How to solve current spike in transformer?

[Patrick66]

Hello everyone, does anyone know how to solve the current spike in transformer?

[Konkedout]

It is VERY likely caused by reverse recovery time of your diode.  Even though it is a "super fast recovery" diode, it most likely has Trr that will bite you in a CCM flyback.

I have been bitten by this when I had a 100V ultrafast diode in the output of a (hardware; not a simulation) flyback power supply.

Try using a schottky diode.  If you need reverse voltage capability exceeding 100V or so, use a Silicon Carbide schottky diode.

My design was fixed by going to a 100V schottky type which SFAIK is the effective upper voltage limit for silicon schottky diodes.

[David Hess]

A clamp or snubber would usually be added across the transformer primary whether the spike was caused by reverse recovery of the rectifier or not, because leakage inductance in a real circuit will cause a primary spike anyway.

[eTobey]

A current spike in an Inductor? I believe this is physically impossible.

[T3sl4co1l]

A current spike in an Inductor? I believe this is physically impossible.

Nearly correct.

K is physically unrealizable.  For the same reason:

A clamp or snubber would usually be added across the transformer primary whether the spike was caused by reverse recovery of the rectifier or not, because leakage inductance in a real circuit will cause a primary spike anyway.

doesn't matter here, as there's no leakage to cause voltage peaking, but a real flyback with nonzero leakage will have voltage spiking on the primary side turn-off, and secondary side turn-on.

Voltage peaking does not show here, because there's no leakage to generate it.

But current spiking is real and finite, because even in absence of loop inductance, transistor dI/dt is set by gate drive characteristics (and any parasitics in the device model if applicable).

Tim

[CosteC]

Current spike is unavoidable as long as D1 has recovery  time or if it has capacitance (if it is CCM). In simulation you got zero leakage inductance, but if it would be there it would add spike at MOSFET turn off.
Better diode will reduce spike, so do slower MOSFET turn on. To eliminate it, move to DCM or change topology.

[slugrustle]

Maybe there's a little bit of both/and going on here.  A better transformer coupled inductor model will give more accurate simulation results, and a diode with lower Qrr will give a lower current spike at primary MOSFET turn on in CCM.

nexperia's website is sadly broken at this time (I hate software more every day), but here is a brochure for their SiGe rectifiers, which are another option beyond Si Schottky / SiC Schottky / Si fast recovery rectifiers.  Their voltage range and thermals might not be sufficient for this converter, but they're a good thing to know about, and nexperia is good about publishing SPICE models that work in LTspice for almost all of their parts.

I've seen Si Schottkys go to 300V, although 200V is more common than 300V.