Hi Steve,
If you are interested I had really good results recently with a homebrew
RLB, in fact so good that I doubted the results and had to redo them several
times before I was comfortable I hadn't made a mistake, and I probably still
have but anyway it works heaps good enough for me. Basically using the non
lab standard test equipment I had available it shows a directivity of better
than 40db from about 1 to 500MHz. I found the best results by using a more
or less current balun with coax, and a balancing extra bit of coax, and
paying close attention to keeping everything balanced/symmetrical and having
the reference connected with a connector identical to the unknown, rather
than just a 50ohm resistor directly connected. See writeup in December issue
of Nerg News
http://www.nerg.asn.au/NERGNEWS/NN200512.pdf

73 Paul VK3DIP.



wrote in message
oups.com...
I have just built an HF return loss bridge according to the info in
"Solid State design for the Radio Amateur". While it works ok, I was
hoping for a bit better performance. Clearly the "balun" doesn't have
enough inductance to operate well below 10 MHz or so (10 turns bifilar
#30 AWG enamelled wire on Amidon T23-43, as specified in the book -
twisted pair, which isn't specified one way or the other in the book).
This can probably be largely fixed by using a larger ferrite core
(T37-43's available in junkbox). I am interested primarily in the
1.8-50 MHz range, though I wouldn't complain if it worked on 2m too.

But even at 30 MHz a reasonably good microwave 50 ohm load gives only
about 28 dB apparent return loss...Not bad, but I might have expected a
bit more.

And an open and short give about 2 dB different signal levels at 30
MHz. That is with big pads (20 dB attenuation) on both the signal
generator and detector. The detector is an HF receiver with a step
attenuator used to maintain a constant S-meter reading.

The circuit is built on a small PC board using construction techniques
typical for the UHF or low microwave range (except that the test port
connector is an SO-239), and is enclosed in a shielded box. Each of
the three 50 ohm resistors is made of two 100 ohm 1206 chip resistors
in parallel and measures between 50.0 and 50.3 ohms at DC.

Has anyone with experience with this circuit any suggestions for how to
tweak it for best performance ? What accuracy level have you achieved
? Do you know where I might find an error analysis for this circuit ?
Or if I am to think about errors myself, does anyone know how to model
the balun in SPICE ?

73,
Steve VE3SMA