[HBR] HBR form factor - "mods"

[Walt Hutchens]

Ian posted:

> Here's a subset of my questions
> (the context is 3-30MHz tube-only receivers, primarily CW/SSB):
>   - best AGC?
>   - best front end at handling large adjacent interferers?
>   - best product detector?

What is the:

- best car?
- best garage?
- best XYL/YL (or OM ...)?

The answers to both your questions and my analogies:

- depend on the meaning of 'best,' which varies from one ham to
another, and,

- are somewhat interconnected so that a full consideration isn't
practical. Having thus freed myself of the need for a really good
answer ...


AGC

The simplest AGC circuit -- 'simplest' is certainly one meaning of
'best' -- is a diode detector driven from the final IF transformer.
The AM detector can be combined, with perhaps a separate detector for
CW/SSB.

The limitations of this circuit, however, are substantial. It requires
high voltages at the IF output -- say 20 volts! -- to get full AGC
action on the strongest signals and it's tough to get distortion-free
IF performance at such high levels. (Use of the 6EH7 is highly
recommended.) The amount of current that can be supplied by a diode
driven from an IF transformer is small, so you can't get very fast AGC
attack to avoid overload at the beginning of a spoken sentence and
control volume of static crashes. The simplest circuit can't have as
high a ratio of decay time to attack time as one would like. And as
one of the other answers pointed out, low audio frequency feedback on
the AGC line will cause intermod distortion in the IF stages.

All these problems can be solved with more elaborate circuits but the
better job you do, the more extra parts you'll need.  One favorite AGC
circuit already mentioned is a solid state design comparable in
complexity to an entire W6TC receiver.

The first requirement for faster attack is a low impedance source for
the AGC voltage. Using a separate diode driven from the plate of the
last IF stage helps but a better answer is a dedicated AGC IF amp
stage and detector. These can be combined in the form of a triode
plate detector with the plate circuit DC-coupled to the AGC line. That
requires negative supply voltage and gets you the stability issues of
a DC amplifier (you must make setup adjustments for changes of tube or
line voltage ...) but is capable of excellent performance and doesn't
add too many parts. I've used this circuit in several receivers.

A separate tuned IF amp driving a diode solves the stability issues
but is more complicated still. A bit of non-linearity here is of no
concern. Still more parts are needed to give the best ratio of
decay/attack times.

Using a separate AGC amp lowers the IF output voltages: Because diodes
become non-linear at low signal voltages that means either adding
another stage of amplification ahead of the audio detector (thus
getting back to concerns about overload) OR using something other than
a diode detector. An infinite impedance detector is good with the low
output made up with extra audio amplification. Or a plate detector can
be used although there may be linearity issues with high signal
levels.


PRODUCT DETECTOR

Any first mixer circuit that will handle large signals well will be
satisfactory. Beam deflection tubes are good, ditto either push-push
or push-pull dual triode circuits.

(Push-push dual triodes connect cathodes in parallel without a bypass,
plates are parallel and filtered to limit output to the audio range.
IF signal is applied to one grid, BFO voltage to the other. Push-pull
dual triodes have cathodes driven push-pull by BFO voltage, grids
driven push-pull by the IF, plates in parallel deliver audio.)

Or the various gated beam tube (6BN6 ...) product detector circuits
can be used.

Use of a product detector forces the use of a separate detector for
AM. Balancing the gains along the various detection paths is an issue,
with some receivers winding up needing four tubes to do the AGC,
SSB/CW, and AM detection jobs.

An additional complication is keeping the BFO from showing up in the
AGC voltage. In the very simplest receivers the problem is solved by
disabling the AGC when receiving CW or SSB. More capable sets need
either very careful filtering/shielding or a neutralization circuit of
some kind.  Dual conversion makes the problem simpler.


FRONT END FOR LARGE SIGNALS

For simple receivers you can't beat a beam tube or dual triode 1st
mixer running without an RF stage. The dual triode is a bit simpler
but for 10 meters the beam tube probably gets the nod for sensitivity.

With the dual triode mixer, a very low gain RF stage not only gives
ample sensitivity but allows isolation of two input tuned circuits
which helps with strong-signal overload problems.  High Q in these
tuned circuits is of critical importance.

The 12AT7 and 6J6 are the best dual triode first mixers. Except for
frequency differences, the dual triode circuits are the same as those
described as product detectors above.

(An 80/40 band imaging 'first receiver' of outstanding performance yet
fairly easy to build uses a push-pull 1st mixer with a single toroid
coil tuned by a dual section TRF-type variable. A separately tuned
push-pull LO drives the cathodes of this mixer. Follow with two IF
stages and a BFO/product detector, use a DC-coupled plate detector for
AGC. AGC voltage can be applied to the mixer grids without perceptible
oscillator pulling. I've described this receiver in more detail,
before.)

We have here departed from the strict W6TC design. For better
stability TC's receivers run the LO on half the injection frequency on
20M and up. To make this work without needing a separate tuned circuit
for the actual injection frequency the oscillator is coupled to the
mixer signal grid so the mixer grid coil selects the harmonic. Using a
mixer with separate LO injection (either of the circuits recommended
above) will require operating the LO at the higher frequency, meaning
either greater attention to stability or more drift.

As usual, simplicity makes its own demands.

Walt
KJ4KV