[HBR] Another Receiver Project -- HBR-4, Part 22

[waltah at earthlink.net]

'Makin' progress.'

Sure enough back biasing the premixer cathode eliminated the 
wierd unstable spurs.   However there were still some S9+30 nice 
stable spurs, particularly the one at 3600 Kcs which is the 5th 
harmonic of the VFO at 5.4 = 27 Mcs less the crystal frequency of 
18 yielding 9 Mcs.

The first question was whether the VFO drive to the premixer was 
too high.  Cutting the VFO plate voltage in half cost only a few db 
of signal strength but reduced the 3600 kcs spur from S9+30 to 
about S5.   The next step took some poking around, but what I did 
eventually was run the 2" lead from the premixer plate to G1 of the 
6JH8 mixer using RG-174/U coax.   That got it down around S1.  
Then I tweaked the mixer balance pot and was able to make it 
nearly vanish -- you can hear it, but the S-meter doesn't move.

In hindsight this should have been much easier to see:  All the 
serious spurs are 9 Mcs signals coming out of the premixer.  
However going into the mixer G1 with output from that tube taken in 
in push pull from the plates, such signals should vanish.   If they 
don't, then either you have a shielding problem -- the signal is 
leaking more to one plate than the other -- or you have a balance 
problem.  I had both.

There's still a just-audible response to very strong signals on 80 
and 20 meters caused by direct conversion (no crystal frequency) 
using the 5-5.5 Mcs VFO a la 'band imaging.'    Jiggering the 
crystal and VFO frequencies would solve this (in a repeat project) 
by moving these responses far enough from the proper one for front 
end attenuation to take them out.   However, impurity of the LO 
signal is nasty, regardless.   Since the lowest frequency out of the 
premixer should be 14 Mcs above the band, I'm going to try a high-
pass filter at the premixer output.

I'll probably do some further investigation of the shielding issues in 
that area.   And then probably increase the VFO plate voltage a 
little to recover the lost gain.   The VFO power level is fine -- ~100 
mW plate input.

I haven't noticed any other spurs but I haven't looked carefully yet.

Meanwhile I noticed that the front end peaked up at quite different 
settings compared to the time before 'it thinks it's push-pull.'   A 
little study showed that I had incorrectly arranged the new winding 
on the RF to mixer transformer, putting it at the hot end of the 
existing winding rather than the cold one. Figuring that that had 
increased the stray capacitance, I did it again, (right this time) and 
adding one more turn.   The bad news is that even with things laid 
out correctly, the tuning is still quite different, making the front 
panel markings wrong and probably hurting the gain on 10 meters.  
However the good news is that I was able to get plenty of gain by 
taking output from the plates of the push pull RF stage directly to 
the deflection plates of the 6JH8. 

I can't prove that that makes anything better, but it's a lot more 
esthetically satisfying.   The receiver does sound teriffic in that 
configuration.

I'll have to further investigate capacitance issues in that stage.  The 
main problem is likely to be the combination of the 6ES8 (RF 
stage) plates and the 6JH8 (mixer) deflection plates, but it's too 
soon to say for sure. The wiring of the bandswitch can certainly be 
improved.   And a trick or two may be possible. 

The one remaining major problem area is the crystal oscillator.   It's 
unstable, hard to tune up ... I think that covers it.   Further work is 
needed there -- those crystals should be 'rock solid.'   I don't think 
it's an excessive drive situation; more likely just that the Q of the 
tank is too high.   But loading that is kind of a mess and I've been 
resistant to getting serious about it.   There's ample gain available --
I'm running that tube at about 60 volts and getting plenty of output.

And a lesser problem -- there's still too much 9 Mcs leakage from 
the antenna.  I have a trap there, but more is needed.   

The shape of the bandpass is not quite the same for strong signals 
as it is for weak ones, indicating some IF regeneration.  

Walt
KJ4KV