Michael Black wrote in
news:alpine.LNX.2.02.1409131746360.24400@darkstar. example.org:

In the old days, one scheme for adapting old receivers for SSB was to use
a local signal generator of some kind on the signal frequency, coupling it
into the antenna input of the receiver. There, it only needed a weak
signal (while by the time an incoming signal reaches the detector in a
receiver, it can be quite strong, and the BFO needs to be stronger unless
there's a product detector in there), and you can adjust injection with a
simple level control on the output of the local signal generator. Another
advantage is that for receivers not meant for SSB, they would tend to
drift, while you could build a decent VFO or something that would be
stable. So the "bfo" would stay in tune with the incoming signal, even if
the old receiver would drift off. So you'd need to retune the receiver
every so often as it faded into the skirt of the selectivity, but the
"bfo" and the incoming signal would remain the same, so none of that oddly
sounding voices.


Thankyou, I like that analysis. I had already wondered about drift, and this
is the way to go, removing any relation to the IF frequency or its stability
(I looked at the datasheet, but it's big, I stopped after a crude search for
uppercase IF, and found no mention of exactly how low...). Avoiding a need
for a strong emission and risk of harmonics getting back to the RF input is
good too. It does mean I have to make some small signal generator with fine
accuracy good for up to 23 MHz, but that is doable. I suspect the Maxim
version of the old 8038 can do at least 20MHz, and maybe some versions of the
4046 PLL oscillator can do it too (but not sine so not as convenient).

All worth doing though, the low cost, small size, the DSP, and the AM filters
in the PL-390 make it very tempting to improve on its all-rounder capability.

One thing I discovered was that while the HF and VHF inputs do not use the
ferrite rod's original coil, I did find that shortwave seemed to benefit from
the 50m length of wire I tested recently, when fed to the 100 turns I added
on the rod's end, instead of to the intended input on the RF jack's tip. So
coupling a small signal to the ring terminal and its extra coil at same
frequency for CW and SSB seems very likely to work with no further internal
modification.