so ive been chasing this problem for about three weeks now and im starting to lose my mind a little. my 857 is putting out maybe 20-25 watts on HF when it should be doing 100. it's not ALC clipping, i checked that first thing. SWR is fine, dummy load reads the same, so its not the antenna or feedline lying to me.

what's weird is it does full power on 2m and 70cm. so the final transistors on those bands seem fine. HF finals maybe? i pulled the top cover and nothing looks obviously fried, no burnt smell, no discolored components that i can see. voltage rails look normal on the meter. i haven't gone in with the scope yet because honestly i was hoping someone had seen this exact symptom before and could save me some time.

radio is about 12 years old, bought it used so i have no idea what it went through before me. anyone dealt with a partial power loss like this on an 857 specifically?

yeah that's almost certainly one of the HF final transistors going soft rather than full dead. RD70HVF1s if i remember right on that radio. when one goes weak it drags the whole pair down because they're supposed to be matched. you dont always get the burn smell if it degraded slowly over time rather than a catastrophic failure.

the thing i'd check before assuming finals is the bias circuit. if the bias trim pot drifted or there's a bad solder joint on the bias board you'll get exactly what you're describing - reduced output, no obvious damage, works fine on other bands because they have separate bias paths on that rig. grab the service manual off the usual sites and look at the APC circuit too while you're in there. sometimes the APC feedback is pulling power back for no good reason and a scope on that line will show you immediately whether the radio is actually trying to transmit or choking itself.

if it does turn out to be the finals, shop around - found mine cheaper from a Japanese parts supplier than from the usual US distributors, took two weeks but saved like 40 bucks.

had almost the exact same thing on my 897 which shares a lot of the same guts. ended up being a cracked solder joint on one of the driver stage transistors, not even the finals themselves. super hard to see without magnification, looked totally fine to the naked eye. reflow the whole PA board if you can, not just the suspicious spots.