[G6Tanuki]

My late-1970s/early-1980s synchronous-demodulation experiments [aimed at getting better reception of medium-wave broadcast stations after dark] focussed on a selective-sideband approach: feed the IF signal through three filters - a USB/LSB pair to separate the two sidebands, and a really-tight [200Hz bandwidth] filter to extract the unimpeded-by-modulation carrier.

Then mercilessly amplify/ clip the carrier to the point where it's a square-wave - and use it to feed a pair of diode-quad mixers [product-detectors] driven by the outputs of the upper- and lower-sideband filters.

And then combine the two resulting audio-signals in appropriate phase.

The theory was fine; the reality was that the group-delay characteristics of the three filters was inconsistent so when combnining the two outputs their phase-differences varied horribly with frequency.

I got far better results by pushing the IF signal through a tighter filter - essentially throwing-away one sideband - and using the 'clipped-carrier' to lock a PLL based around the NE561 chip to provide the carrier-coherent component for the demodulator.

If doing this today I'd feed the 450KHz signal into an analog-to-digital converter and after that the rest becomes a few hundred lines of non-challenging software....