My 2x 3-500z homebrew never did a very good job on input matching on 10m;
couldn't get it below 2:1 VSWR and the power output showed it...750W out
with 2500V on the plate.
I have a network analyzer at home so I decided to start sweeping the input
circuit one "leg" at a time and see why no setting of the variable inductor
in the input pi would solve the VSWR problem.
For one thing, don't ever use Octal-plug based relays for T/R switching. I
seem to remember seeing that somewhere but I verified it via measurement...
Anyway, so that's out of the way. I replaced all the coax and started from
square one, the input connector. Here's some of the measurements I made,
all values at 28 MHz.
1. 47ohm resistor directly connected across the backside of the input
SO-239 connector: 29dB RL.
...well, you'd hope so! For those not conversant in return loss, that's
<1.1:1 VSWR.
2. Used a 15" run of RG-58/C (MIL-QPL version) from input connector to
input circuit. For grins, connected same 47ohm resistor across far end of
this coax. Approximately 3/4" shield pigtails on either end of coax and
resistor leads trimmed to 1/4". Found that you gotta chassis-connect both
ends of the coax shield (no surprise there either) for it to work.
Measured 18dB RL. (That's about 1.2:1).
So it's a bit disappointing to lose 11dB of RL just in a 15" run of
coax...surely 3/4" pigtails aren't THAT big a deal?
The input switch to select the tuned pi's is a standard Centralab-looking
ceramic wafer switch, double pole.
3. Connected center of coax to 1st pole of input switch common, grounded
pigtail, put 47ohm resistor from selected position to ground. Measured
14dB RL.
So there's some reflection in the switch wafer/metel.
4. Jumpered selected position across to 2nd pole of input switch, put 47ohm
resistor from selected (2nd pole) position to ground. So now we're going
thru both poles (wafers) of the ceramic switch. Measured 11dB RL. That's
roughly 1.75:1 VSWR.
On top of that, the phase of the reflection is nearly zero. So the input
impedance "seen" at the input connector is around 85 ohms, resistive. I
can't even compensate with a single component!
Well shoot...if the switch screws it up that bad all the tuned circuit in
the world won't matter! Is there some reason why not to use this type of
switch on the input? It seems a natural. I do have CRL-9000 series
switches (the old Radiokit catalog seems to think they're pretty
good)...should I use these instead? Kind of an expensive switch for a
low-power input section. Or just straight to PIN diode switching?
Any comments, observations, experiences, greatly appreciated. Thanks for
the read!
------------
Scott Townley
nx7u@primenet.com
------------
Collector of:
Stoddard Aircraft EMI/RFI receivers and accessories
Big Parts for that Big Linear Amp
70's era RF test equipment HP/GR/Tek
Radio-related technical reference material 1940+
...anything else that will keep me off the streets at night
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|