[HBR] Partially balanced mixer

[Ian Wilson]

I've been playing around with mixers for a while. One thing I want to do is
to make
a heterodyne VFO that has reasonable spectral purity. I am sticking to
tubes, so
am not ready to cheat and slip in a diode DBM. My target band is 40m.

This evening I finally found a setup that shows signs of promise. This is a
slight
modification of a circuit in the Collins SSB book - a 12AT7 differential
amplifier:
cathodes connected together with a 470 ohm tail resistor; plates connected
together through a parallel LC tank to +250V; grids each have 100k to
ground.
You get some suppression of the inputs because each tube sees more or less
the negative of what the other tube sees. The suppression isn't great
because
the tail resistor isn't very large compared with 1/gm of the tubes.

Results using a 6J6 with 2.7k tail resistor are quite good. The 7MHz
component
(I am mixing a 3MHz xtal osc with a 4MHz VFO which is quite challenging) is
much larger than with the 12AT7, and is larger than the 6MHz second harmonic
of the 3MHz input.

The two halves of the 6J6 have grid bias of about +40V. B+ is 150V (I
designed
the circuit for +200V but forgot to change the PSU). The individual plate
currents
are about 7.5mA and the gm at this operating point is something like 4mA/V.
The drive level from the xtal oscillator is probably about 3V peak, and the
VFO
input is something like 1.5V peak. Output across the tank is about 4V peak
(I imagine that the Q is being lowered by the scope probe).

The signal needs a second stage of filtering but this is the best
performance I've
seen so far from a single-ended partially balanced mixer circuit. I do
realize that
using a 10MHz xtal and a 3MHz VFO would be an easier mixing task.

73, ian K3IMW