[HBR] That General Coverage HBR Project -- 2

[Walt Hutchens]

Bill wondered:
> You mention a 'simple minded synthesizer' that sounds like a
> "crystalpexer" scheme.

Welcome to the list, Bill!

"Crystalplexer" is a term I haven't heard but the scheme in the GC-HBR
is to use the tenth harmonic of the 120 crystals of the WW-II surplus
BC-1335 FM set.  They run 5675-8650 kcs I think, by 25 kcs steps so that
gives you 56.75 -86.5 Mcs.  Subtract from that 49.2 Mcs yielding
7.55-37.3 Mcs.  Use that as the crystal mixing frequency to get to a 1st
(tunable) IF of 5.550-5.80 Mcs with coverage of 0.75-30.5 Mcs.  (Front
end probably won't tune the bottom end -- I don't need the BC band and 
the first few crystals show up in the tunable IF so they'll give large 
birdies.)  The 2nd (fixed) IF is 4.096 Mcs (clock crystals are 
available) and LO tunes 1.454-1.704 Mcs.

That's all from memory so if it doesn't quite add up ...

The modern cheap clock crystals are amazing.  Out of 20, I got two 
groups of five matched within a total of 7 CPS each -- that's perfect 
matching for an SSB filter at 4096 Kcs.  The BFO will be a VXO and must 
tune completely across the filter range; that's what I'm working on 
right now.  Fallback is two BFO crystals but I don't want to have to 
switch them.

Anyway there's one crystal multiplier and one mixing to get the 1st 
mixer injection frequency = 'simpleminded.'

> You referenced R8040A. Where can more information and maybe some
> block and schematic diagrams. Searches for R4080 have turned up
> empty.

R8040 was a band imaging design 80 & 40 meters, five tubes, series 
filaments, that I did in the mid 90's -- it worked pretty well and was 
documented in an article in Electric Radio at the time.  A search of 
their on-line index should locate it.

Last year I revisited the design, changing just about everything. 
R8040A (which I am listening to as I write) turned out to be an 
excellent performer for 7 tubes+rectifier.  And easier to build than the 
forerunner.   It's not written up anywhere at this point, though I have 
most of the circuit diagram and -- since most is being recycled in the 
current project -- I'll obviously be finishing that up.

12AT7 push-pull mixer.  This has a single toroid push-pull tuned circuit 
in the grid and is on the AGC buss.  There is slight leakage of 40M 
signals onto 80M when using the set with a large antenna that's resonant 
on both bands -- I use a 160M FW loop for everything but it hasn't ever 
been a problem.  An antenna tuner or separate antenna for each band 
would eliminate this if necessary.

A push-pull triode mixer with no RF stage is 'strong' enough that a 
single tuned circuit of high-Q preceding it is adequate.

12AU7 push-pull LO.  Decoupling is enough that there's no pulling at 
all, even with the AGC controlled mixer.  IF is 5400 kcs (I think) and 
LO tunes 1400-1900, giving coverage 3500-4000 kcs and 6800-7300 kcs.

There's a two-crystal half-lattice filter in the plate of the mixer, two 
stages of 7GH8 IF.  The triode in the 1st IF is used as a diode 
rectifier for -160 VDC bias line needed for the AGC; the triode in the 
2nd IF is an AGC plate detector (cathode biased negative so the plate 
runs negative from 0 VDC for AGC).  No problems with attack speed -- 
clicks at the start of a sentence or such -- and it works the same on 
SSB and AM.

Half of a 12AU7 is used as a plate detector on AM, with BFO injection it 
becomes an adequate product detector.  The other half of the 12AU7 is 
the 2nd audio -- it's enough power for a 3" speaker.   The pentode half 
of a third 7GH8 is the 1st audio stage, the triode is the BFO, crystal 
controlled on the high side of the IF.

The conversion scheme renders LSB as LSB in the IF for both bands so 
there's no need to tune the BFO.  The only control is ON-OFF.

A 35W4 delivers plate voltage.  You could use a diode rectifier and a 
35C5 audio output stage, but there's really plenty of audio for anything 
you'd want to use a relatively simple receiver for.  The S-meter runs 
from the 1st IF cathode/screen using the common bridge circuit.  7 tubes 
total.

Every receiver is a compromise.  This one is compact, stable, has 
excellent selectivity, doesn't use much power (little warm-up drift) has 
great AGC (3 controlled stages), few controls, has very good overall 
audio quality, and isn't too hard to build.  The negatives are 
non-excellent audio because a triode isn't a perfect product detector, 
the 40M breakthrough, AM signals are a bit distorted unless received 
with the BFO on because the crystal filter is optimized for SSB and 
isn't quite wide enough to pass both the carrier and sideband.

For the GC-HBR I'll use the R8040A design as a tunable IF with the 
obvious tweaking of the tuning range.  Other changes are a 12AU7 with 
the sections in push-pull for audio output, a 6BN6 product detector, and 
  the VX0 BFO as mentioned.  Most of the tube types have to change to 
get a 300 mA filament string instead of 150 mA.  It's conceptually not 
much more complicated but there are quite a few pieces to make.

Today I hope to put together the ladder filter, then I can go back and 
finish the VXO with exact knowledge of the tuning range I need.

Loosely a VXO will tune from the series resonance of the crystal up to 
the parallel resonance but this space can vary considerably -- from say 
1 kc to 5-6 kcs at 4 Mcs, depending on the crystal parameters.  And it's 
common to put an inductor in series with the crystal in order to lower 
the series resonant frequency.  If the inductor is too big you can lose 
stability or even oscillation.  The usual VXO tuning cap has a 
substantial minimum so you don't get too near the parallel resonant 
frequency either.  You could put an inductance across it but man, this 
thing is starting to get complicated!  The oscillator output varies 
considerably over the tuning range -- that's a problem for the product 
detector ... and so on.

Walt Hutchens
KJ4KV