so i finally got around to finishing this DC receiver ive been breadboarding for about 3 months now. its based loosely on the Neophyte design but i swapped out the detector stage with a pair of 1N4148s in a ring configuration and im running the LO off a si5351 module i had lying around from another project. overall its working way better than i expected honestly, pulling in signals on 40 pretty well considering its just a cardboard box chassis right now lol

but heres the thing thats bugging me — im getting noticeable LO leakage out the antenna port. like if i put my SDR dongle near the antenna lead i can see the carrier sitting right there at whatever frequency the si5351 is putting out. i know DC receivers are inherently going to have some of this but is there a standard approach people use to knock it down? i was thinking about adding a low pass filter on the RF input but i dont know if thats really the right tool here or if im better off just improving the shielding on the LO side. the cardboard chassis isnt exactly helping i imagine

also as a side note the audio chain sounds a little muddy on SSB — running it into an LM386 amp which i know isnt ideal but its what i had. wondering if thats a filter issue or just the LM386 being the LM386

yeah the LO leakage thing is pretty much just the nature of a direct conversion design, especially if the detector isnt super well balanced. the ring detector should help versus a single diode setup but youre never going to get it to zero. a simple low pass or bandpass filter on the antenna input wont really stop it from leaking out because the problem is more about isolation between the detector and the antenna port — the LO is basically backdriving through the mixer. better shielding on the LO and making sure your si5351 output level isnt way hotter than it needs to be will probably do more good than a filter. what level are you driving the si5351 output at? those things can put out a lot more than you actually need for most detector configs and backing it off sometimes helps

the LM386 thing — yeah that chip is notoriously noisy and has a pretty weird frequency response without some help. the classic fix is the 0.047uf cap between pins 1 and 8 to set the gain and a bypass on pin 7, theres a million variations of the decoupling network online. wont make it sound like a nice audio amp but itll clean up some of the mud

cardboard chassis lol i built a 20m regen in a cracker box once, worked great until humidity got to it. anyway for the leakage i always just wrap the detector section in copper tape and make sure the ground connections are solid, made a noticeable difference on mine. not a perfect fix but cheap and easy to try before you commit to a real enclosure