Hello again Willen,
Here is a schematic for a one transistor radio. Once again, this is just a circuit idea and not proven but I think it has a lot of potential for development and should outperform the single diode type. By biasing the transistor on with around 20μA collector current, the diode VF and IR problems should be greatly reduced, hopefully!
Another advantage is that there will be a better drive for the earpiece and the antenna matching should be better, especially if tuning is designed in.
This approach does need a power supply, although a button cell will do. A few solar cells connected in series would be another possibility, unless you need to listen in the dark that is. Pretty much any supply voltage will be OK as long as the VCE of the transistor is not exceeded. To change the power supply voltage R3 needs to be changed in proportion but the value is not critical. Here are the values:
Voltage____R3
3V_______2M2
6V_______4M7
9V_______7M5
If the high value resistor for R3 are hard to get, let me know; I can change the circuit to suit what you have. Q1 should be a small signal type: BC109 etc. Once again not too critical, although a radio frequency type would be ideal, but it must have a current gain (hFE) of at least 20 at 20μA collector current. It would probably be best to divide the normal value for R4 used in single diode crystal radios by 10 and multiply the parallel capacitor, C4, by 10 to lower the impedence but keep the same time constant.
View attachment 95052
ERRATA
(1) The 0.5V at the base is incorrect: I forgot to add the voltage across R4
(2) It may be better to put a capacitor between the emitter of Q1 and the earpiece, say 10nF upwards, but not an electrolytic type.