Synchronous rectification problems

Question

I am currently designing a LLC resonant converter and am trying to perform synchronous rectification but am running into some issues with it.

This is my design and note I am currently using the IRFZ48SPBF MOSFET and the FAN3111ESX MOSFET driver as the desired ones were out of stock.

Here is the waveforms I am getting (drain current in green, gate voltage in yellow, drain voltage in blue, and other mosfet gate voltage in purple).

I was pretty sure that this was happening because the secondary was resonating with the parasitic capacitance(s) from the mosfet but the ringing is happening at 2.7 MHz and when I measured all the capacitance values of the MOSFET (Cds = 6.3 nF, Cdg = 3.5 nF, Cgs = 3.7 nF, Lcentretap = 3.9 uH, Lsecondary = 7.5 uH), the largest resonant frequency I could get was 1.4 MHz; This confused me a little bit.

I replaced the mosfets with schottky diodes and the ringing went away (at least the large ringing), so this reinforces my theory, but I am wondering if just replacing the mosfets would work or maybe the drivers have something to do with it too?

Edit: I am concerned because the diodes are much more efficient than the MOSFETs which shouldn’t be the case; the MOSFETs heat up very fast and I think this is happening because they dont fully close causing shoot through... maybe. Oh and the dead time is around 200 ns if that helps.

Answer 1

Mosfets that run as switches, as long as they are otherwise correctly selected for the application, heat up due to two common reasons:

1. Shoot-through.

2. Linear operation.

In your case, the culprit most likely is #2. As a quick check: double the gate voltage. If the mosfets don't heat anymore, then insufficient gate drive was the problem. That's at least how it works for me 90% of the time.

Answer 2

Cds = 6.3nF, Cdg = 3.5nF, Cgs = 3.7nF

The MOSFET data sheet clearly states that the output capacitance is 332 pF but, for a better figure to use, the effective output capacitance (time related) will be more representative (633 pF): -

If we used this figure and the 3.9 μH we get a resonance of 3.20 MHz and this might easily be as low as 2.7 MHz given that I have used typical figures. For instance if your device was more like 900 pF for the effective time related value then you would get 2.7 MHz.

I am wondering if just replacing the mosfets would work or maybe the drivers have something to do with it too

You will get some ringing with this type of design and what I see in your waveforms does not concern me. However, if you see an improvement in power delivery efficiency when replacing the MOSFETs with diodes, then that would concern me.