On Feb 11, 8:24*am, phaedrus wrote:
HI,

I have an old HP network analyzer that goes up to 1.3Ghz. I need to
make up some more cables for it, though, as the existing ones (not
original) are far too thick and inflexible for comfort.

I have to cut three identical 2-foot lengths (for phase matching
purposes) of some 50 ohm coax for this purpose, to connect the VNA to
its bridge. Can anyone suggest a suitable, readily available coax for
this job? The terminations will be N-types, by the way. I'd like
something flexible and easy to terminate, ideally. Just don't want a
cable that will change its Zo every time it gets a bit bent!

It's possible to buy very nice, reasonably flexible instrumentation
grade coaxial cables, but they are quite expensive. Hundreds of
dollars each for a 1 meter long cable is not unheard of. In general,
you don't want to bend your cables in too sharp a radius. That's
asking for trouble with pretty much any cable. If you can arrange
things so that you can install the cables and leave them alone, you'll
stand probably the best chance of having a stable system. You might
consider a bit longer cables if it lets you arrange them with larger
radius bends. I'd go for cable like RG-213/U, or better, RG-214/U if
you can find it. If you can get cable with solid Teflon dielectric,
that will likely be a bit more stable. In any event, I'd certainly go
for cable with solid dielectric, as foam dielectric lets the center
conductor move around too much.

Is you analyzer an HP8754, with the transmission/reflection test set?
If so, I understand exactly the issue you're dealing with. ;-)

Since you're only going to 1.3GHz, it's possible to build a quite
respectable set of calibration standards from SMA connectors and 0.1%
(or even 1%) 100-ohm 0805 SMT resistors. They'll serve quite nicely
for pretty much any ham work I can think of. Icing on the cake is if
you can get someone to measure them for you with a freshly calibrated
precision network analyzer. I have built these sorts of loads using
thru-hole SMA connectors intended for vertical mounting on a PC
board. I cut the center pin off so practically none is showing above
the Teflon insulator for an open, cut the center pin off to perhaps
0.01" and solder a pair of 100 ohm 0805 resistors from the pin to the
shell, radially symmetric, for a 50 ohm load, and just add either a
disc of thin copper (copper tape) or four radially symmetric wires,
about AWG28, from the pin out to the shell, down flush against the
Teflon insulator, for a short. I've made a few such 50-ohm loads and
find that it's practically trivial to get at least 40dB return loss
out to 1.3GHz, which is better than I need for any ham work I've ever
done. (Bewa "50 ohm" coaxial cable typically has a tolerance of
+/- several ohms...)

Cheers,
Tom