[HBR] Another Receiver Project -- HBR-4, Part 14

[waltah at earthlink.net]

Adjusting the premixer DC balance greatly reduced the strength of 
the spurs, though it didn't entirely get rid of them.   They're 
tolerable now -- either in places that don't matter much to me -- 
3600 kcs, 21,200 kcs, or very weak.   Doing the design from 
scratch I'd move the crystal frequencies and VFO range down by 
125 kcs which would make things even better but its not worth it for 
this one.

The only things I haven't tried are selecting the best VFO tube (for 
balance to give the lowest odd-harmonic output) and adding a high 
pass filter between the premixer and the the mixer.   If built with a 
suitably moderate Q and tuned to about 17 Mcs (between the 
injection frequencies for 160 and 80 Meters) that ought to clean up 
the 9 Mcs and 5 Mcs leakage and eliminate the spurs almost 
completely.  However there'll be some loss of signal -- we'll have to 
see.

Doing it again I might add two more switch sections and another 
tuned circuit between the premixer and mixer.   That would 
definitely clean things up and with a little overcoupling the drive 
wouldn't fall off toward the band edges.   (One or two S-unit loss of 
sensitivity, depending on the band.)

The crystal oscillator problems of low output, critical tuning, and 
not working at all on 15 and 10M were solved by:

1.  Optimizing the plate coil for the bands 80 through 20M and then 
switching another identical coil in parallel on the higher bands to 
halve the inductance, and,

2.  Replacing the 12AT7 Butler oscillator tube (gm ~5000) with a 
6922, gm ~12,500.   The 6922 is identical in basing and 
characteristics with the 6DQ8 but has a 300 mA filament instead of 
365.  Basing of these tubes is the same as the 12AT7 except for 
filament arrangements so the swap was almost trivial.

There's still some tinkering to do: The crystal oscillator drifts on 
some bands and it's *not* due to crystal heating because it 
continues when you switch away from the band.   That puts the 
blame on the tuning of the tank circuit, which drifts some as the 
coil warms up.   The solutions are (a) moving the coil so it will stay 
cooler, and (b) lowering the coil Q by loading it with a resistor on 
the bands that have the problem.

When the crystal is stablized I need to revisit the VFO temperature 
compensation.   Drift is just a little high since I changed the 
feedback ratio a couple of chapters ago.

Other things -- Adding a locked multivibrator to give 25 kcs 
calibrator steps.   I kind of hate to mess with the calibrator 
because the circuit I have now is the best ever -- very level markers 
at about S-9 all the way to 10M -- but I really want the smaller 
calibration interval.

Tidying up a couple of places in the circuitry.  Detector output on 
AM wound up a little lower than on SSB.   

Trying filament chokes on the VFO.   This is a hot-cathode (for RF) 
design and the high harmonics have audible 60 cps FM; using 
chokes to float the filament might give a slight improvement in 
audio clarity.   Or maybe not.   

Adding more shielding in the IF stages; the IF passband shape is 
slightly different at very low signal strengths (little AGC action) than 
for strong signals, suggesting some regeneration -- though it could 
just be Miller effect.

The connection for the internal speaker needs to be installed.   This 
time I arranged things so the FT-101 internal speaker is a drop-in.

Paint and mark the front panel.  

Make a sheild plate to fit over the trimmers so the set isn't detune 
when installed in the FT-101 cabinet.   I may be able to trace some 
holes from the FT-101 shield but I moved most of the trimmers so 
those will have to be laid out.  

Buy a crystal for 160M -- 16 Mcs 3rd overtone -- and install that 
band.   Didn't want to get into that until I could see that the receiver 
was going to work but now that it is ...

Measure performance including 3rd order IMD dynamic range.   
Possibly another whole set of challenges there, although I don't 
know any reason it shouldn't be good, with no RF stage, a beam 
tube mixer, and single-conversion design.

The thing has all the front end selectivity you can tolerate.   On 
some bands the S-9 calibrator is inaudible if you're seriously 
mistuned and on all bands it's necessary to peak the front end 
when tuning from end to end -- even 28.0 to 28.5.  

Fortunately I carried over the FT-101 preselector dial and drive so it 
can be marked for the various bands.   

The IF breakthrough performance is probably marginal but the only 
thing I've heard when listening on 9 Mcs is a weak RTTY signal so 
I'm not going to add a trap at this time.   The calibrator 90th 
harmonic is inaudible except when the preselector is tuned to the 
bottom of the 30 meter band.

Another month of work.  I think I already said that.   

Did anyone see Ray Osterwald's (N0DMS) very advanced receiver 
design in the September Electric Radio?    What a piece of work!

Walt 
KJ4KV