Crystal sets? Any interest?

[Doodlebug]

Did anyone build a crystal set to receive 405 TV sound? :-)

[Mike Watterson]

They don't work well enough at VHF Band 1 never mind, Band II or Band III.
Also I doubt at Band1 you could get enough Q and selectivity to separately receive the sound without vision with just one tuned stage, The pre-WWII TV TRFs had a lot more tuned stages than a radio TRF.

At 1MHz a modulation of +/- 4kHz is  0.8% bandwidth, that's about 400 kHz at 50 MHz. Likely a crystal set can't easily differentiate two adjacent channels without a giant loop aerial and nulling one, if it's a suitable angle between the two.
At 50 MHz a modulation of +/- 10 KHz  is 0.04% bandwidth. A crystal set might receive over 1MHz bandwidth, assuming you were close enough to the transmitter with a big yagi and a fantastic Q of tuned circuit.
At 200 MHz it's 0.01% bandwidth. A single tuned circuit might pick up 5MHz or more.

I've built one transistor super-regenerative receiver tuned from about 60 MHz to 130 MHz or more with a straw in the coil that had a brass grub-screw and a ferrite slug separated by slightly more than the coil length.
   
This is a similar one built about 1990 to fit in a Swan Vestas matchbox. Runs off an AAA cell. The extra transistors are very low current drain (thus poor gain per stage) audio amp for 32 Ohm earpiece.

The limit to Q on a crystal set is the diode load, either 2K + 2K phones or ceramic beeper with 10K in parallel with about 5nF (the beeper self capacitance). The old crystal earpiece might have let you get to 22k, but the capacitance was still high. A better Q single tuned TRF needs either a FET, valve or low bias transistor like in the ZN414 first stage (not as good as jFET or valve).

All the fancy stuff on more complex crystal sets is simply to match long wires or loops. Any method of increasing selectivity dramatically decreases sensitivity. A 2m loop that is the coil, tuned with air spaced cap, a 1N60 and a ceramic earpiece with 10k Ohm (to flatten response and give DC load) is about as good as it gets. You'll get more Q in the coil using more wires in parallel, or massively fat Litz (3mm), or copper ribbon. I've used screened Cat 5 cable to make switchable loops for LW/MW/SW. The screen is a SW loop. Screen floating on MW and one end earthed on LW. The 4 pairs are switched to give all 8 wires in series on LW and only pairs in series on MW. About 5 to 15 turns in a square, even wound on a cardboard box, depending on size. Marvellous on valve sets that have no built in to reduce interference compared to a long wire. Not all sets will match to a loop rather than long wire without a transformer.