This was an unexpected surprise to be working on.
Handed to me to be repaired by a machinist friend whom bought it apparently pre restored some years back because of the striking cabinet.
And I can concur, I too like the look of this one.
But it was really the insides that surprised me!
2 valve, that made me scratch my head. And a 32 volt tube! More scratching.
Well yes they managed to achieve a lot with a 32L7 and 12B8.
I knew it was too old for a selenium rectifier so to find the pentode output had a seperate diode built in was rather interesting.

Other than a little hum, it was stone dead. It took me some time to identify what it was and find a hand drawn schematic.
Indeed it has been subject to it's fair share of work underneath in the not so distant past. Though some values of components used are a smidge off I doubt they'd make a difference to such a set.
No, after tracing things back it turned out both tunings capacitors were shorted in various spots. I have no idea how that comes to happen but it did.
Now the set works as good as I'd expect, not great and poor selectivity. Yep it's a TRF.
That's my show and tell anyway.

That will be very interesting to see.
It may be the set can be 'made' to work better than at the moment, if it uses some subtle tuning coil magic that needs tweaking.

Fre

Not counting the rectifier, R&H published an ingenious 2 valve superhet design in about 1949 that uses a 6AN7 and a 6AB8.

The triode of the 6AB8 was used as a regenerative detector in the IF amp position. Advantage was the "IF - detector" was fix tuned so the reaction control as set and forget.

Nifty!

Article with a 2-valve set, using 25A7GT and 12B8GT:

http://messui.polygonal-moogle.com/valves/americanmidget.pdf

Photos uploaded.

For those playing along at home, this is the website where I finally identified the radio

https://www.vintageradio.co.nz/model/hmv/midget_1

Cute little radio built to a price.

Very clean little set, plenty of room underneath with modern parts.
Be wary of any random bits of wire that don't appear to do anything.
Particularly around the tuning coils.
Resonant coils are very strange things particularly when they don't have a damping can around them.
There can be deliberate coupling or neutralising 'capacitors' that perform weird functions that 'tweak' the Q of the coils or the coil gain.
In that EA article the author almost says it where the open coils have 'coupling'.
I must do more work on coil resonance and small TRF sets that work just as well as superhets.
I stumbled into that with my 'Reinartz' and my single valve superhet efforts.
Pity you cant see RF fields in these old sets.
Fred.