so i've been slowly piecing together a direct conversion receiver for 40m over the past couple months, mostly following Wes Hayward's stuff from Experimental Methods but modifying a few things based on what parts i had on hand. got it working last weekend and honestly pretty stoked about it, can hear SSB and CW just fine, selectivity is obviously limited but its a DC receiver so what do you expect

anyway my issue is the LO leakage. i'm using a simple Colpitts oscillator running around 7.1 MHz feeding a doubly balanced mixer i breadboarded up with some 1N4148s and a couple of small toroids. the mixer seems to work but i can pick up my LO signal on another receiver sitting across the room, like a good 10-15 feet away. i know some leakage is basically unavoidable but this feels like a lot. my shielding is pretty minimal right now, just a tin altoids can for the oscillator section, not sure if that's the problem or if the mixer balance is off. anyone dealt with this before? the DBM balance is kind of fiddly to adjust and im not sure im doing it right tbh

also separate question — whats the practical limit on image rejection with a direct conversion design, i know its fundamentally limited but trying to understand if what im seeing is normal

the leakage across the room is definitely a shielding problem more than a balance problem id say. altoids cans are fine for audio stuff but at 7 MHz you really want a proper copper or brass enclosure with feedthrough caps on every line going in and out. the ground bond on the lid matters a lot too, a lot of people just set the lid on and wonder why it doesnt help much.

that said DBM balance does affect how much LO gets coupled back out through the RF port, so its worth spending time on. with hand wound toroids on a breadboard you're probably not going to get better than 30-35 dB of isolation anyway which is part of why the leakage is showing up. i built something similar a few years back and ended up just using a Mini-Circuits SBL-1 for the mixer and it made a huge difference even though it felt like cheating on a homebrew project. if you want to stick with the diode ring you can try matching the diodes more carefully with a DMM checking forward voltage.

on image rejection — a pure DC receiver has essentially zero image rejection, both sidebands fold on top of each other which is why everything sounds kind of mushy on SSB. you can do the phasing trick with a quadrature LO and two matched audio channels but thats a whole other rabbit hole

yeah what he said about the shielding, i had almost the exact same thing happen with a 20m DC receiver i built last year. what finally helped me was running the oscillator off its own small regulated supply separate from the rest of the board, the LO was coupling out through the power rail which i didnt even think about at first. once i added a simple RC filter on the VCC line to the oscillator section the leakage dropped noticeably even before i touched the shielding