[R-390] R-390A sensitivity measurements

[DJED1 at aol.com]

A shield room is not a bad idea, but it may be too expensive for most of us.  
 I played around with a couple of items tonite, and the issue of shielding 
came to the fore.   first, I made an autotransformer on a toroid which matched 
125 ohms to 50 ohms (both unbalanced).   I soldered it to a piece of coax and a 
connector and put it on the back of the receiver... the external noise coming 
in on it was 10 dB higher than that of the receiver.   Definitely not useful 
unless built in a well-shielded case.   So I've put that on the back burner.
I thought about the very good comments from you guys, and did a few 
calculations which turned up some interesting results.
First, I concluded that adding an 82 ohm resistor in parallel with the 
receiver, to give the generator a 50 ohm load, and adding a 75 ohm resistor in 
series, to give the receiver a 125 ohm load, both form 2 to 1 voltage dividers 
between the generator open circuit voltage and the voltage across the receiver 
terminals.   Thus either will allow us to use the generator meter to read the 
sensitivity.   I haven't tried a pad which matches both, but requires a 
correction to the meter reading.   That also awaits a shielded box.
I also did some calculations on the noise floor assuming a 125 ohm resistor 
at room temperature, and the 9 dB noise figure I measured on my receiver.   It 
comes out to a voltage of 0.4 microvolt for 10 dB S/N in a 4 Kc bandwidth.
I then reran some of my measurements using both the URM-25 and the 8660.   
For measurement with the carrier turned on or off, either with the BFO on or 
with 30% modulation, I got measurements of 0.3 to 0.4 microvolt.   So everything 
seems to hang together, including agreement between the URM-25 and the 8660.   
The only difference is that below 0.5 microvolt on the URM-25 the readings 
don't drop down much as I turn the attenuator.   The 8660, on the other hand, 
just keeps going down into the noise.   So shielding is definitely an issue for 
the old generator.   However, I think I'll keep her.   I was pleasantly 
surprised at the accuracy- the two generators agreed within 1 dB at levels of 1 and 
5 microvolts.
I got sensitivity of 1 microvolt in 4 KC bandwidth when I measured with the 
modulation turned on and off.   This is the specified procedure for AM.   I'm 
now satisfied that I understand the methods and results.   The only unexplored 
issue is whether a balun will make any significant difference.
Ed