Looking at how I would configure a 7360-type mixer in the front end of
a low-band (say 80M and 40M) receiver.

Method A: To me, the most obvious way to do it is to feed the RF into
the 7360's grid, and feed the deflection plates with balanced local
oscillator input at circa 10 or 20 V P-P. Put a center-tapped
transformer to the anodes and pull the output out of a winding.

BUT... that's not what I see being done in the various app notes.

Method B: I see, for example, feeding RF into a single deflection
plate, and putting a pot for fine-tuning the balance on the other
plate. LO is injected into the 7360's grid. (e.g. page 6 of
http://www.ab4oj.com/1st/hb/7360/7360.pdf ). I think the intention
here is not to saturate the deflector beam inputs into switching mode,
but to keep the amplitudes small where it'll be linear.

What's the advantages and gotchas of doing it each way? Method A takes
a lot of LO amplitude and has the highest gain for the RF. Method B
takes less LO amplitude but I think has lower gain. Method A cancels
out the LO, and Method B cancels out the original RF. Right?

Of course two 7360's would let me do a doubly-balanced mixer... can't
afford 7360's but I do have some 6AR8's!

Tim N3QE.