[HBR] Another Receiver Project -- HBR-4, Part 19

[waltah at earthlink.net]

'More progress and another realization.'

I had planned a parallel trap in series with the antenna lead placed 
between the BNC connector and the link on the coil.  Unfortunately 
in that position the reactance changes so wildly, both due to 
reflection from the antenna coil and the antenna itself, that there 
was no one correct setting of the trap -- it depended on the band 
and the exact tuning of the antena coil.   Using a series trap 
across the antenna connection (with much higher reactances) 
make things much better -- good enough to be satisfactory.

I put in a cathode-coupled 6ES8 RF stage.   The antenna coil feeds 
a cathode follower which drives the second half as a conventional 
triode RF amp.   After fixing a problem caused by inadequate 
bypassing of the AGC line to the second triode (the c.f. is not on 
AGC) it is perfectly stable on 80 through 20M;  later today I'll tune 
up the other bands and see if there are problems there.

The gain seems ample; if I now reduce the IF gain, the mixer noise 
should disappear.  Have to add a pot to adjust the gain of the 
second IF stage from the back panel so that S-9 = 100 uV. 

It was nice this morning to be able to hear a VK on 75 and realize 
that our bands were back in operation after a week of absence due 
to a solar flare.

In addition to adjusting overall gain and checking stability, I need to 
check the sensitivity (which should be improved) and the dynamic 
range (hopefully unhurt and with luck I might now have enough 
signal generator power to measure a higher number than before.)

Then it will be time to plow ahead with returning the premixer to the 
two-6J6 configuration.   With the higher sensitivity, the spurious 
response caused by direct conversion by the VFO (without the 
intervention of the crystal oscillator) is now easily heard.   The 
former configuration -- a doubly balanced premixer with four triodes -
- killed that, and I need to go back there again.   Another couple of 
days work, I expect -- have to make a plate to mount the two 
sockets, completely rewire, replace the trimmer board, wind a new 
coil ... like, *total* makeover.

At the same time I want to increase the output of the premixer.   
The mixer should be operating in switching mode and something 
like 10V p-p is required to do that.

This premixer layout will be a little different because driving G1 of 
the mixer requires only single-ended output.   So that puts all four 
premixer triode plates in parallel -- relatively a lot of capacity for a 
circuit that goes up to 39 Mcs.  The good news is that the 6J6 
plate capacitance (to all other electrodes) is only 0.4 mmf/section if 
you don't use a shield and only around 1.2 if you shield it.   That 
won't be a problem.

Another possibility is a single twin triode with independent 
cathodes (say a 12AT7), driving the grids from the VFO and the 
cathodes from a center tapped winding on the crystal oscillator 
plate coil.  However it's harder to drive cathodes than grids and 
there may be a problem getting enough power.   Probably I should 
start with a 6922 premixer -- that's like a 6DJ8 but with a lower 
current filament.

The premixer doesn't really have to be balanced against the crystal 
oscillator.   That frequency is 14 Mcs above the channel so not 
likely to cause problems.  But because 'a mixer is a multiplier,' 
adding signals in the plate circuit for single ended output requires 
that both input signals be supplied in push-pull.   

And the realization:  with the shift from G1 to the deflection plates 
for the antenna input to the 6JH8 mixer, I have the choice of push-
pull signal drive.   Meaning that a push-pull RF stage could be 
used, as in the G2DAF design.  *That* ought to be a real killer 
design, since push-pull stages tend to cancel out even harmonic 
distortion and the second harmonic plays a major role in third-order 
intermodulation distortion.  

It can't be done in this set without major changes.   A push-pull RF 
stage needs balanced tank circuits and having used the FT-101 
slug tuning front end parts makes that impossible.   But a similar 
overall configuration using a four-gang tuning cap and fixed coils for 
the front end (as in the G2DAF) would do it and would fit just fine in 
the same basic layout.  

There's miles to go before that becomes an issue.  I don't even 
know for sure that the present single-ended RF amp will be the 
limitation of this design, though with a mixer capable of handling 
~40-volt swing, the limit certainly should be the RF amp.   

Walt
KJ4KV