[Lowfer] SWR meters for LowFERs

[Clint Turner]

In the mid-late 1980's, Mark Mallory (WB7CAK) operate the MPM beacon 
from the Salt Lake City, area.  This was likely one of the very first 
beacons to reach that "1% efficiency" rating on the antenna system.  He 
accomplished this several ways:

- The antenna was on a large building with a metal roof.  The footprint 
of the building was far larger than that of the radials.  Having a large 
metal roof under you reduces ground losses somewhat.
- The loading coil was made from heavy Litz wire, appropriately sized 
for the frequency.
- The transmitter was a Class-E type and was quite efficient.  The 
output tank circuit was simply tapped to match the feedpoint resistance 
at the loading coil, thus sparing the losses of yet another "matching 
network."  A "web reprint" of the original Western Update article may be 
found here:
http://www.ussc.com/~turner/mpm_class_e.html
- The antenna consisted of 35 feet of Rohn-25 tower (welded to the 
building's roof and superstructure) with 15 feet of aluminum mast atop 
it.  The transmitter/loading coil was placed at the top of the tower - 
and the mast was insulated from the tower.
- There was a tophat on the top of the mast.  Mark considered the tophat 
to be part of the matching network and not part of the antenna.  (This 
is just one interpretation of the "15 meter" rule - a point that has 
seen much debate in the past, but for the purpose of this posting is 
irrelevent...)
- He used an "auto-tuner." - which *is* the point of this post.

As I recall (having seen and used the circuit) it was very simple:  A 
sample of RF voltage and a sample of RF current were taken, their phase 
was compared, and the error output sent to the servo driver.

The voltage was simply capacitively (or resistively) coupled from the 
transmitter output while the current was taken from the same line using 
a "1 turn" toroid transformer (not unlike one seen on broadband VSWR 
bridges or clamp-on ammeters.)

These two signals were then limited (taking care to do the threshold at 
zero-crossing, to avoid a phase shift) and then applied to a 4-quadrant 
multiplier.  I *believe* that he used just a 1496 (run at very low 
levels - just a few hundred millivolts - it will function as a 
4-quadrant multiplier) or an EXAR part to do this.  (I suppose that if 
you have square waves, an XOR gate would work, too...)

The result is a voltage that is somewhat proportional to the phase 
relationship between the voltage and current.  At resonance, these 
should be in phase with each other - a condition that is pretty easy to 
detect.  It is worth noting that the inductive coupling scheme typically 
introduces a phase error - but this simply amounts to an offset voltage 
on the output of the multiplier.

In his case, he just used a simple counter circuit to operate a stepper 
motor to adjust the variometer.  There was also an "enable" line so that 
the tuner didn't go bonkers trying to find a match when the transmitter 
was unkeyed.

The only trouble he had with this was that, occasionally, the antenna 
would get too heavily loaded with snow and go out of range - and the 
tuner would get "lost."  It could never automatically find resonance 
again because it was unable to detect the current component - because it 
was so badly detuned.  Some simple logic could have fixed that problem 
(which was fairly rare) I suppose.

The entire point of this is "If you want to tell if your antenna is 
matched, then look at the phase relationship between the voltage and 
current."

Getting a resistive match is another matter, however.  On the Class-E 
transmitter, it was pretty simple:  Load it at the design impedance (or 
the tap-point impedance) and set it up for the proper current 
consumption.  (e.g. 10 volts at 100 milliamps = 1 watt input.)  When you 
connect it ot the antenna, simply adjust the tap point on the 
amplifier's tank circuit (keeping the system at resonance) so that the 
target current consumption is once again obtained.  If you have done it 
right, then it usually comes means that a proper resistive match has 
been obtained.  If you are really clever, you can count turns on the 
tank inductor and get an approximation of what your feedpoint resistance 
*really* is...

Clint