Phase shifting in HF amplifiers

[John W. Brosnahan]

In response to K1GQ's question about phase shifts through amplifiers, I can
shed some light.  In the course of building many 49MHz 50KW pulse 
transmitters for wind profiling applications, we have built systems that
used as many as 64 transmitters.  Each transmitter drove 1/64 of a phased
array of approximately 13,000 dipoles.  (This was in Alaska)  An active
phase monitoring and correcting system was used to track the phasing of
all of the amplifiers.  But I was curious just what the phase variation
of one of the transmitters might be as its tuning changed with time/temp and
phase variations through the transmitters in general.  I performed
phase testing on a next generation version of this transmitter which
consisted of a 4-stage solid state 300 watt predriver, a 3CX-800A7
driver (4KW pulse) with typical T-network input and Pi-network output with 
air variable capacitors, and a final stage of a 3CPX5000A7 (50KW pulse)
using an air variable T-network input and vacuum variable Pi-network output.
 
Either of the tube stages could be thought of as a typical ham amplifier and 
with two stages in series any phase shift variations would probably be
worse than with a single stage.  I deliberately started mistuning the tube
amplifiers and watched the phase shift through the system.  To my surprise
the phase would only vary about +/- 3 or 4 degrees before the output
power started dropping very significantly.  To get +/- 10 degrees phase
shift (total,through two stages) required pretty severe mistuning. 
 
Based on my experiences with amplifiers at 49 MHz, I would expect two
Alphas to track within +/- 5 degrees or better when tuned in a normal
fashion.  This is small enough to have negligible effect on typical 
stacked arrays at HF.  I have two Alpha 87As and later this spring,
when life gets a little less hectic (trying to get another 49 MHz radar
into the field for a research project at White Sands Missile Range), I
will rig up a test set up and measure them for phase shift using an
HP vector voltmeter.  Actually only one is required to measure its
sensitivity to tuning adjustments, but I want to confirm that they both
have the same phase shift.  (I see no reason why there should be any
differences for the same type of amplifier unless major design changes
have been made in one.
 
Actually I recall seeing a couple of things in QST over the last 
3-4 decades on combining RF outputs of amplifiers both at HF and VHF.  
I think there was something about combining two of the old Heathkit 
mobile KW amps somewhere. 
 
Hope this provides some help.    73   John  W0UN