Current sense transformer leakage inductance

[Faringdon]

Hi,
We are doing a 100:1 Current sense transformer for a Buck with 24Vout and 750W. (Vin=42VDC)
The CST uses an EFD15/8/5 core.
The saturation current  for the secondary is 47mA (as you know , that’s magnetising current)
https://ferroxcube.home.pl/prod/assets/efd15.pdf

We need the saturation current to be this high because of the leakage inductance, and the way it results in far higher core magnetising current.
This is demo’d in the  attached LTspice sim, where if you make the transformer coupling coefficient equal to 0.9, then the magnetising current peak is 34mA, whereas its only 1mA  peak if you simulate it with perfect coupling.

Can you shed light on why the magnetising current peak is so much higher when k=0.9 as opposed to when k=1?

[T3sl4co1l]

Treat the core as a virtual third winding.  Which (physical) winding does it couple more closely to?

Tim

[mag_therm]

Hi Faringdon,
I am not sure what you are trying there. It looks like an attempt to measure the peak value of the current pulses,
 in the presence of a high DC offset, without introducing  impedance to the circuit.
My comment would be that is usually this is not feasible with any known ferro magnetic cores.

If the core is driven into magnetic saturation ( tending to U_effective =1 ) due to the long term "average DC" in the primary, then the core ceases to have any effective function on the AC electrical side.
That is, the primary and secondary can be considered to be coupled by air. So the core can be removed without much effect.

Without going into integrals etc, using SI units you can use these on the simplistic  DC analysis of core:

H = N*I /length_effective  [A/m]
and similiar to the above:
phi = N*I/ ( Reluctance _Ferrite + Reluctance_airgap)      [Wb]
 
B = u_effective * H. [T]
and similiar to the above:
B = phi/Core_Area

Ferrite toroids usually go into hard saturation at B > 0.3 to 0.4 Tesla.

references for above include DC choke design, C/T design with DC offset allowance, and my favorite ferrite reference book, Snelling (section 4.2.2 +more)

Other ways to measure, could be a sensing resistor, a high-side differential amplifier, , Hall effect device, and Rogowski coil.

[Faringdon]

Treat the core as a virtual third winding.  Which (physical) winding does it couple more closely to?

Thanks, i would say the core is more closely coupled to the secondary.

The attached shows the CST (Current Sense Transformer) magnetising current for the case of k=0.9 for CST transformer coupling, and k=1.
The comparison is striking. The magnetising current peak is far higher when k=0.9.
In real  life, coupling factor k  in a CST would be more like 0.9 than unity...since coupling between a 100 turn coil, and a 1 turn coil, is inevitably poor.

As you know, the secondary referred magnetising current is the [secondary current] minus the [primary current x 0.01], since its a 100:1 CST.

[Faringdon]

The thing is, that this high magnetising current (which occurs when leakage inductance is high), can be offset with a PNP current source as in the attached. This  significantly reduces the peak magnetising current. The LTspice sim demo's this.
Why is it that nobody uses this "PNP magnetising current reducer" circuit?