Current transformer: Magnetising current has negative slope in secondary?

[treez]

Hello,
In the attached LTspice sim of a full bridge smps, we see that magnetising current in the secondary of the current transformer has a negative slope.
This seems counter intuitive.
But is the wisdom involved in its analysis, concerning the fact that the field due to the secondary is the opposite of that due to the primary coil, and so the magnetising current is reducing as the current pulse goes forth?

 

[FvM]

I wonder where you observe magnetizing current of the transformer and how you assign it to either the primary or secondary?

 

[godfreyl]

Are you talking about this - the current through L5?

 

[treez]

Yes it is L5 current.
But ultimately , I am referring to page 3.175 of "switchmode power supply handbook" byBillings & Morey, where they say that the magnetising current subtracts from the secondary of the current sense transformer, and the pulse can have a negative slope if the secondary inductance of the CST is not big enough.

 

[godfreyl]

In the attached LTspice sim of a full bridge smps, we see that magnetising current in the secondary of the current transformer has a negative slope.

OTOH, the current in the primary has the same waveform as the current in the secondary, so in this case it looks like the CT is working as it should.

Yes it is L5 current.
But ultimately , I am referring to page 3.175 of "switchmode power supply handbook" byBillings & Morey, where they say that the magnetising current subtracts from the secondary of the current sense transformer, and the pulse can have a negative slope if the secondary inductance of the CST is not big enough.

I don't follow that explanation, but... If the current in the primary was a square wave with flat tops and bottoms, and the secondary inductance was too low (or the resistance too high), I would expect the waveform of the secondary current to slope in the opposite direction to the waveform in your example.

 

[FvM]

I don't understand the term "secondary inductance" related to the problem. A transformer without leakage inductance is characterized by only two parameters, a main (or "magnetizing") inductance and a windings ratio.

Whatever the "negative slope" statement means, the present simulation isn't an example for it, because the magnetizing current is only a small fraction of the total transformer current.



P.S.: You get a negative slope of the secondary current if reducing the CT inductance by a factor 5 or more.

 

[godfreyl]

I'm assuming "secondary inductance" is what you'd get if you measured the inductance of the secondary winding. Should be equal to the primary inductance times the square of the turns ratio.

For clarity, do the waveforms posted show a "positive" slope or a "negative" slope?

 

[FvM]

Negative slope in secondary current looks like below:

 

[treez]

The following shows the unfortunate distortion in the secondary waveform (full bridge smps ) due to magnetising current.
https://i40.tinypic.com/be652f.jpg
..the red shows that the gradient of the secondary waveform is significantly less than the green waveform which is what we would like it to be.

I conclude that the magnetising current, when "seen" from the secondary, rises in the negative direction, resulting in this subtraction from the ideal waveform.
The shocking thing is that the secondary inductance is very high in the 1:100 trafo and yet still there is such distortion in the waveform.


Schem and LTspice simulation attached

 

[FvM]

The shocking thing is that the secondary inductance is very high in the 1:100 trafo and yet still there is such distortion in the waveform.

Why shocking? Just normal behaviour.

Refering to the original question, I understand that negative slope or no slope would be a problem for a current mode controller when the current transformer is used in the current feedback path. But some 10 % relative error as such won't matter.

It's a different thing if you want to make quantitative current measurements through the current transformer. But the error can be well calculated, it's up to you to choose transformer parameters as required for specified accuracy.