[BC] 50 Ohm Load...

[RichardBJohnson at comcast.net]

Switch-mode RF amplifiers do not work well into a
capacitive load. This is because the peak current
while switching a voltage onto a capacitor is exceedingly
high. That is why these RF amplifiers run best into a
slightly inductive load. The rate-of-rise of the current is
limited by the inductance.  If the inductance is too high,
the current will not increase to its peak value before
switching occurs again (for the next cycle) so there is
a limit of how much inductance is allowed. Note that I
am writing about the switch-mode transistor circuit itself,
not the output of the transmitter. This slightly inductive
load, plus the requisite harmonic rejection is accomplished
by fixed value components within the RF output circuitry.
They should have been adjusted at the factory for the correct
values when the amplifier is terminated in a 50 ohm load.

Normal aging of the output network components can change their
values slightly. This can change the terminating impedance into
which the amplifier will operate most efficiently. Because of
the numerous sign changes within the output network, it is not
possible to conjecture at a distance whether the load reactance
should be positive or negative to provide the best possible
load for the RF switches. Use of an additional tuning network
greatly improves ones opportunity to find the "sweet spot"
at which the transmitter runs best.

Note that the sweet spot is well known to broadcast engineers
who tune vacuum tube transmitters as well. The best efficiency
always seems to be when the plate tuning is on the inductive side
of resonance and, in fact, Nathan Sokal actually patented the
effect! http://www.arrl.org/tis/info/pdf/010102qex009.pdf
I leave it to the reader to consider whether or not such a
patent is valid.

ATUs are normally quite broad unless the tower is extremely short
or extremely tall. The sidebands from HF modulation would not
normally show up on a reflectometer at the ATU. However, once you
add additional tuning networks, such as an external match-box,
you increase the Q of the entire system and could possibly have
some sideband rejection. Note that 1/3 of the power is in the
sidebands so a large amount of reflected power during modulation
should be investigated. It might be a broken guy insulator that
arcs over during modulation, considerably detuning the antenna
system  during modulation --or worse. This kind of problem has
been found to be the reason why one AM station had bad sounding
modulation and "nobody could fix it." When I investigated, I
found that the base insulator was cracked, water had seeped
in...,etc.

Directional antenna systems are quite another problem. It
has been claimed that it is impossible to not have some kind of
sideband rejection in a complex array. Certainly, using modern
network theory to design one, rather than the old-fashioned
"tapped-tank" construction will go a long way towards reducing
the problem.

Cheers,
Richard B. Johnson
Book: http://www.AbominableFirebug.com/

----- Original Message -----
From: BOYDSIRBOYD at aol.com

 
bill

I would then read the input of the ATU at Fo and adjust it for  50J0, 
which was all I could do on TX's that didn't have a reflected power  
meter.

After I set the ATU input to 50J0 I would tune / load the TX  into the 
ATU on the TX's that were tunable.

So far the TX's with  fixed outputs (BE AM-1A etc.) have had reflected 
power meters and I use that  to tune the ATU on CW for min reflected to 
accomadate the whole tunnle thing  that you suggested, which I think is 
a very good explaination for why the TX  doesn't like 50J0 at the ATU 
input.

Ron  D