[HBR] Progess report

[Walter A. Hutchens]

The little 7-tube (6 + rectifier) receiver which I am calling the R8040A 
since it is basically a major makeover of one I build in 1993 is pretty much 
done; the diagram needs only the BFO and how hard can that be?

I do need to study the software around here to see what I've got to do a 
computer drawn diagram.   If anyone has a cheap or free suggestion, I'd be 
glad to hear it.

I measured drift: starting a minute after turn-on it went down about 100 
cps then back up about 70 at about half an hour, then very slow wandering.  
This is excellent for such a casually laid out oscillator.   But it's very high-C 
and tunes 1350-1950 kcs so the tube is all but out of the circuit.   A bit 
more tinkering to get the heat to the compensating capacitor just a bit more 
quickly will finish that up.   

Intermodulation distortion measurements are in order and I should check 
the sensitivity.   The latter is good enough though -- you can copy stronger 
signals with no antenna at all in the evening.   The S-meter doesn't move and 
you have to turn the audio way up, but the 4" or so of wiring to the front 
panel terminal for a short antenna is enough.  

The basic plan for a general coverage receiver using the R8040A design as 
an IF seems to be complete.   The plan is to cover 1.75-31.75 Mcs in 250kcs 
bands.  250 kcs wide bands (rather than 500 kcs or 1 Mcs) make 
management of spurs easier as well as allowing a lower tuning rate without 
using the most expensive gearing.  

They key ideas of the G2DAF Mk 2 design are (1) use of push-pull front end 
to minimize intermodulation distortion, and (2) a below-unity gain RF stage, 
included only to allow AGC ahead of the first mixer.   The push-pull front 
end is incorporated in the GC-HBR; with AGC on the mixer stages (it worked 
fine in the R8040A) an RF stage isn't essential.

The fixed (2nd) IF will be 4.096 Mcs (instead of 5.4, I have some surplus 
4.3 Mcs IFTs that will work), the LO will tune about 1.454-1.704 mcs, 
making the variable (1st) IF 5.550-5.800 Mcs.   I have a four section AM-
FM BC tuning cap; two sections are 32-370 mmf which will be used to tune 
the push-pull oscillator; the other two sections are ~30-50 mmf and I'll use 
them to tune the variable IF.   A big toroid will have a Q of > 200, meaning 
quite a bit of rejection of unwanted signals into the 2nd mixer.   

Up to that point the circuits are identical to the R8040A.

The first mixer will be a 6922 -- a dual triode very similar to the 12AT7 but 
with a Gm of 12,500 which is about 3x higher.   Like the R8040A the grid 
circuit will have a single tuned circuit feeding the grids in push-pull.   A 
double section BC cap will cover the range ~1.75 Mcs - 31.75 Mcs in three 
ranges, something like 1.75-6-18-32.

With only one tuned circuit ahead of the first mixer, there will be some 
break through at the first IF but this is entirely tolerable in the R8040A -- 
one can either use an antenna that doesn't have high gain on the unwanted 
frequency or use a trap in the antenna lead.   

The first conversion (to the 5.550-5.800 Mcs IF) will be crystal controlled.  
For a ham band only receiver you'd switch in a different crystal for each 
ham band.   What I plan is to synthesize a crystal controlled frequency from 
7.550 to 37.300 in 250 kcs steps by multiplying a selected crystal from the 
120 crystal set for the BC-1335 (5.675-8.650 Mcs in 25 kcs steps) by 10 
and subtracting a fixed frequency of 49.2 Mcs.   Overtone crystal for that 
one -- no need for a whole 'nuther chain of harmonics.

Compared to the usual overtone oscillator circuit this very simpleminded 
synthesizer adds considerably to the complexity (two more tubes) but it's 
worth it to be able to get continuous coverage with crystals that are on 
hand.   It's impossible to calculate all the spurious signals in a set of this 
kind but there aren't a heck of a lot and none of the obvious ones 
(harmonics of LO passing through an IF, fundamental crystal frequencies) 
are in a ham band.

The plan is to provide internal sockets for six crystals covering the five real 
ham bands 160-10 with two on 10.  That will be two ham bands per front end 
range, plus a socket on the panel for any desired other crystal.  

It will be necessary to gang tune three circuits in the synthesizer according 
to the selected crystal.   These will also have to be bandswitched.   

The layout is something of a mess because of shielding and connection 
requirements but I think the issues are manageable; the toughest looks like 
finding room for all the stuff that has to be hooked to the bandswitch.   A 
fallback position is to just provide the one crystal socket on the panel as 
that would eliminate the switching of crystals, mounting of the sockets in 
the scarcest part of the chassis real estate, and the need for an 
INTERNAL/EXTERNAL switch.   

I think there'll be room for the internal sockets.   That's a more polished 
scheme.

I have a Miller MD-7 dial which would work (and takes very little space), 
but the two-speed tuning is annoying and I'd rather have a more positive 
drive for this.   I also have an Eddystone 898 but 100:1 (=5 kcs/revolution) 
is a rather slow rate.  I want easy tuning of SSB signals, no backlash, etc.   
I've looked at assorted military junk, thought about string drives -- 50:1 
doesn't  seem wise with a string and the 1-1/2" drum I have on hand.    Still 
thinking about the dial -- probably the MD-7 will get the nod.

Does anyone have experience with string drives at ~50:1?   What about a 5:1 
ball drive and a 10:1 string drive?   Not too complicated and with a 1-1/2" 
drum the string drive shaft would be a bit over 1/8" -- not unreasonably 
small.

(I did consider 500 kcs bands but ruled them out because I wasn't able to 
get rid of the spurs without using a higher LO frequency.   You just can't 
get the stability at 8 or 9 Mcs that you can at 1.5 Mcs.)
 
There are always complications.   We'll see how it goes.

Walt 
KJ4KV