[HBR] Re: Another receiver project

[waltah at earthlink.net]

Bill Higgs wrote:
> I was noting with interest the HBR-4, and also noting with interest the
> late W1KLK's design in Feb '72 QST (an experimental receiver for 75-meter
> DX work, p. 41). Doug used a low-gain RF stage with a grounded grid dual
> triode . . . the RF stage used dual tunable Cohn filters, so the point was
> only to make up the loss in the filters, about 5-6 dB. He used a 7044
> "computer" tube; your beloved 6J6 might well work here also. Interesting;
> should be stable in grounded grid mode. Mixer is the venerable 7360,
> driven unbalanced by a series-tuned 6U8A VFO/buffer.

Single band receivers are a whole 'nuther ballgame.   The '75-meter 
DX work' receiver looks excessively complicated for the job.   I 
really think that sensitivity better than about 1 uV is wasted on 
75/80 meters -- and that was *before* computers brought us whole 
new spectrums of RF 'dirt.'   

For a single band receiver you don't have much of a problem with 
high drive levels -- just run a push-pull oscillator using a 6J6 or 
perhaps that 7044; there should be enough gain even with no RF 
stage and even with a bit of loss in the input circuits.   I probably 
need to read the article to know why he used a bandpass input 
circuit.   Yes, it eliminates tuning but by his judgement it forced 
another stage.

Geez ... it's not a problem to *gang* the tuning for a single band 
receiver and doing so makes the absolute best use of a wide-
dynamic range mixer.
> 
> The design of the BFO is of interest, as it looks as though it might be
> modified a bit for push-pull output by using the buffer (tetrode) stage of
> the 6U8A as a phase splitter, replacing the RFC with a resistor and taking
> outputs from the plate and cathode. Probably would reduce the output by
> half, but you'd get it back in absolute voltage difference at the
> deflection plates. 

What you get into is that tube capacitances effectively bypass the 
phase splitter outputs.   This isn't a problem at low enough 
frequencies but at a few Mcs it gets you, unless you use such low 
load resistances that output goes to nil.   (The plate load is in 
series with the tube's plate resistance.)  The frequencies for which 
push-pull output might be needed (and not easily available just by 
using a push-pull oscillator circuit) are 16-44 Mcs in this project ... 
that's out of the range for phase splitters.

> Dunno if you've biased the deflection plates in your mixer or not
> (the 7360 requires 25V or so bias, the 6JH8 allegedly does not) . .
> . I don't know if biasing the plates positive affects the overall
> conversion gain of these tubes. 

The data sheets for the 6JH8 claim that maximum deflection 
sensitivity occurs at -14 volts on the deflection plates but that's at 
250 plate volts.   I'm running mine at 130 plate volts and -7 volts on 
the deflectors.   I don't know if that's optimum or not, but it works 
okay.   

Conversion gain is proportional to the amount of deflection, so the 
slight negative bias is good for gain, as well.   And an added plus -- 
with some negative bias, the tube draws negligible deflector current 
so the deflection *power* and loading of the driving tuned circuit 
should go down considerably.

> My NEXT project after the HBR-??(WAY ahead of myself here) will be
> a project using the 6JH8 as a mixer and a pair of 6EH7 frame grid
> tubes in the IF; probably for 20 meters with another 6EH7 as an RF
> stage. 

That sounds like a good plan.   I don't know what would be the best 
possible 6JH8 noise figure on 20M but considering the greater 
difficulty of getting good tuned circuits on this band and above, an 
RF stage sounds wise.   

I don't understand all I know about noise and the beam deflection 
mixer, but the input noise of the 6JH8 in this project is *inaudible*.  
That is, with the antenna disconnected there is nothing you can 
peak by tuning the grid through resonance.   The tube definitely 
generates some noise (in fact nearly all of it, in this application) but 
it's just the electron stream.    And the noise doesn't vary with the 
drive level -- I can pull the premixer and the noise remains the 
same.   So increasing the drive level just raises the output without 
a corresponding increase of noise.   

I'm using the 6JH8 with no RF stage and *3* 6EH7 IF's in the HBR-
4 and that seems the right amount of gain, with the higher drive 
levels I'm now using.
 
> Doug's IF is 455Kc with a Collins filter, followed by a Q Multiplier stage
> set up for notch mode. 

455 kcs IF's and mechanical filters both seem like anachronisms 
now.   (*So* 1950's ...)  You can't possibly get adequate image 
rejection at that frequency so you're forced to a multiple conversion 
design.   And crystal filters for the low HF are both easy to design 
and dirt cheap, using microprocessor clock crystals.   

One can either go with an existing conversion scheme, trusting the 
designer thereof to keep you out of trouble, or take paper and 
pencil and strike out on your own.   Right now, I'm wishing I hadn't 
had 9 Mcs filters on hand, because I'm certain I could have gotten 
that 3.6 Mcs spur out of the band with just a slight jiggering of the 
IF and VFO frequencies.    Probably I can get the amplitude down 
considerably with some adjustment of operating conditions, but it's 
hundreds of uV -- it's not going completely away.

Humm ... maybe that's what I'll do eventually -- make my own 
filters in the same size packages but for a slightly different 
frequency.   

Actually, you can do this kind of adjustment by varying any two of 
the IF, the VFO range, and the band crystal frequencies.  

> Work on the HBR-?? is going to be suspended for several weeks, pending the
> relocation of my shop. The termites finally won.

They do that ...

Walt
KJ4KV