BTW, real air trimmers seem to be far better for temperature stability
than normal plastic dielectric trimmers! In an AM broadcast
transmitter using a VXO, a plastic trimmer in series with one of two
crystals(mixer setup) gave instability and a deep hetrodyne drone that
got worse over several days(unattended remote pirate rig) even though
the outside temperature was the same.

Replacing it with an air variable trimmer whose plates were apparently
cut from two solid pieces of brass mude for utter stability, with no
sign of drift over at least a 20 degree temperature change without
having to reset it.

For noncritical applications, those little plastic variable caps from
car stereos are pretty good. Have not used one in an ultra-critical
application in a role giving a lot of control over frequency, but when
I used one in the 2004 rig to shift a crystal maybe 2KHZ out of 16MHZ,
it didn't seem to drift.

Of course, a VXO(pulled crystal) is a hell of a lot more stable than
any VFO! A VXO and a similar but not the same frequency fixed crystal
can give a suprisng tuning range with stability. Put in a crystal oven
or even a heated/air conditioned room it would leave little to be
desired in stability.

For a VFO for any application, the better your parts, the better
your results. Wind coil on "air" or unity permeability cores such as
wood or ceramic, and epoxy the windings in place. Blow on an
oscillator's coil while listening to the beat note with and without the
epoxy, and hear the difference for yourself. Wood cores seem to work
fine with the epoxy covering.

With an air-core coil,a good tuning capacitor, and a circut that
minimizes active device contribution to drift, you end up with a lot
less chasing drift to do. Best active device for any VFO and probably
any VXO as well is a JFET. Almost no heat(unlike a tube), and no
junctions in the current path to change characteristics with
temperature(unlike a bipolar). If you must use a powdered iron core,
keep DC out of the windings as changes in the DC current change the
permeability opf any ferrite or powered-iron core. Ferrite cores of any
type have been named as an especially bad source of drift in
oscillators, so don't use modified IF transfomers as tuners in
oscillators expected to be stable. They are fine in tuned small-signal
amps, just not in oscillators.

In that VXO with the bad trimmer, I got lucky and found the bad part
first try, but this is unusual. If you have to track down drift expect
hours of work. That's why it takes less tiem to use the good stuff from
the start, unless it takes hours of driving or biking to obtain it, of
course.