so ive been at this for about three weeks now, built a direct conversion receiver for 40m loosely based on the NorCal design but i swapped out a few parts because i couldnt source the exact values locally. getting audio out of it which is exciting but the problem is there's this massive hum riding on everything and when i touch the coil form it shifts the frequency noticeably. like i can tune in SSB signals and actually hear voices which feels like a miracle honestly but the hum is bad enough that its hard to copy anything weak.

the LO is a colpitts oscillator running around 7 MHz, wound on a t50-2 toroid. power supply is just a wall wart i had laying around, regulated 12v but maybe thats the issue? i havent put it in a box yet either, everything is just sitting on a piece of copper clad with the components tacked down. i measured about 80mV of ripple on the supply rail which seems like a lot but im not sure what the threshold should be for this kind of circuit. any ideas where to start chasing this down?

80mV of ripple is definitely going to cause you grief in a direct conversion rig. the problem is the audio frequencies you care about are literally the same frequency range as the ripple artifacts so they mix right into your demodulated audio in a way that a superhet would mostly reject. slap a 78L09 or even just a LM317 set to around 9V with some decent filter caps right at the board, like 470uF followed by a 10 ohm resistor followed by another 220uF, and see if that clears up the hum. seriously that alone might fix 90% of it.

the hand capacitance thing on the coil is pretty normal for an unshielded oscillator sitting out in the open. once you get it in a metal enclosure and maybe put a small can around the VFO section that'll settle down a lot. i built a similar thing a few years back and spent two weeks chasing hum before i realized my bench supply had a bad filter cap. always suspect the power supply first with audio stuff.

yeah whats your bypass situation look like on the mixer chip or however youre doing the mixing? sometimes people forget to bypass the supply right at the IC and the trace inductance is enough to let noise couple in. also what are you using for the product detector stage, a diode ring or one of the SA612 type chips?

not trying to pile on but the unboxed thing really does matter more than people think, ive had rigs that worked fine on the bench and then picked up all kinds of garbage once i put them near a computer or a switching supply. shielding a direct conversion is kind of annoying because you have to be careful not to change the coil geometry when you close the lid but its worth doing early rather than retrofitting later.