[HBR] Pullen mixer

[Walt Hutchens]

Donald said:

> RF stages can hide a bunch of mixer problems

Mixer characteristics -- which may or may not be problems in a particular
application -- include internal noise, linearity on the signal path, and
gain.   An RF stage can make up for low gain (though usually that is better
done in the IF section), and hide the mixer noise.

An RF stage generally has three functions:

1.  Raising the signal level above mixer noise.

2. Isolating two signal frequency tuned circuits, thus simplifying their
design; and,

3. Acting as a controlled attenuator for signals much stronger than needed
to overcome mixer noise that would otherwise drive the mixer into the
non-linear range.   Generally this attenuator function needs both automatic
control (AGC) and manual (AGC threshold or RF/IF gain) control to cope with
the various situations.

Examples of ways this can play out are:

1.  The many 1940's designs with extremely noisy pentagrid and hexode mixers
that were nonetheless tolerable sets because of a high gain RF stage;

2. The G2DAF receiver with a 6ES8 RF stage for isolation and AGC purposes
operating at less than unity gain ahead of a very quiet 6DJ8 push-pull
triode mixer; and,

3. The Squires-Saunders SS-1R set with a beam deflection mixer (very quiet
AND able to handle very large signals) and no RF stage.

When you break new ground -- I'd guess that a Pullen circuit with ~2V on the
plate of the mixer scores as 'new ground' -- you get to figure out how the
rules apply to your creation.   I gather that the noise is low and the gain
ample but I don't see how it could handle very large signals while remaining
linear.  Maybe NO RF stage?   Donno ...

You can do the standard interference free dynamic range (IFDR) test with two
signal generators 20 kcs apart to get a number for this.   Or just set it up
and listen to 80M SSB in the evening traffic peak: If everything sounds just
as good then as it does at 2 AM local time, you don't have a dynamic range
problem.

(IFDR test: What is the signal strength that produces a chosen S/N ratio --
say 10 dB at 3800 kcs?  What is the signal from two (equal output)
generators when they are tuned to 3820 and 3840 that produces the same S/N
ratio at 3800 kcs?   The ratio of the two signals is the IFDR.

(There is some fussiness about connecting the two signal generators so they
don't see each other and cross-modulate in their own circuits.   There is a
fairly simple circuit to prevent this; the details are in the more recent
Handbooks.)

IFDR testing generally puts greatest stress on the last stage before the
really sharp selectivity.   That's because once you go through the
'knothole' the signals that are 20 and 40 kcs away from the desired
frequency are so much attenuated that they can no longer cause
non-linearity.  

In a set like the HR-10, the stress would be on the mixer because the
crystal filter is in that plate circuit.   In a 'true' HBR it could be in
the first or second mixer or even in the first LF IF because 20/40 kcs isn't
that far off for these stages and there is gain at each stage, as well.

Poorly designed RF stages -- too much gain, particularly if the tube itself
doesn't handle large signals well -- can also be the limiting factor in
IFDR.  The antenna coil Q is often limited by the load of the antenna so it
gives little help.   Basically an RF stage should have just enough gain to
cover the mixer noise and the mixer should have just enough to cover the
noise of the 1st IF.   If the 'knothole' is in the mixer plate circuit
you're home free after that and as much additional gain as needed can be
supplied in the IFs, detector, and even audio stages.

One question would be how the mixer job was done in the auto radios that
were designed around the space charge tubes.  Auto radios MUST handle large
signals so that ability should be a given with that tube/circuit.   And most
auto radios are reasonably sensitive as well.

The reason you found the push-push triode circuit to be high noise is that
it's not a good choice for use without an RF stage.  It does, however, have
a high dynamic range, meaning that if you boost the signal above the tube
noise you still have room before you drive the tube into non-linearity.

Somewhere I was hanging out a decade or so ago somebody was joking that an
833A might be an ideal receiver RF stage or mixer.  It wouldn't, really, but
that guy did understand that the ability to handle some power is a
requirement for HF receiver front ends.

My nominee for the least-well exploited mixer circuit is the beam deflection
tubes -- 6AR8, 6JH8, 6ME8, and 7360.   While there are some design hurdles
to get over they're not beyond anyone who could build a W6TC set from
scratch and you will have a radio with very few peers -- it should equal or
even beat the R390s for IFDR.   Unfortunately no ham design that I know of
made good use of these tubes.

The Miser's Dream and related sets (the Advanced Six Tube Amateur Receiver
in the '69 handbook) are the only ones I know of and the designer of the
Advanced Six ... really didn't get it: He turned out a single band receiver
(80M) using a 7360 mixer with no RF stage -- okay so far -- but used a 455
kcs IF, presumably so that he could use a Collins mechanical filter for
selectivity.   Great on 80, but images will be a problem if the front end is
tuned for higher bands.  Well, he knew that, so he put a crystal controlled
converter in front of it.   Image problem solved but the 1st mixer is now in
the converter and is perfectly conventional in design.

It's not bad as such things go but a showcase for the possibilities with a
beam mixer it isn't.   The SS-1R design is the one to follow:  If stripped
of all the complexity it will do the job and the circuit is out there.

General information on the SS-1R is here:
http://www.w1vd.com/SS-1R.html

Look at those performance numbers!   Yeah, the 0.5 uV sensitivity isn't
remarkable but it's all you need on HF and the numbers you do need, beat
everything including the R-390.

(My hat is off to anyone with the tools to MEASURE an IFDR of 94 dB!)

The schematics can be found here:
http://bama.edebris.com/manuals/squires/ss1r

In the five files with 'sch' in the file name.

There's probably some sort of beam tube HBR in the KJ4KV future.   Maybe on
the second HR-10 chassis?   Maybe three years from now?

Walt 
KJ4KV