[Moved] Ideas for 60+ MHz MOSFET Gate Driver Circuits

[RF_R_US]

This my first post to this forum and I hope some people might have some "out of the box" ideas for this project. This topic borders between switchmode and RF as we are designing a class D half bridge RF generator at 60MHz and beyond. We have already made one work fairly well at 40MHz, ~80% Efficiency, but now are having difficulty in identifying the gate drivers. We've looked at the Intersil ISL55110 and it appears to be the fastest on the market, but it cannot drive the gate capacitance we anticipate. FYI - We are a power semiconductor manufacturer and supply power devices to our customers who actually design the circuits, but we are assisting on the driver section for the higher frequency units. We have new power MOSFETs coming out in a "pizza mask" from our fab, with several new technologies, that we think will get the Ciss down to ~750pF. I'm thinking about using the ISL55110 as the pre-driver or perhaps an LTC4441 as all of the CMOS based driver IC's probably have the same Trise/Tfall into the same load from what I can tell from the data sheets. The IC would then drive a complementary pair of high frequency, high current devices to drive the MOSFET gate. My problem is that I can get high Ft, but not enough current (need to source/sink ~ 8A) or enough current but low Ft. Also, this is NOT a real square wave input signal (defies laws of physics ;-), we can live with ~30% duty cycle, which at 60MHz would be a Trise/Tfall of ~ 3nS. Anybody have an idea of what semiconductor technology would fit the bill for the complementary pair gate drivers?

Thanks,

John

FYI - I'm a microwave engineer by trade, so all this semi-switch mode stuff is new to me.

 

[Dan Mills]

Have you had a look at the IXYS stuff, the DEIC515 will get you to 45MHz with plenty of drive current, it may be worth an enquiry.

If this is a narrow band application you may be able to tune out some of the gate cap with a suitable inductor, time to fire up ADS and have a look at some Smith charts (Watch for stability issues, you really want to plot the stability circles if you try this)?

You could always use a couple of modest RF power devices to stiffen the ouptuts of a more reasonable gate driver, I am thinking some of the smaller ldmos parts may fit (Freescale, NXP both have VHF power ldmos parts in the several watts range that may serve), transformer coupled gates are nout out of the question and if you swamp with a shunt resistor you can get a reasonably broadband match to the driver stage.

IMHO, you need to be thinking broadband RF power amp rather then digital style squareish waves, and should probably be thinking class E output.

Regards, Dan.