How about considering the Cgs (gate to source capacitance for better working on HF.
?May be it might matter as one would try other higher frequency bands.
Incidentally,? I saw the ebay design of 45watt linear PA kit that used a pair of IRF520 with 2sc1971 as driver and 2sc3357 as? pre-driver.
?
Few people have tweaked the design for almost flat gain across the HF band.
regards
?sarma
?vu3zmv
On Friday, 17 November 2017 5:45 AM, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:
I'll agree that Rds-on is mostly lost in the noise here. But for the sake of starting a good argument, concede that it represents the inherent resistance of the FET's silicon.?
If we have 25 watts of clean fundamental RF going into a 12.5 ohm load from the FET's drain the RF current is 1.414 rms, and power lost due to 1.5 ohms of Rds-on is 3 Watts.
The FET is typically under 50% efficient in generating those 25 Watts, I'm not really sure how that plays out with respect to Rds-on, but suspect it adds another ~12% hit there as well.
These IRF parts have limited thermal transfer from die to tab. I suspect if the gate voltage is high enough to put it into this avalanche condition, it will already be smoking due to the heat anyway. ? So just keep that drain current in check.
On Thu, Nov 16, 2017 at 03:11 pm, John Backo wrote:
Rds(on) applies mostly to very low frequencies (15 KHz) in switching power supplies. We are working in the small linear region which is never "full-on" operation. Rds(on) is incidental for RF applications. In fact, it applies in the "avalanche" condition.
