Richard Clark wrote in
:

At 1MHz, this 50 KOhms
would be a capacitance imbalance of 3 pF. Looking at the wire
dressing of the coils in the photo would suggest 3 times this easily
(and no attention has been paid to this at all).

That's the bit that seems most important, and I can also relate to it. It
doesn't take much to cause a few pF difference.

A few points occur to me though...

First, to get it out of the way.. if this was causing serious bother, which
it surely might if it is that bad at the lowest end of the SW range, what on
earth induces a designer to persist in thinking it's working? Either it
isn't, and he's deluded; or it is, so why?

Second, I imagined balance to pretty much relate to symmetry. I hadn't seen
that file yet (even though I'd actually grabbed it with intent to), and the
looseness of the coil wiring isn't lost on me. My idea was to omit the amps
and just wind the toroids with neat symmetry to reduce obvious causes of
imbalance and take it from there. If it works, I use it, if not, I try
something else. Given that I have more than once been told that I might
overdrive the input on the ATS-909 radio with a 18' whip in the back yard, I
decided that I might as well omit the amps as they only boost a few dB, and
instead rely initially on the built in attenuator for first efforts to see
what's out there, then consider building (or getting lucky with on eBay) a
preselecting filter on the input.

Third, the amount of effect a few pF has on a circuit would depend also on
the inductance, or resistance, or any delay in the circuit. I hadn't looked
closely at that (I'd want to see what happened with a simple test first), but
I am guessing that the simpler idea of a twin wire with transformers and no
amps would have a smaller risk of imbalanced signals, so a better common mode
rejection. Why not do the amplification after the second transformer where
balance is clearly irrelevant, if it has to be done at all?

I may still be missing something other than a grasp of quantities, but
whatever I do, it has to be something that aims to do what that scheme was
said to be able to do. I don't have a lot of space, and that radio ideally
needs a single antenna to cover all of its AM range, at least for initial
efforts. Ideally some scheme that can be improved rather than thrown out when
I need something better.