so ive been building this direct conversion receiver for 40m from scratch, nothing fancy just an NE602 mixer feeding an LM386 audio stage with a simple LC bandpass filter up front. got it receiving signals last night which was exciting but the audio sounds like everything is coming through a wet blanket. LSB stations sound kinda intelligible but there's this low frequency rumble underneath everything and the audio just doesn't have any crispness to it.

the LM386 circuit is basically straight out of the datasheet, 250uf cap from pin 7 to ground, gain set to 200 with the 10uf cap between pins 1 and 8. i'm wondering if that's actually too much gain and it's just amplifying everything including noise. or maybe my audio filter is wrong. i did slap a 0.047uf cap from the output to ground which i thought would kill the RF but maybe it's also cutting too much of the audio band.

anyone built one of these and can tell me where the muddy audio usually comes from? power supply is a 9v wall wart, i haven't checked how clean it is yet

the wall wart is almost certainly your problem or at least a big part of it. those switching supplies dump a ton of garbage into the audio on sensitive low-level circuits like this. i had the exact same symptom on a similar build and adding a 78L08 regulator plus some filter caps on the supply rail cleaned it up massively. also try running it off a 9v battery just to confirm that's the issue before you go changing anything else.

the LM386 at max gain is also just inherently noisy and prone to oscillation, some people put a 10 ohm resistor in series with the output to stabilize it. but honestly diagnose the supply first, the muddy low frequency rumble is classic switching supply hash.

yeah what he said about the supply but also your 0.047uf cap to ground on the output might be rolling off too much of the high end. for audio you usually want that cap to be like 0.001uf or even smaller, 0.047 is pretty big and will definitely take the crispness out. the crossover frequency with a typical 8 ohm speaker load would be way down in the mud.

also the NE602 wants to see a fairly high impedance on its output pins, if you're driving the LM386 input directly that might be loading it down. a simple emitter follower buffer between the two can help a lot. i built a 40m DC receiver a few years back and spent way too long chasing audio issues before i figured out it was just impedance mismatch between those two chips