the CT one could increase the burden resistor about 4 x, so keep the original maximum voltage (if one ignores the winding resistance, which is often a good approximation). How much maximum voltage is OK with the transformer depends on the design.

To get more output signal voltage one would need more turns on the secondary. More primary windings don't help - the same effect can be had with a larger burden resistor.

Yes I have 100A CT and need only 25A range. As wish to connect that direct to ADC need as high as possible voltage. When I try single loop and 200 ohm burden seems that I have more noise than with 4 loops and 50 ohm burden. I did try to even raise the burden to see what distortion I got when saturate but seems that core has some sort of soft saturation as I can't see distortion with naked eye on scope. I have two types of transformer available. Preferred one is yuanxing CTSA016 100A/40mA, and other is yhdc SCT013 100A/50mA. The saturation knee in SCT is clearly visible. The CTSA core cross section is about 4 times bigger than SCT.

[eslavko]

There is datashet of CTSA016 (100A/40mA) and this is best I can find on net.
There are no charts of burdens but maybe some hidden information I missed.
I believe the ratio is 2500:1.
There is schematic with P6Ke transient diode. I have this type. I test transformer with open secondary and primary at 50A for about 10 minutes just to check if something will overheat, but nothing happend. On the output I see near square wave about 12V peak/peak. (I know that CT should not operate without load and transient diode kick in)
There are rated voltage of 0.333Vac, but not max voltage. 0.333Vac means about 0.94Vpp, and I wish to have about twice of that.

As with open secondary device doesn't overheat I assume it will be safe to operate with burden to get 0.7Vrms. Probably not accurate/linear on upper half. In reality I need to be as precise I can get up to 4A, and distortion over that is not so important.
Somebody see some risks?

[Kleinstein]

Not a surprise to see not much heat with an open secondary, as there is not more current flow and the core losses at 50 Hz are usually low unless for a really poor material (e.g. bulk steel).
Getting a 12 V square wave suggests that there is still quite some reserve towards saturation. So using the transformer with a little higher than nominal voltage should be OK. This especially help with the low current and if at all it would add a little error to the larger currents (e.g. > 40% of the range).

If the performance with low currents over a larger dynamic range is important, one could still consider amplification. This could be like 2 ranges: one direct with no gain and a 2nd ADC input with amplification (e.g. ~ 10 x). This could help as the µC internal ADCs tend to have a somewhat limited resolution.

[eslavko]

Not a surprise to see not much heat with an open secondary, as there is not more current flow and the core losses at 50 Hz are usually low unless for a really poor material (e.g. bulk steel).
Getting a 12 V square wave suggests that there is still quite some reserve towards saturation. So using the transformer with a little higher than nominal voltage should be OK. This especially help with the low current and if at all it would add a little error to the larger currents (e.g. > 40% of the range).

If the performance with low currents over a larger dynamic range is important, one could still consider amplification. This could be like 2 ranges: one direct with no gain and a 2nd ADC input with amplification (e.g. ~ 10 x). This could help as the µC internal ADCs tend to have a somewhat limited resolution.

As ADC have just 12 bit (XMEGA128) I allready split two ranges. The burden is two resistors in series with 1:5 ratio. The 1st ADC have complete burden voltage and other ADC have 1/5 of burden voltage. There is series resistor and protection circuit on ADC side to protect MCU. I sample 3 full periods and if any sample from 1st ADC has value 0 or 4095 then I use ADC2 with higher range.