[R-390] R-390A sensitivity measurements

[DJED1 at aol.com]

I tried making some sensitivity measurements on my receiver, in which the IF 
gain was reduced per Chuck Rippel's suggestion.   I tried several different 
scenarios:
First, per the MIL spec, sensitivity is measured with an 8 Kc bandwidth and 
adjusted for a 10 dB change in output by switching the modulation on and off 
(30%, 400 cycles).   They also call for a resistor in series with the radio 
input.   I couldn't see how this made sense, so I followed the directions in the 
URM-25 manual.  It calls for a series or parallel resistor such that the load 
the generator see is 50 ohms.   I know both the URM-25 and the new HP need to 
be terminated in 50 ohms to provide the voltage indicated on their meter, so I 
put an 82 ohm resistor in parallel with the nominal 125 ohm receiver input, 
resulting in a 50-ohm load.   The results were interesting:
Using the method of modulation on and off gave a sensitivity of 1.9 
microvolts.   Not the result you often hear discussed with these radios, but credible 
for a radio where the specification is 3.3 yo 4.4 microvolts.   
I then tried the method I used in the past, of turning the modulated carrier 
on and off, and using a 4 Kc bandwidth.   Big difference-   0.23 microvolts.   
I tend to like this method because it seems more representative of an actual 
signal.
Finally, considering that we listen to CW and SSB as well as AM, I measured 
the sensitivity with the BFO on and a 2 Kc bandwidth.   Got about the same 
result- 0.22 microvolts.   I then reduced the signal level until the signal was 
just detectable in a 1 Kc bandwidth and got down around 0.01 microvolts.   
That's in the ballpark of the -143dBm noise floor that is quoted by some for the 
radio.
All in all it was an interesting exercise.   The measurements made with the 
specified method explain the specification of 3.3 microvolts, while I believe 
the method of switching the carrier on and off is more realistic.   Definitely, 
the CW measurement is more representative of real world conditions.   And I 
think 1/4 microvolt is adequate for almost all application in which I would use 
the receiver.   I'd be interested in hearing from anyone who has duplicated 
the test setup and gotten results better than mine.
(Incidentally, I did check the generator for leakage prior to making the 
measurements.   Detected a very low level with the receiver connected to a wire 
near the generator, but nothing when connected to the generator by a shielded 
cable.)
Ed