EMRFD Message Archive 4709

Message Date From Subject
4709 2010-05-24 08:03:58 Tim Power MOSFET gate drive
I've been playing around with class E RF amps using power-supply-type power MOSFET's. e.g. IRF510's but some others too.

Certainly one issue with RF Amps done this way is driving the gate on the MOSFET effectively.

The IRF510, despite being a little long in the tooth, is actually quite good in terms of gate drive. Spec-sheet wise it has an effective switching gate capacitance of 180pF or gate storage charge of 8.3nC. I compare it with other parts in the same power and current class (e.g 43Watt, 100V, 5.6Amps ballpark) and it still is among the best among the power-supply class part for this voltage range.

The IRFZ24 used in the latest Homebrew challenge winner (50W for $30) has a higher gate capacitance (370-590pF) and lower max drain source voltge, but a much lower on resistance (eight times better than the IRF510). Maybe the figure of merit you want to optimize, is smallest product of gate capacitance times on resistance.

Smaller parts (e.g. 2N7000) of course have less gate drive requirements because they have less gate capacitance (roughly 50pF) and more on resistance (5 ohms). Bigger parts have more gate drive requirements, e.g. the FQA11N90 popular among class E QRO'ers has a gate capacitance of 3500pF but of course much larger power capability (due to higher voltages).

Seems to me that making a 3-D scatter plot, of say Rds vs gate capacitance for parts in a particular power class, could be effective for picking out parts likely to be good candidates.

The QRO'ers are fond of incredibly beefy gate drivers, e.g. IXDD414, 14 Amps of gate drive! Wow! But I think they're going the right way, using square wave gate drive where possible rather than starting with sine waves.

Tim N3QE
4711 2010-05-24 11:43:27 Graham Re: Power MOSFET gate drive
Tim:

An even bigger problem can be the Miller Effect, which is basically an
'amplification'
of the gate-drain capacitance. And, it is non-linear, dynamic
capacitance effect.

Gate capacitance is usually specified without the drain voltage in
motion, so
you only see part of what you have to deal with if you just look at gate
capacitance.

Since it usually scales with gate capacitance, your figure of merit is
still reasonable.

But, the actual effective capacitance that you have to charge and
discharge every
cycle is actually bigger than just the gate capacitance.

4712 2010-05-25 07:18:41 nm0s_qrp Re: Power MOSFET gate drive
4716 2010-05-25 20:01:01 ajparent1 Re: Power MOSFET gate drive
4717 2010-05-26 06:07:23 Tim Re: Power MOSFET gate drive
4718 2010-05-26 06:18:22 Tim Re: Power MOSFET gate drive
4724 2010-05-27 14:33:47 Ashhar Farhan Re: Power MOSFET gate drive
i have used a half wave filter at the gate input, eating up the gate
capacitance as a part of the half-wave. this makes the designs narrow
band but the amount of drive required is reduced considerably.

- farhan