Towertalk
[Top] [All Lists]

Re: [TowerTalk] amplifiers for phased arrays

To: "Tom Rauch" <w8ji@contesting.com>,"Dave Bowker" <dbowker@mail.sjv.net>, <Towertalk@contesting.com>
Subject: Re: [TowerTalk] amplifiers for phased arrays
From: Jim Lux <jimlux@earthlink.net>
Date: Fri, 04 Mar 2005 18:02:11 -0800
List-post: <mailto:towertalk@contesting.com>
At 05:08 PM 3/4/2005, Tom Rauch wrote:
> > 4 SDR-1000 digital radios  (w/100 Watt output) - 4@ $1300
> > 4 SGC500 500W amplifiers - 4@$1200
> > 4 high power computer controlled tuners  4 @ $500
> > 4 radiating elements (one nifty approach would be a pair
>of switch
> > selectable horizontal dipoles at the top of each of 4
>middling high posts)
> > PC with necessary multichannel sound card interfaces $1000
>
>Sounds like a great system if you want less significantly
>than 1500 watts output, pee-poor IMD, and have time and
>money to burn.

Why would you have less power with multiple amplifiers than with a single 
amplifier (assuming comparable radiating elements, of whatever form)?  The 
"antenna loss" is mostly geometry determined, and in an apples to apples 
comparison (i.e. no fair comparing a 10 foot whip on the ground against a 
dipole at 100ft) the losses will be
comparable.  I contend that if I were to put 3 aluminum tube dipoles at 100 
ft, and drive them with 3 separate amplifiers of 500W each, the radiated 
field would be comparable to that achievable with a single 1500W amplifier 
driving the same 3 pieces of metal (whether coupled by a network, or by 
mutual impedance).

IMD is mostly a function of what's available (off the shelf) for 
amplifiers.  Essentially all solid state amps run about $2/Watt, so I used 
the SGC as a specific example.  Most of the tube amps available in lower 
powers have fairly low gain; the anti CB-linear legalities, drives design 
decisions about what people sell.  And, in fact, that's what prompted my 
original post.  I'd love to find a suitable modular amplifier for this 
application that has better performance (and doesn't involve sending many 
kilobucks to Amplifier Research (Model 500A100A for a mere $32K would do 
nicely) http://www.amplifiers.com/post/500A100A.pdf although they don't 
give any real performance specs for IMD, etc.)




> > Probably $15K by the time you get it all put together.
>Compare that to a
> > big tower and a SteppIR or two and a 2kW linear.  It's in
>the same general
> > ballpark.
>
>Except the single amp and antenna would have much better
>IMD, the power would be much higher, and it would be a less
>complex and less expensive system. It also would not cost
>near 15K to do this with a conventional setup.


Less complex, yes.
Less expensive, probably.
But, the conventional set up isn't going to be, say, half as much.




> > 10-15 years down the road, the potential is for small
>modules, costing
> > perhaps $500-600 each incorporating the digital stuff and
>a suitable 100W
> > or so PA, integrated with some reasonably compact
>radiating element.  Put
> > 10 or 15 of those modules on your roof, along a fenceline,
>on a pair or
> > three towers, etc.
>
>You'd still never have 1500W power. Not unless the only
>phasing you intended to use was a bi-directional pattern
>with high sidelobe levels and low gain. And any compact
>antennas would still be inefficient.

I don't think so.

In a compromise installation (the roof of a house), ANY antenna is going to 
be inefficient, big, small, compact or otherwise. I contend that in such a 
situation, the phased array has more potential to be useful than any single 
antenna connected to a single amplifier.  (you have the ability to reduce 
the average reactive near field energy, so the loss due to surroundings is 
less).

As far as efficiency goes, there's no requirement in amateur radio for high 
"power added efficiency".  Worrying about efficiency is only because you 
have a "max power to the feedline" regulatory limitation, and you'd like to 
radiate as much of that as possible.

All that matters is the power to the feedline(s), and in the case of an 
amplifier where the radiating element is part of a resonant circuit in the 
amplifier, one can make a case that IR losses in the antenna shouldn't be 
counted against your 1500 PEP (any more than IR losses in the plate circuit 
of a conventional amplifier should be counted).  The regulatory precedent 
is in things like cell phones, garage door transmiters, RFID systems, and 
the like, many of which use components in the final amplifier as the 
radiating element (to reduce cost)).  Near field losses are a bit more 
dicey in a regulatory sense.

I don't think I'd ever claim that an antenna (of whatever kind) 10 ft off 
the ground is going to be competitive with an antenna 100 ft off the 
ground.  However, there are large areas of the country, with lots of hams 
(and potential hams) who cannot put up a 100 ft high antenna, but CAN put 
up a 10 or 20 ft high antenna, particularly if it's visually inconspicuous.

The phased array approach provides a way, when constrained by such 
restrictions and  circumstances, to get more performance.  Sure, it's 
expensive (today). Sure, it's complex (today).  Both of those could change 
in the future.  Maybe not.  Maybe the only future for hams is for those who 
live on 10 acres with accomodating neighbors.  One thing is certain, 
though.   There WILL be more installation restrictions on ham 
antennas.  There WILL be more environmental RF noise to deal with.  If you 
want to operate on HF, new approaches are needed.  I won't say that phased 
arrays (whether amplifier per element or not) using insitu calibration are 
the panacea, however, it seems that working on this approach is as useful 
for keeping HF operation alive as any other activity.




>A simple four square with near optimum phasing has a power
>division of 34%, 31%, 31%, and 4%. What does that do to the
>idea of using 4 375w modules? It kills it.

"near optimum" by what criterion?  Forward gain? Radiated field strength in 
the desired direction?  Have you attempted to find a phasing which radiates 
375W from each element?  How much worse (if at all) is this than what you give?

The "simple 4 square" also isn't a good model for the kinds of installation 
I'm aiming at.  Unevenly spaced elements. Interacting objectes in the near 
field.

When you start to throw non-ideal locations and installations into the mix, 
the "penalty" from constraining the optimization to "equal power per 
element" isn't all that great.



>You'd need three 500 watt modules and one 50 watt module.
>They would also have to tolerate RF impedances varying at a
>cyclic rate on the output ports without undue distortion.

Why would the impedance be varying?  The frequency is constant. The phasing 
is constant. The mutual impedances are constant. No cyclical variation to 
be seen (except when you scan the array, but that's another story).


>Also if this fantasy array were ever built, you could NEVER
>use an array with negative resistance element without
>intentionally dumping (wasting) power.

This is true.  The real question is whether you can always find a phasing 
that doesn't have negative power.  This isn't something that can be easily 
answered analytically (except in trivial cases..)



>  You also face the
>severe technical problem of the amps being subjected to
>power returning on the feedline due to mutual coupling from
>other elements.

Clearly, one needs to supply reactive power to each element (i.e. you need 
a tuning network of some sort).


>I can easily imagine the headaches that
>would cause, especially when SS PA's by themselves are awful
>IMD producers!
>
>I can't think of any advantage in using an overcomplicated
>system like this in an HF application. Not one.


Constrained environments are your advantage.




Jim, W6RMK 


_______________________________________________

See: http://www.mscomputer.com  for "Self Supporting Towers", "Wireless Weather 
Stations", and lot's more.  Call Toll Free, 1-800-333-9041 with any questions 
and ask for Sherman, W2FLA.

_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk

<Prev in Thread] Current Thread [Next in Thread>