[R-390] R-390A IF gain setting and S/N measurement

[Charles Steinmetz]

David wrote:

> It's been stated many times by Roger Ruszkowski - who ought to know - that the 150uV/7V IF gain spec in the manual is far above the gain that yields best S/N.  (My radio agrees.)  Were those over-hot IF's simply following the manual, or were they cranked even higher?

I concur with Roger that 150uV/7v is too hot.  However, you can't back 
the IF gain all the way down to the point that maximizes S/N without 
also compromising the overload performance of the receiver.  This is 
because the stage that overloads first is upstream of the IF.  If you 
set the IF gain lower, the AGC just cranks up the overall gain to 
compensate.  This means that the RF gain runs higher and, therefore, 
that the mixer overloads on weaker signals than it did before the IF 
gain was lowered.  Since the 390/390A front end is prone to overload to 
begin with (despite what sellers say in their ebay listings), you don't 
want to compromise its overload performance any more than absolutely 
necessary.

This is the art of balancing a radio's DR -- to get it sensitive enough 
that the practical limit is always atmospheric noise coming in on the 
antenna, never the front end's self-noise, while keeping the overload 
point as high as possible.  This is the fallacy of pursuing maximum S/N 
and sensitivity -- you throw away overload performance that you 
desperately need, to gain sensitivity that is useless in practice 
because the atmospheric noise was already 10-30dB louder than the 
radio's own input noise.

Once some basic design choices are made -- in particular, the number and 
general type of active devices and the impedance levels of the plate 
loads and tuned circuits -- the best possible DR has been preordained. 
All you can do beyond that is (i) make sure you actually get the best 
possible DR out of the parts (i.e., make no design blunders), and (ii) 
slide the DR up and down the input signal range to optimize the tradeoff 
between weak-signal and strong-signal performance.

In the real world of antennas, QRN, and QRM, there really isn't much 
room for debate about what constitutes the best compromise for any 
particular value of DR the radio has.  And to change DR, you need to 
change some really fundamental things about the topology you are using, 
such as the impedance levels and standing current at each stage.  As 
Larry is finding, simply changing the stage gains by using different 
tubes doesn't buy you anything if the radio was well designed in the 
first place.  After you adjust everything so it works again, you're 
right back where you started -- at the limits of the fundamental circuit 
elements and stage impedances the original designers chose.

I once spent a few days with several 390s and 390As, a suite of signal 
generators and combiners, a spectrum analyzer, and a distortion analyzer 
to determine the optimum IF gain.  I can't find my notes right now, but 
I do recall it was less than the Collins spec but not by all that much.

As to the IF gain settings I've observed in radios "as found," it 
appears that once someone gets it into their head that more IF gain is a 
Good Idea, they are not bound by any sense of moderation.  The ones I've 
found with too-high IF gain are generally 6 to 20dB higher than 
150uV/7v.  I'd be interested to know how a radio set up according to 
Roger's "alternate procedure" compares to the "by the book" 150uV/7v 
with respect to input level and AGC voltage -- just how much lower is 
the IF gain?  Has anyone checked this after performing the alternate 
procedure?

Best regards,

Charles