[HBR] Beam deflection mixers

[Walt Hutchens]

Peter commented:

> Those mixers are a good way to pick up extra rejection of IF signal blowby,
> especially for sets using 1st IF frequencies in the new expanded AM band.
> 
That's a good point.  To do this you apply the (antenna) signal to the
control grid, provide push-pull LO to the deflection plates, and take the IF
output from the two pentode plates in push-pull.

This means no cheating:  Simple designs often (RF) ground one of the two
pentode plates and take the IF output from the other, single ended.   After
all, it's going to a single ended IF, right?   But if you do it this simple
way, then whatever unwanted IF frequency manages to get to the beam tube
control grid will be coupled to the one plate you're using both by the
electron stream and by capacity:  In other words, no rejection.

Even the signal band gets amplified and goes through the tube, possibly
causing problems in the filter or the 1st IF stage.

This is an even bigger deal than it may seem because one of the great
advantages of these beam tubes as mixers is that they're quiet enough (and
tough enough) to stand alone with no RF stage, meaning that in simple
construction they will often have just one tuned circuit separating them
from the antenna.  

Push-pull output is the best way to go in these situations.   With careful
(balanced!) design you'll get another 20db-up of rejection for IF blowby
(and other antenna) signals and provisions to trim the balance can add
another 10+ db to that.   These tubes were very often used as balanced
modulators and although carrier rejection of over 30 db was hard to maintain
in practice (because of drift in tube parameters) they did generally achieve
that, I believe.

Note also that if you use a half lattice filter following the mixer then
push-pull to single ended conversion is built in.   You just need a nicely
balanced (center tapped) tuned circuit in the mixer plate.

Walt 
KJ4KV