I have noticed couple times old folks to talk about stabilising VFO using voltage divider with NTC-resistor and varicap back in the day, before cheap PLL controls. I know these has no real usage nowdays, but im iterested to hear about it, maybe even try it if I have some time.

I guess that main idea was to stabilise rig that is located in outdoors, maybe on autumn weather when temperature might go from days +20°C to -10°C at night.

Fast Google search didnt return anything, but what old folks have been talking, NTC-resistor has been located very close VFO coil, and that voltage divider has had couple trimmers, I think one would be on the divider itself, to set voltage between it and NTC, and second would maybe have been parallel with NTC or between varicap and voltage divider.

What I know, they have been carrying VFO/rig out to the cold, and back indoors, or have used freezer/fridge to get it cold, and then warmed it, while tracking its freqency.

I dont mean old trick to get couple wirewound resistors inside insulated VFO casing, and transistor with NTC to keep that VFO stable at like +40°C.

Any older folks here who have idea bout that.

The best thing to get a stable VFO is to start with a known stable design for an oscillator, like using a J-Fet or Mos-Fet (because their thermal behaviour is beneficial to avoid drift, instead of BJT's), and the topology of the circuit matters, the best known is the Vackar type, or other types that load very lightly the oscillator L-C tank.
Then there are constructional tips, which include strong, rigid air-cored coils (wherever possible), mica trimmers, and latest trick is to combine ceramic caps of different temperature coefficients (some are positive and others negative, or zero (NP0)).
Closing the circuit in a metal box and having well stabilized and decoupled power also matters a lot.

Yeah, but Im just interested about that old folks way, never seen it in action.

Here's a sketch of what we used to do:
There were no hard and fast values to use, because the thermistor sensitivity could be very variable, and because the oscillators would also vary somewhat. The V+ on the circuit above would be zener stabilised (with three 5V1 zeners in series, which have no temperature coefficient). The voltage applied to the middle of the tempco varicaps ("Vt") was set to be roughly 7V. The variation in capacitance applied to the oscillator would be tiny - just 1 - 2 pF at most. We used to take hours getting the compensation right!

As soon as PLLs became possible - locking our oscillators numerically to a very stable crystal - we all migrated to that relatively painless solution. I used to build my PLLs from discrete logic parts, but as integrated PLL ICs became available (later on, usually ones from TV Tuners) the circuits became simpler!.

I did do a few medium wave rigs that locked themselves to the old 200 kHz signal from Droitwich. These were frequency accurate to better than 5 Hz! The circuit for this was relatively simple, but the only really tricky part was keeping the MW signal out of the 200 kHz receiver!

Nice! What I know, that way has been used from MW to FM here and there.

Thats just another part of old useless information that I like to collect.