SN6505B: Use with cascode transistors for high Vin application

I've found in simulation that the drain-source capacitance of the CSD18543 devices shown seems to pull up the drains of the SN6505 to well above tolerance.  I think this particular CSD18543 cascode MOSFET is just too big for this application; not sure if TI has a more suitable part.  At any rate, I added the BAT54C diodes as shown to clamp the drains and this seems to help quite a bit.  The BAT54C comes in a dual diode (common Cathode) package that would work well here and is inexpensive:

As far as the specific diodes, etc... that isn't settled yet.  Trying to prove this concept now and will get more particular later.  Unfortunately, this transient simulation runs very slowly.

Ideally, we would like to source 10-15 Watts on the secondary side, which will push this topology a bit.

Hi Steve,

How long were the overvoltages when using CSD18543? In any case, DMN6140 is an appropriate choice since it also features low Rds_on and a drain-source voltage spec that can tolerate > 2x the 15V supply.

When selecting the Schottky diodes, ensure they can withstand > 2x the transformer output voltage since the push-pull converter action causes voltage swings of 2 x Vin x N on the diode cathodes [where N is the transformer turns ratio]. If this system is expected to operate in high temperature, we do recommend using the RB168MM-40.

The transformer V-t product should also be sized appropriately, but I understand that might not be concern yet. Was the transformer model from your simulations imported to TINA?


Thank you,
Manuel Chavez

Using CSD18543 had large, square wave pulses exceeding about 10 volts on D1, D2  despite cascodes.  The Cds of those parts is just too big.  I put some shunt 100 Ohm resistors on the sources of these cascodes to try to snub this artifact, but they didn't do much- just made an RC decay out of a bad problem.  Then I added the BAT54 snubbers, and that helped, but still didn't work nearly as well as DMN6140.

The transformer is the same as what came with the default SN6505 simulation.  Would hope to use one of the off-the-shelf option from Wurth or another vendor.

I wish we had a higher current version of this part- I'm worried that this solution might not have enough headroom or margin in it.

-Steve

Hi Steve,

Understood about why CSD18543 cannot be used in this circuit. Some TI models are inaccurate compared to their physical devices, but parasitics on D1/D2 lines are best minimized.

What is the target current output of this isolated supply?


Thank you,
Manuel Chavez

Need Vout ~= 15V, Pout 10-15W. As simulated, it's doing 13.8V, 12.7W and the D1, D2 currents are right at 1A.  We could get more power with higher Vin, but not sure that's going to be available.