[HBR] Beam deflection mixers

[Ian Wilson]

Slightly OT for an HBR mailing list but I hope for forgiveness.

My personal holy grail is to come up with a hollow-state receiver
architecture that can
compete with modern solid-state HF receiver performance.

It appears to me (and I am no expert in these matters) that the most
difficult aspect
of all this is handling close-in interference. This is reflected in the IP3
figure - a
parameter that predicts the largest term in the Taylor series generated by a
mixer
when presented with a pair of signals differing by a small relative
frequency. This is
some sort of measure of linearity (if everything was linear, a close-in
signal would
not interfere at all).

Beam deflection tubes seem to me to offer a lot here. They combine a pentode
with
a switch. A pentode is more linear than a BJT (I think). And the
beam-deflection
mechanism seems to be a mostly linear process.

The Squires-Sanders receiver used a very familiar architecture: no RF
amplifier,
strong mixer, crystal filter following the mixer. Unfortunately it was never
developed
after its original introduction.

There are two ways to use beam-deflection mixers:
  a) single-ended signal to pentode grid, double-ended LO to deflectors
  b) double-ended signal to deflectors, LO to pentode grid

I have seen it stated that (a) is more linear than (b). This is opposite to
what I would
have expected. However, if it's true, that's good news - it's easier to
construct
balanced oscillator drive to the deflection plates than to bandswitch a
balanced input
signal.

Under scheme (a), I assume that the beam deflection tubes still offer
advantages
over pentode mixers. You can set up the operating conditions of the pentode
to handle
the signal levels you expect (and reduce large signals with and attenuator
upstream
if necessary). The mixing mechanism then reduces to a switch, much like how
diode
balanced mixers operate in the solid state domain.

OK - the point behind this long rambling point is to ask whether any
measurements
are available that would lead one to believe that a high-performance front
end is a
possibility? I do not have access to the kind of test equipment that would
enable me
to determine this for myself.

I have read in various places that solid state devices are intrinsically
better than
hollow state - but this is never backed up with anything concrete. I suspect
that there
are areas where the large-voltage handling capability of tubes, plus a
transfer
characteristic that has smaller high-order terms, may well give tubes a
significant
performance advantage over solid state devices.

73, ian K3IMW