Re: Franklin VFO ?
 

I think that the next step would be to make an oscillator using a couple of low idss SMD JFETs in parallel with the drains soldered direct to a metal ground plate on a heat sink.

Then run from a negative supply and make a regular Colpitts or Hartley circuit. Hopefully, this would be even better, with low phase noise, fast warmup and low drift over time. The SMD package would hopefully give direct thermal contact from the JFET channel to the heat sink and this would give good thermal stability within a few seconds.