I have been thinking about this TX phased array idea for a couple of years
now, but as I said before, I have been working like a crazy man trying to
get a software business going so I have been putting it off. I was not as
brave as Jim in my vision, however. I made a number of simplifications that
may (or may not) make it much more practical to build with simple
electronics and a lot of software. These limitations are:
- CW only. For me this is fine, since this is the only mode I work.
- 160 and 80 meters only. These are the bands where verticals are usually
the only solution.
- QRP. This is my major interest these days.
- At the base of each vertical element is an rx preamp and a final TX amp.
- Each vertical element has a separate ADC/DDC stage which all feed into a
single DSP.
The CW, QRP, and upper frequency limitations greatly simplify the amplifier
electronics. I envisioned a separate opamp-like current source feeding
directly into the base of each non-resonant vertical element. I am assuming
that one can build a clean current source amp that can deliver a few watts
of 1.8 Mhz CW into a reactive load. For QRP, the sum of the output of all
the amps will be 10 watts or less, so you can design with lots of overhead
and little concern for efficiency. Below 2Mhz, you might be able to think
more like an audio amp designer than an rf amp designer.
As for pattern generation, I figured that you would try to place your
elements in a desirable configuration, but if real estate concerns preclude
that, you do the best you can. Then you measure the position of each
element in respect to each other. Then you apply a signal to each element
one at a time while receiving on each of the other elements. Since you know
their positions, you can have the software solve for all the phase delays
due to coax, etc from the phase measurements in each RX channel. From that
you can easily optimize for power levels and phases for tx beamforming in
any desired direction based on whatever your criteria (i.e. f/b, gain, null
in a particular direction, etc.). Similar optimizations can be used for rx
beamforming as well.
With each element being non-resonant and being fed with a current source
allows for a broadband system and reduces mutual coupling between the
elements.
Each vertical element would have a preamp/buffer at its feedpoint which
mostly serves as a broadband impedance buffer between the non-resonant
element and the coax. The preamp may only need to run at unity gain, so it
might not be the weakest link in the rx signal chain. For 160m, where the
atmosphere dominates the noise figure, perhaps the tx amp can be the rx amp
as well (semi break-in by switching in different amp feedback components for
RX than for TX).
All the coax runs go back to a single box which contains the DSP, the
ADC/DDC RX channels for each of the elements, and the UPConverter/DAC tx
channels for each of the elements. The ADC/DDC channels all funnel into the
DSP where the complex demodulation and signal combining take place.
Naturally, what I just described is too limited to be commercially
interesting (except for the half dozen 160m/QRP/CW people out there), but it
would be fun to play with. It also might give me an edge for 160m QRP.
Since all the beamforming is being done in the DSP, you are free to form
more than one beam or null at the same time. You could be forming a maximum
gain lobe in the desired direction while interactively nulling strong
signals from other directions and doing active noise elimination.
Oh yes, this is Towertalk and not Antennex. One can use this arrangement to
phase up a set of existing towers, regardless of how they are arranged
(within reason). You could augment the towers with additional wire elements
and other things to make as many elements as you think you need. Then let
the system I described figure use the towers and wires for beamforming.
It's a nice pipe-dream, of course. Besides the functional limitations I
have already described, can anyone see other weaknesses? Is this too far
off topic? Perhaps we need a Software Defined Antenna email list.
Regards,
Dudley - WA1X
----- Original Message -----
From: "Rick Karlquist" <richard@karlquist.com>
To: "Jim Lux" <jimlux@earthlink.net>
Cc: <towertalk@contesting.com>
Sent: Saturday, March 05, 2005 9:41 PM
Subject: Re: [TowerTalk] amplifiers for phased arrays
> Jim Lux said:
>
> > that some ham, who IS an amplifier designer, would take that on.
There's
> > a
> > lot of new amplifier design approaches out there that are efficient,
have
> > better performance, etc., but the ham market (being relatively tiny)
seems
> > to be willing to buy the same old triode, bipolar, and fet designs
> > that have been around for decades.
>
> Besides the DSP predistortion you mentioned, what are these "new
> amplifier design approaches?" Do they work with an arbitrary load
> impedance? What kind of efficiency are you thinking of?
>
I was thinking, in general, of things like envelope elimination/replacement,
or various modifications of outphasing, all of which can use high efficiency
stages (like, for instance, Class E/F or inverse). The real high efficiency
stuff is narrow band, but maybe it might be more cost effective to, for
instance, have a separate driver for each band with its own tuned tank. The
semiconductor technology isn't quite there for real cheap transistors at 30
MHz, but down at 7 or 14, you have things like Rutledge's kW amp using a
high power switching FET. You might need some clever way to retune the
resonant circuits across the band.
Another intriguing approach, for signals with multiple carriers, each coded
(like COFDM), is an amplifier for each carrier (since the carrier is
constant envelope, even though the overall signal isn't, you can get higher
overall efficiency)
On the subject of impedance matching.. If you're closely coupled to the
antenna, there's no particular reason to ever go to 50 ohms. Even more
intriguing is the possibility of supplying the reactive power with the
amplifier, rather than passive components (probably make your efficiency go
to heck, though).
I think a reasonable efficiency goal would be 70% or 80%, wall plug to RF.
For all I know, these techniques would be impractical, or have some
fundamental problem. However, looking over the recent literature in MTT,
for instance, it seems that there might be something to tinker with. You're
going up against decades of historical development on tube amps and fewer
decades, but still substantial time, on solid state. And, ham modulations
don't resemble modulations currently popular in the development world.
There HAVE been some hams poking at new amplifier topologies, but, as far as
I know, they've been looking at CW (or AM), and not the general SSB linear
amplifier.
> BTW, DSP predistortion is not, strictly, an amplifier design technique.
> It has to be applied to the whole transmitter. It won't help with an
> amplifier module intended to amplify the output of an arbitrary
> exciter.
Certainly true, unless you do something crazy like digitize the output of
the exciter...but then, modulation is the the easy part of a transmitter..
it's the amplifier that's the challenge.
>
>
> Rick N6RK
> (who IS an amplifier designer)
>
_______________________________________________
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
|