The problem with homodynes and synchrodynes is that their selectivity depends on the profile of the audio filter following the synchronous detector.
Stations which are 9kHz apart in Oz can interfere with each other because sidebands of the modulation of one station can stretch out into the bandwidth of another station, so you can get whistly monkey chatter that a 9kHz notch filter cannot remove. There needs to be good selectivity in the RF amp first stage but with no attenuation of the modulation sidebands; this means a TRF input stage with special tuning, and boy is thatv difficult to get right if you have never built such a thing; a complete pita.
A superhet using tubes and IF trannies is a lot more fun although hard work to build; the high IF voltage of about 10Vrms at 455khz is easily attainable with a pentode and thus a maximum of 5Vrms of audio signal at 100% modulation can be obtained and there is negligible thd generated in the diode detector buffered by a cathode follower.
But also a wide band ceramic IF filter could be used with solid state to get the ideal passband of 20kHz at 445kHz but with very steep roll off each side of the pass band.
There are other methods using 2 IF amp stages and having the IF at 2.5MHz, so that its easier to get 20kHz bw after 6 tuned circuits in 3 IF transformers. Winding the IFTs is easier as F goes higher, but again its a lotta effort and pain to teach oneself how to do it.
Its probably easier to digitise the antenna signal and use a PC to count out the AM station you want.
The simplest tube set I have uses 6BE6 input mixer after a ferrite rod antenna,
6BA6 vari µ pentode IF amp at 455kHz, then a 1/2 12AU7 cathode follower buffer to drive the diode plus CR peak detector, followed by another 1/2 12AU7 cathode follower to prevent loading the first RC filter so that you get very low thd in the diode detector.I get about 4Vrms average signal level and very clean from the set. The 6BA6 runs with about -5 to -8 V of avc voltage which makes the the amplification of the IF envelope fairly linear, just a bit of 2H.
I have also used an RF amp in another set which accepts all the avc, so that the IF pentode is a 6BX6 sharp cut off pentode which is more linear than the vari µ tube, and that sounds even better.
Audio bw can be increased by altering the distance between the two IF windings on IFT No 1.
Unfortunately most IFTs you find in old valve sets have very high Q so that only
10kHz of IF bw is possible per IFT, so that after two such stages bw falls to only 7kHz, thus giving 3.5kHz of audio.But in some setys I have cut the former of the IF coils in the centre and placed a cardboard tube sleeve over them to allow the coils to be slid slightly closer to further flatten the response of the IFT. Placing 100k across each IFT coil also reduces the Q somewhat and broadens the bw but as some cost to IF amp gain.
Fiddling around with all these things are terribly prone to oscillations at 455kHz which make the station tuning whistly across all or part of the band.
Every old radio set you may find has had the layout carefully considered and trialed and tweaked to get rid of stray parasitic oscillations, so if you get an old set to play with expect major amounts of time to get it working better.
Patrick Turner.