so this has been going on for about three weeks now and i cant figure it out. the rig transmits fine for maybe 5-10 minutes and then just... drops. ptt still activates, the tx light comes on, ALC is moving, but no rf out. sometimes it comes back on its own after 30 seconds, sometimes i have to key up and unkey a couple times. happens on all bands, 40m seems worst but that might just be because i spend more time there.

i pulled the finals and they look fine visually, no discoloration, solder joints look okay to me. the pa bias hasnt been touched since i bought the radio used about two years ago. i was thinking maybe the low pass filter relay is sticking? i swapped to a dummy load and it still does it so i dont think its antenna related. voltages on the 13.8v rail look stable on my meter while its happening but i havent put a scope on it yet.

anyone dealt with this before on an 857 specifically? ive seen some stuff about the finals going soft but this doesnt quite match that pattern to me.

had almost exactly this on an 857D i picked up at a hamfest, different symptom pattern but same frustrating intermittent thing. turned out to be the PA thermistor giving flaky readings and the thermal protection was kicking in even though the finals werent actually hot. you can check if thats whats happening by watching the ALC behavior right as it drops -- if the ALC slams to max for a split second before rf dies thats probably not it, but if rf just quietly disappears with no ALC weirdness the thermal trip is worth looking at.

also worth checking the driver stage transistors not just the finals. a lot of people go straight to the finals but ive seen the driver board cause exactly this kind of intermittent dropout. the 857 service manual has the voltage charts for the driver board and its not hard to check if you have a decent meter. the relay thing is possible too but i'd rule out the thermal stuff first since its a quicker check.

does it happen faster when the shack is warm? asking because mine did something similar and it was literally just a cold solder joint on the pa board that would open up once things got to temperature. took forever to find because it looked totally fine visually. had to reflow basically the whole pa section before i found it with a thermal camera a buddy let me borrow.