Another Rod pentode Pantry TX

[Kalee20]

I used NFB in mine, see here, now 11 years running 24/7.

Though, I'd certainly accept that your trapezium figure shows only a gentle curve, and yes I'll bet it does sound good!

[Murphyv310]

Hi Peter.
I remember that project from then, I was seriously impressed.
I spent some time earlier on and made a suitable Diode detector and filtered out the RF, got around 10v P-P of detected audio. Fed this back to the audio input grid with R C coupling, I tried many different combinations with very little success.
I did though manage to get a little improvement in the trapezium by slightly increasing the negative grid bias on the modulating section of the valve, - 0.5v inserted into the grid from a CR2025 via a potential divider made the difference. Also looking at the RF waveform with 1khz modulation now gives a perfectly balanced waveform even at 95% and lower levels, RF output drops slightly but there is plenty of reserve. I also found this circuit still functions on 18v HT and is perfectly happy at that voltage, obviously output is much less but it still fills my shed adequately. With lower volts lower drive can be used.
I also added a small trimmer of 50pf in parallel with a 68pf to the crystal which is now lifted from the G1 and the two components are inserted between the crystal and G1 now, this allows a small adjustment to the crystal frequency and removes DC from the crystal.
I think now I'll leave it as is.
I want to now transfer it to matrix board and make the PSU, fully regulated LT & HT.
It may be two or three weeks before the finished article is done.