I guess that there are two design mistakes:
1. the supply voltage to the drains of the transistors is fed
via individual chokes, not the "balancing" dual winding choke
2. the input cpacitance of the FET´s do a big harm
Please have a look on the often cited WA2EBY amp: he uses the
choke (T2) on the output and in addition he is using inductances
from the gate (L1, L2) to the bias feeding resistor (27 ohms).
To my understanding the inductor forms a parallel resonator with
the input capacitance of the FET thereby flattening the
In Part 2 of Mike Krossors article (fig 4) the "bad" frequency
response of "normal" IRF510 amps is clearly shown, so it seems
that L1 and L2 togehter wit the series inductance of the input
balun are upgrading the performance.
Please have alook on the Philips (NXP) app note NCO8701. In
this app note they are also using inductors in the input and
even from the drain to the balun on the output. Design formulas
are also given there.
I am not sure if the drivers are "good" enough for driving the
high capacitive input of the IRF510.
I recently read an article (EDN) about the 2N2219 als vidoe
amps, they do not perform very well, althoug I remember that
this transistor (two in parallel) was used in a 2 m AM
transmitter (german company Semcoset, MBS22?, about 1969, the
year I got my ham radio license). A 2N3866 is preferred even for
video amps in the 50 MHz range!
What about a BFG135 or similar for the driver? Or: RD00HS ????
I am urgently waiting on my µBITX ordered on Dec 24 2017.
Am 25.01.2018 um 09:01 schrieb
I know it will increase the price, but how about
switching over to the RD16HVF1 which would allow the rig to
run from 12V doing away with those wanting more power to
increase the voltage to the PA?
Would also do away with the PA heatsinks if the case is
metal and it is used as the heatsink.