Roy et al:

I've always wondered why no one ever tries to use the constant resistance LC
lattice all-pass
networks to make variable phase [delay] networks for HF antennas.

In the simplest LC all-pass lattice there are two matching inductors in the
"through"
arms and two matchng capacitors in the "cross" arms. [Or an equivalent
version with two
"through" capacitors and two "cross" inductors.]

When terminated in it's designed constant resistance R such a network also
has a constant
input driving point impedance of R.

In practice, the transfer function of the LC constant resistance lattice
network has a flat magnitude
over very wide frequency ranges, and twice the phase shift characteristic of
a single LC network.

Potentially the phase shift could be made variable by mechanically or
electrically
ganging the two "cross" capacitors together so that they could be tuned in
synchronism
to continuously vary the phase shift while maintaining the constant driving
point
impedance/resistance. Or vice versa, the inductors could be ganged.

For some range of pole-zero locations the full lattice can be unbalanced
into a bridged-Tee
network which is perhaps easier to tune than the lattice. Indeed the
transformer assisted
half-lattice is also a candidate with easier tuning.

Good information on LC lattice networks is available in the "older" network
synthesis litterature.

cfr:

Louis Weinberg, "Network Analysis and Synthesis", McGraw-Hill, New York,
1962.
Shelved at you technical library under LC Shelf Call TK3226.W395

or...

Ernst A. Guillemin, "Synthesis of Passive Networks", John Wiley, New York,
1957.
Shelved at LC Shelf Call TK3226.G84 or Dewey 621.319.

For some reason folks interested in phasing antennas have not seemed to be
interested
in the constant resistance all-pass LC networks. I don't know why? They
seem to meet
all of the relavent requirements. i.e. a constant resistive driving point
impedance, and a
potential of continuous 360 degree variable phase, while maintaining the
constant driving
point impedance.

--
Peter K1PO
Indialantic By-the-Sea, FL.

"Roy Lewallen" wrote in message
...
The only variable delay lines I've ever seen with the delay variable
more than a couple of nanoseconds are some units used in an old
Tektronix instrument of some sort. They're awfully rare.

Of course, even transmission line delay lines produce a delay equal to
the electrical length of the line, and a voltage or current
transformation ration of one, only if terminated in their characteristic
impedance, or a few special cases. And the impedances of elements of
efficient arrays vary all over the map as phasing is changed. But a
variable delay line (most often implemented as binary-weighted lengths
of transmission line that can be switched in and out) can be practical
if the array elements are electrically very short and/or lossy, to swamp
the effects of mutual Z(*). That means it's practical for a receive
array but not so practical for transmitting(**). You'll find more about
this in Chapter 8 of the ARRL Antenna Book.

You might be able to make use of some of the tapped digital delay lines
(for receive only or very low power levels, of course) if you can
compensate for the insertion loss and frequency-dependent
characteristics. Another option is L networks. Although these can be
adjusted for various delays, it would be tricky to make an adjustment
arrangement that would adjust the delay without also changing the
voltage and current transformation ratio and the equivalent
characteristic impedance -- both elements would have to be adjusted
together in a particular way. Of course, you could make a series of L
networks with binary weighted delays and switch them as you would
transmission lines. The disadvantage is that the L network delay and
equivalent Z0 will change with frequency, while those of real
transmission lines won't.

(*) It's also practical if there's a very large number of elements, so
that each element is effectively in the same environment as all the
others. I worked on a radar using this principle back in the '60s. It
had switched delay lines to steer a transmit array of 10,000 separate
transmitters, each with its own antenna, and a receive array of over
5,000 separate receivers, each with its own antenna. It was actually the
phasing of the Tx and Rx local oscillators which were controlled,
though. Last I heard, it was still in operation.

(**) If you build a receive array where the signals are combined in a
low-loss fashion, you encounter exactly the same problems as for a
transmit array. Lossy summing networks can stabilize the impedances and
simplify the job just like lossy elements can.


Roy Lewallen, W7EL

Jack Twilley wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm interested in experimenting with phased arrays, but I can't find
any variable delay lines. I don't want coils of transmission line --
what I'd like is a little widget that can be tweaked to change the
amount of delay, but I don't see any of that for analog stuff like
transmission lines, just digital stuff like TTL logic. Help?

Jack.
- --
Jack Twilley
jmt at twilley dot org
http colon slash slash www dot twilley dot org slash tilde jmt slash
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (FreeBSD)

iD8DBQE/NBmVGPFSfAB/ezgRAgskAKCbQFMITmGhGItOA3RksbCzp1D+hwCg2AgT
uzxFUsx4WhIZ9U6SkYJ+j1k=
=UNrg
-----END PGP SIGNATURE-----