so ive been working on a direct conversion receiver for 40m for the past couple months, mostly following the NorCal 40A schematic but modified a bit because i couldnt source a couple of the original parts. got it mostly working now — can hear stations, SSB sounds reasonable — but im getting this really annoying carrier null issue where i have to constantly retune even after it warms up. like i'll peak it on a signal and 10 minutes later its drifted enough to sound hollow. VFO is a Colpitts design i wound myself on a T68-2 toroid.

im thinking its either thermal drift in the LC tank or maybe my supply isnt regulated enough, running off a wall wart right now. tried adding some thermal mass to the coil by potting it in hot glue which didnt really do anything useful. anyone dealt with this on homebrew DCRs, is it just a thing you live with or is there something im missing in the build

yeah that drift on a Colpitts is almost always the caps, not the coil. hot glue actually makes it worse in my experience because it has a pretty bad tempco and it'll stress the windings as it cures. if you used NP0/C0G ceramics in the resonant part of the tank you'll be in much better shape. polystyrene are even better if you can find them but those are getting hard to come by. also your wall wart is almost certainly contributing — any ripple on the supply modulates the varactor-like behavior of any junction capacitances in the circuit. slap a 78L08 or something on there and see if it tightens things up before you go crazy chasing the VFO.

i had almost the exact same problem on a receiver i built a while back and spent way too long trying to fix the VFO before i realized the mixer was the culprit, one of the 1N4148s in my diode ring was leaky and it was doing weird things to the balance. not saying thats your issue but worth pulling the VFO signal and checking it standalone on a counter to see if the drift is actually in the oscillator or somewhere downstream. took me forever to figure that out and i felt kinda dumb after