[HBR] Another Receiver Project -- HBR-4, Part 7

[waltah at earthlink.net]

I raised the IF gain until the mixer noise (no RF stage) is plenty 
loud.  IF noise (with mixer pulled) is almost inaudible even at full 
volume -- all that seems right.

I believe all the pieces work as they can be expected to, but sure 
enough the system doesn't work.  Namely, the receiver's sensitivity 
is far too low, apparently because of inadequate drive to the 6JH8 
mixer.    There are two answers, either of which should work, but 
first, the details of the problem.

The VFO (5-5.5 Mcs) output is about 1 volt p-p.   The Butler crystal 
oscillator (crystals 14 Mcs. above upper end of each band) has an 
output modestly greater than that.   These signals are mixed by a 
pair of 6J6's each operating as a push-push mixer with opposite 
phase outputs coupled to the deflection plates of the 6JH8 mixer.   
The deflection voltage (premixer output) is about 1 volt p-p from 
plate to plate.   Squires (of Squires-Saunders SS-1R fame) said in 
a QST article on 7360's as receiver mixers that about 10 times that 
value is optimum.  

The SS-1R is the only commercial receiver I've found using a beam 
deflection mixer.   It's double conversion with each mixer driven 
directly from an oscillator -- a much easier way to develop the 
necessary voltage.   Like my design, there's no RF stage -- 
desireable for best strong-signal performance.

The beam-mixer ham receivers that I know of either use a 
preceding RF stage or are double conversion above 80 meters with 
another mixer and RF stage ahead of the beam tube.   The only 
premixed (single conversion) design is W1OMX's, which has a 
pentode RF stage.   There's a no-RF stage beam tube single 
conversion design in a hybrid receiver in the 5th edition RSGB 
handbook but it drives the mixer directly from a PLL'ed VFO.   

In other words, I'm not the first to get here.

The 2 x 6J6 mixer is doing about what it can do -- roughly zero 
conversion gain.   The crystal oscillator can't be flogged any harder 
as the HC-49 crystals have a tiny bit of drift -- ~10 cps -- when you 
switch bands, as it is.   I might get somewhat more from the VFO, 
but the price would be reduced stability: this is the best VFO I've 
built and I want to keep it that way.  And in any case, driving the 
pants off of a mixer of this type to get more output voltage will lead 
to distortion and more spurs.

One solution is to add an RF stage.   I left room to do that; it 
wouldn't be a big deal.   The stage wouldn't have to have much gain 
and could be optimized for large signal handling.   But every db of 
gain there pushes you closer to the limit of the 6JH8's signal 
handling capabilities -- that limit is independent of the deflection 
voltage, so running the tube with low deflection voltage doesn't  
help.

The other approach is to use a much higher gain premixer.   There 
are various ways to do that but the tidiest by far is to replace the 
two 6J6's with a single 6JH8.   These tubes were used in an 
occasional ham SSB transmitter and are able to develop enough 
output from oscillator-size signals to drive tubes delivering 10-20 
watts; they're not your dad's 6K8.  Input and output arrangements 
are the same as the pair of 6J6's, namely one single ended and 
one push-pull input with push-pull output.

That's the way I'll try; the only small problem is cutting out and 
replacing that chunk of chassis with a plate having a single 9-pin 
socket.   An hour's work, plus the (modest) wiring changes.

I did add the DC balancing voltage to the 6JH8 mixer deflection 
plates.   The RTTY signal at 9000 kcs is now gone.  A trap may 
still be needed but that's for later.

We'll see ...

Walt
KJ4KV