[600MRG] Class E Amplifier Performance

[Ralph Hartwell]

> The second item attracting my interest is paralleling active devices, be
it
> tubes as we did in the last century, or solid state devices as we
presently do.
> Off the cuff we used to say "use matched tubes."  I assume designers are
still
> using that off the cuff statement.  I recall in the days of "hollow-state"
> amplifiers some designs used means to equalize the load when paralleled
> amplifying devices were used.


Running MOSFET's in parallel is fairly easy of you take care with the
circuit layout.

Here's a photo of a 3.1 MHz switching amplifier I designed for a plasma tube
driver.

http://w5jgv.com//special/Dual_MOSFET_Amp.jpg

Running into a 50 Ohm load, it will produce 500 watts key down for as long
as you want it to.  The output is not a sine wave, but for this application,
it does not need to be.

To ensure that the devices were balanced, I used a thermal camera to view
the temperature rise as the amplifier was operated at different power
levels. Because the MOSFET's have a positive temperature coefficient, that
makes balancing almost a no-brainer. It is not a perfectly designed circuit,
as the amplifier must operate with loads that depart considerably from the
nominal 50 Ohm design point. Because of this, the efficiency at full power
is only about 87%.

About measuring the device current - Inserting anything in series with the
source lead of the MOSFET will degrade the gain of the device and distort
the expected operating conditions, unless the inserted item has been
designed into the circuit from the start.  It's usually better to split the
power feed to the drains and measure the current to each device there.

Best DX & 73,

Ralph  W5JGV - WD2XSH/7