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

[Charles Steinmetz]

Larry wrote:

>     I believe that the real S/N of an RX is determined by the first 4 or 5 tubes and circuits in the RX, and those after that point really don't affect it (if they are working correctly).   Since the IF gain is in the 7th stage of the RX (3rd stage of IF), changing it should not affect the real S/N.  I have no way to measure it's real affect, but I don't hear any difference.

It should darn well be dominated by the first stage.  (Note that 
"dominated" does not quite mean "determined exclusively by," although in 
a well-designed receiver the subsequent stages should contribute only 
minimally.)  There are two provisos:

First, radios have sections and stages with different bandwidths.  If a 
narrow BW stage is followed by a wider BW stage, the later (wider) stage 
may dominate the noise in the portion of the passband outside the narrow 
passband of the earlier stage.  This is not generally an issue in 
well-designed receivers, because the progression of bandwidths is like a 
funnel -- it should get narrower as the signal proceeds through the 
receiver.

Second, in recievers that apply AGC to both RF and IF stages, the gain 
distribution may change dramatically from weak-signal to strong-signal 
conditions.  Again, in a well designed receiver, this should not be an 
issue -- but there are LOTS of receivers that do not get this right. 
Setting the IF gain of a 390/390A is a necessary step in ensuring this 
balance is right in terms of both overall S/N and overload performance 
(as you have found).

>     So why does it make a difference in the measurement?  And do we really need to consider S/N when setting it?

Yes, for the reason noted above.  It is critical to ensure that the gain 
distribution is correct for best overall S/N and best overload performance.

Note that you keep changing things and comparing apples to oranges.  As 
long as there is enough gain everywhere (and there is, in a 390/390A 
with all original tubes -- even with thoroughly whipped tubes that test 
"bad," except in the case of a completely dead first RF amp), you cannot 
change the overall DR by any significant amount by adding gain to any RF 
or IF stage(s).  All you can do is trade off low signal S/N against 
strong signal overload performance.  The DR remains stubbornly the same. 
  If you change tubes and then adjust the stage or section gains so that 
the overload point is where it was (in a properly operating, properly 
adjusted receiver), the S/N will be within experimental error of where 
it was with the factory setup.  You're right back where you started. 
Any difference in S/N you can accomplish by swapping tubes (with its 
tradeoff in overload performance), you can also accomplish much more 
easily simply by tweaking the IF gain control.  If you think you have a 
good reason for striking a different balance than the Collins engineers 
struck, then have at it.  I have played that game, and always come back 
to the factory setup.

Best regards,

Charles