[R-390] R390A sensitivity and noise floor measurements

[jgedde at optonline.net]

After weeks of fiddling, tube swap optimizations, fixes, etc. I've got my
R390A performing extremely well.  I have an HP 8644A signal generator I
bought from the local scrap dealer for scrap value (it works fine) and was
able to make the following measurements:

 

Sensitivity: 0.55 uV for 10 dB S+N/N with a 30% modulated input.  It'll do
better with 100% modulation..

 

Noise floor: better than -137 dBm.  I can't go any lower with the 8644A
without additional attenuators (they appear to add noise of their own) and
my noise environment isn't great.  A screen room would be needed to go any
lower.  I fiddled a bit with the IF gain and got an improvement from -132dBm
with the standard IF gain setting at -7V at the diode load to -137 dBm.  The
judging factor was whether the signal could be detected.  At -137 dBm you
could just make out that the sig gen was on or off.  My spectrum analyzer
can go much lower in noise floor and confirms I wasn't just measuring RF
leakage in the sig gen: it can see -137 dBm no problem if I set up the BW
just right.

 

While I am not able to repeat Chuck Rippel's -143 dBm and Sherwood
engineering's 0.2 uV, I consider this quite good and I am very pleased.  But
to be fair, Sherwood doesn't state what they used as a source.  If it was
100% modulation, then I was able to get there.  I used the 30% modulation
standard.  I tend to doubt -143 dBm noise floor to begin with as I have
stated before.  I wonder if his sig gen is leaky?  Sherwood got -137 dBm as
did I.

 

Another good R390A day!

 

John