[TowerTalk] measuring power, reference planes, and all that

[Jim Lux]

At 11:21 AM 9/5/2003 -0400, K3BU at aol.com wrote:
>In a message dated 9/5/03 10:22:30 AM Eastern Daylight Time,
>jimlux at earthlink.net writes:
> >>I'd like to know more about that.  Indeed, phase and amplitude matching
>(particularly with respect to nonlinearities and differing distortion
>products from each amp) is a challenge. It's a challenge today with any
>amplifier that uses multiple amplifying components, which is why the usual
>commercial design rule is use the biggest single device you can find that
>can put out the power you need.  I don't know about reradiation being an
>issue.  You're already reradiating in any multielement array, passive or
>active.<<
>
>Big difference between passive and active. Passive there is no problem, we
>are dealing with one signal being split and phased. With active array 
>there is a
>problem with induced and
>rectified/amplified signals from other elements causing distortions.

You're referring to signals from the output of Amplifier 1 getting into the 
output of Amplifier 2 and causing intermod or other problems?  Sure, 
isolation is always an issue with multiple amplifiers.  However, that's 
basically a design problem, not a basic theory of operation problem.  I 
make no claim that a box stock design as used today would be appropriate 
for an active phased array.  Certainly, in the similar, but somewhat 
simpler, case of combining amplifier modules in high power solid state 
amplifiers, you have to deal with module interactions (particularly under 
failure conditions). This is exactly the problem Wilkinson was solving way 
back when.  It's also a problem that phased array designers have to deal 
with on a day to day basis.



>Next
>time you hear station with mushy and wide signal, he is probably running two
>antennas with two amplifiers closely coupled (except propagation induced
>distortion).

I'd add to your statement: "probably running a poorly implemented 
combination of amplifiers and antennas".  Poor design and implementation is 
poor design and implementation, whether you're talking about SSB exciters, 
speech compressors, spurious oscillation in a tube amp, or an active phased 
array.  Phased array design is substantially more complex than a casual: 
power divider to two amps approach.

Look at all the horrible passive phased array designs for four-squares that 
have been published over the years (including the Wilkinson divider 
approach in the ARRL handbook a few years back).  And that's a "simple" 
design problem.  Even feeding two verticals is nontrivial.



>Again, what about receive??? Forget the argument of using piece of something
>and preamp. If there is no signal from the antenna, there is no amp in the
>world that can amplify it.

Kind of depends on whether front end noise or sky noise dominates.  Either 
your antenna physically intercepts the incident (desired) RF signal or it 
doesn't, and that's essentially determined by the antenna's effective 
aperture (or gain/directivity).  Whether you intercept the power with a 
single antenna of aperture X or two antennas with aperture X/2 doesn't make 
a heck of a lot of difference.  The two smaller antennas intercept half the 
power, but also only collect half the noise. A physically large antenna 
could be viewed as a number of smaller antennas with a power combiner. 
Whether you put an amplifier after the combiner, or before the combiner, as 
long as the amplifier noise is lower than the noise incident on the 
antenna, it doesn't matter.

At UHF, this is a big problem though... amplifier noise IS bigger than the 
pretty quiet environment, so physical aperture helps (why radio telescopes 
and the deep space network use BIG dishes).  You can combine preamps and 
small apertures at higher frequencies, but instead of getting a factor of N 
improvement, you only get sqrt(N) improvement(the noise combines 
incoherently, the signal combines coherently).




>Then there is requirement for Signal to Noise and QRM
>discrimination (pattern control) that is so crucial on receive. This is why
>antenna and its phasing is so important, works on TX and RX.

You bet.. you need to build a phased array for receive.  However, a receive 
phased array, in some ways is easier, in others harder.  You don't have to 
handle high power, but you DO have to deal with strong adjacent signals, 
front end intermod, etc.  As we in the digital radio business like to ask: 
Why doesn't someone make a cheap high performance 24 bit A/D converter at 
150 Megasamples/second so we can do away with all those horrible mixers, 
local oscillators, filters, and IF strips....


>It is easy to use the argument about the progress, but the reality is that
>things catch on if they are better, cheaper, easier and other -er.

When people first started using SSB, it wasn't any of those things.  Over 
time, it became so.  You're right, though, I don't expect every ham to rush 
out and buy a super duper phased array tomorrow, or even next year, or even 
in the next few decades.  I recall sitting in my grandfather's shack(the 
original W6RMK) when he and my father (ex W6CZA) were discussing whether it 
was worthwhile to invest in buying/building a SSB rig, when he already had 
a perfectly good AM rig with Class C PA and modulators, etc. right there, 
and most of the folks on the air were working AM.  That was over 30 years 
ago, and today, I doubt anybody sells a production transmitter that does AM 
but not SSB (CB excepted..)


>We are
>looking for that active, miniature, super-duper antenna that will fit on 
>the front
>lawn or in the attic and beat stacked beams. There are those miserable 
>laws of
>physics that are in our way.

Indeed there are laws of physics that can't be beat. The laws of physics 
basically tell you what the minimum physical extent of the antenna needs to 
be (i.e. the effective aperture), but don't tell you anything about what 
the configuration inside that extent is.  It's really more of a classic 
engineering problem where you have to trade off cost, efficiency (because 
you're legally power limited on Tx), complexity, etc.

Interestingly, one of the most significant physics problems, and one that 
fancy phased arrays can't help, is the effect of ground losses.  The only 
solution is height (or that proverbial salt marsh on top of a hill).  So, 
while the antenna on the tower, and the kind of cables we run up there 
(power and Cat5?) might look pretty different, you'll still need a tower.

The virtue of the phased array is that in a compromise situation (i.e. no 
tower), the active phased array can potentially provide better performance 
than existing solutions.



>Good luck and more power to you, we are standing by and waiting for that
>magic solution and the bottom line.

No magic involved, just electronics.. And, for the first few goarounds, 
while we're on the learning curve, the bottom line's going to be big.