[Hammarlund] Why 2nd RF stage?

[James A. (Andy) Moorer]

Oh boy - good question!

One of the biggest problems in high-sensitivity receivers is front-end 
nonlinearities (a.k.a. intermodulation). These produce images where 
strong stations may show up several times on the dial when you know they 
only transmit on one band. This is most common on higher frequencies. I 
have several receivers that exhibit this behavior - one of them is a 
very expensive and respected receiver as well. My usual test was to 
check how many times Radio Russia (RIP) showed up on the band. 3 is not 
an uncommon number.

There is one place where a nonlinearity is introduced deliberately in 
the circuit, and that is in the mixer stage. The idea is to deliberately 
form an image at 455 kHz (or whatever).

Fixed-frequency gain stages, as in the IF stages, are the easiest and 
cheapest to make, so designers often economize by having fewer tunable 
RF stages and more IF stages.

One could argue that tuned RF stages have the effect of lowering the 
overall IM because they provide additional suppression of strong nearby 
stations before you get to the nonlinearity (the mixer stage). Following 
this logic, the ideal would be to have all TRF stages, but that would be 
a devil to do mechanically, and it would be very difficult to make sure 
the tuning of each stage tracked the other stages precisely. Note that 
this assumes that the TRF stages themselves don't introduce much 
nonlinearity. This is probably true.

Thus, many (most?) of the "high-performance" radios have multiple TRF 
stages before the mixer. The Hallicrafters SX-88 takes the cake with 
three stages. Some folks consider it the best-performing receiver Halli 
ever made, and the prices on eBay, on the rare occasions they come up 
for sale, seem to reflect that belief (I've never had a chance to fool 
with a working one).

Note that in the move from the R-390 to the R-390A, the engineers 
eliminated the 2nd RF stage for cost reasons. This is my guess, but I 
would say this was justified because that they felt the factor of ten 
gain of the 1st RF stage was enough sideband suppression to get through 
the 1st mixer OK.

As Ulrich Rohde notes in his series of books on communications 
receivers, the first problem the engineer has to solve is what is the 
overall "floor plan" of the receiver - that is, how many of what kind of 
stage.

Almost all modern receivers use NO TRF stages. Instead, they have banks 
of fixed, passive (LC) bandpass filters going right into the mixer. One 
could argue that this places a lower limit on how good the IM could 
possibly be, since nobody can make enough front-end filters to really 
damp out competing stations. These filters are generally several MHz 
wide, and thus are filled with as many as, say, 200 stations. A receiver 
with even 1 TRF stage is theoretically capable of better IM performance 
than any modern receiver. Having said that, I'm not sure that the good 
ol' receivers with 1 or 2 TRF stages actually manage to achieve that 
kind of performance. Just a theoretical observation.

Good question.

-- 
James A. (Andy) Moorer
www.jamminpower.com