[HBR] GC-HBR Project -- Was 'Navy guy ...'

[Walt Hutchens]

I said:
> The plan is to take the 120 crystals of the BC-1335 WW-II portable
> FM set, 5675-8650 kcs (25 kcs spacing) and multiply the
> frequencies x10 to give 120 frequencies 56.75-86.50 Mcs.  
> Subtract 49.2 Mcs yielding 7.55-37.30 Mcs by 250 kcs steps.  
> Adding a tunable oscillator 1.704-1.454 Mcs to a fixed IF of 4.096
> Mcs (the ladder filter with 4.096 microprocessor clock crystals)
> gives a tunable IF of 5.800-5.550 Mcs.   The difference between
> that and the synthesized band crystal frequency is the tuning
> range for the band:  for the lowest band, it's 7.550 Mcs - 5.550-
> 5.800 Mcs = 2.0 - 1.75 Mcs. 

Today's quiz question is "how many times can the same person 
look at the same numbers and miss a problem?"  

Answer:  A bunch.   

The synthesizer starts with those fundamental crystals 5675-8650 
kcs; after multiplication and subtraction of 49.2 Mcs, the result is 
mixed with the signal from the antenna.   However, I overlooked the 
fact that the lower end of the range of crystal frequencies overlaps 
the tunable IF, 5550-5800 kcs.   So the five bands using crystals 
below 5800 kcs (1.75-3.00 Mcs) would all contain a birdie from the 
frequency of the crystal for that band.   

There will be two or three three tuned circuits to attenuate these 
signals, but they'll be audible.  The band 3.000-3.250 would have a 
band-edge marker at the low end, where the 5800 kcs crystal for 
this band is picked up by the high end of the IF tuning range.  

If you want to look on the positive side, these frequencies supply 
some calibration points for the LO.   Not enough to do the whole 
job though -- they're all at the low end of the band.  So I probably 
can't get away with claiming this was just a clever piece of design 
work.

You could move the band crystal range up 150 kcs to start at 5825 
kcs -- or just skip the lowest frequency crystals in the set and 
change the 49.2 Mcs subtraction frequency to start the coverage 
where you like.  Neither idea was really terribly attractive.  The 
receiver is planned around a set of crystals that I already have. If 
starting with purchased ones, you would go with overtone units 
($20 x 120 = $2400 ... Ah don' think so ...) and avoid the whole 
issue, as well as getting rid of the need to hetrodyne to get the 1st 
mixer frequency.   And I'm too obsessive to really like throwing 
away part of the set.

I also considered moving the tunable IF by moving the LO down 
from 1.454-1.704 Mcs to 1.304-1.554 Mcs, giving an IF 5.400-5.650 
Mcs.   The fourth harmonic of the LO would then cover 5.216-6.216 
Mcs so it would sweep entirely through the tunable IF range, going 
in the same direction.   The LO and 2nd mixer (which gets the 
tunable IF as the input) are in the same box.   This would be a very 
strong spurious signal tuning slowly through the same spot on 
every band -- a definite loser.

Or you could keep the LO range the same but move the fixed IF 
down to somewhere in the upper end of the 75 meter band.  You'd 
have the BFO in a ham band and the least bit of feedback would 
make the IF passband shape go west when tuned near that 
frequency ... The idea was dead before I even looked at available 
clock crystal frequencies.

Obsessive or not, skipping the first six crystals in the set seems 
the best approach.   This means starting with the 5825 kcs 
(channel 276) crystal and changing the subtraction frequency to 
keep the lowest synthesizer output frequency at 7550 kcs so the 
tunable IF remains 5550-5800, This requires an overtone crystal at 
50.7 Mcs, but I have that crystal -- it was used in the PRC-6 handy-
talky.   You lose the coverage above 30.50 Mcs, but that's not a 
big deal.   

The harmonics of the LO (1.454-1.704) also must be considered.  
None of them are in the fixed IF, tunable IF, or the lower ham 
bands but the 9th harmonic goes through the 20M band. It's going 
in the reverse direction (higher LO frequencies -> lower dial 
frequencies) so it'll be going fast.   

Some other (non-ham) bands will be hit by LO harmonics.   Since 
the LO comes after the front end, these harmonics should be close 
to inaudible if the LO/2nd mixer box is closed and well-filtered.   
Also the layout there (and tuning cap design) makes it easy to 
return all the grounds to the tuning cap frame, thus minimizing 
chassis currents.  

With 30 Mcs of total coverage it might seem logical to start lower 
than 1.75 Mcs -- say down at 500 Kcs.   However doing so would 
require tuning the front end range in four rather than three bands 
because it's tough to get more than a 3:1 tuning range with 
conventional tuning caps.  In addition, extending the tuning range 
down would make it overlap the LO range.  You'd pick up another 
(probable) birdie per band from the LO and there could be trouble 
with more complicated spurious responses caused by strong BC 
signals near the LO frequency.  

It could be done on a 'take what you get' basis, just by using the 
six omitted crystals and adding the fourth range to the bandswitch. 
Given the strength of typical BC band signals, that probably would 
be useful.  However the goal was to keep this project fairly easy to 
build.  Adding BC band coverage probably exceeds my level of 
ambition.   

Commercial grade general coverage receivers beat these sorts of 
problems with complexity -- adding or taking away a conversion on 
some bands, for example -- or by making other types of 
compromises such as using single conversion with phase-locked 
oscillator.   In a practical homebrew project, I think it's better to 
accept that there will be some garbage and try to leave it in places 
you won't often visit.  

It's still a paper receiver (plus what I hope is now a complete 
bucket of parts) but I think I'm closing in on the design.   

Walt 
KJ4KV