so ive had this 857 for about 4 years now, bought it used from a guy at the club and it's been solid until a few weeks ago. started noticing my signal reports on 40m were noticeably worse than usual, finally got around to checking it with my dummy load and a power meter and yeah, its only putting out around 18-22W when it should be doing 100W. every other band seems fine, checked 20m and 10m and both are doing 95-100W no problem.

the radio itself doesnt throw any fault or protection mode, ALC looks normal on the meter, SWR into the dummy load is fine. i did the usual stuff first — reseated all the connectors i could get to, checked the power supply voltage under load (sitting at 13.5V solid), cleaned the band switch contacts with deoxit. nothing changed.

my first thought was the low pass filter board for 40m but i dont really know how to verify that without swapping in a known good one. anybody been down this road with an 857 before? wondering if its worth sending to an authorized service center or if this is something a reasonably careful guy with a scope and soldering iron can trace down at home.

had almost the exact same symptom on a 897D a couple years back and it turned out to be one of the driver transistors starting to fail — not fully dead, just weak on that band. the 857 and 897 share a lot of the same PA architecture so it might be worth looking there before assuming the LPF. the way it showed up on mine was inconsistent output that would vary a bit with temperature, like it would be closer to 30W when cold and drop as it warmed up. does yours do anything like that?

if you have a scope you can probe the driver stage output and compare it against what the final PA is seeing — if the drive level into the finals looks low on 40 compared to the other bands then thats your answer and its upstream of the LPF. if drive looks right and output is still low then yeah id suspect the filter board or possibly a bad solder joint somewhere on the PA board in the 40m path. yaesu's service manual for the 857 is floating around online and actually has pretty decent block diagrams for tracing this kind of thing.

not sure if this applies but i once chased a similar issue on a different radio for like two weeks and it ended up being a cracked solder joint on one of the switching diodes that selects the band filter. super tiny crack, looked fine visually until i hit it with good lighting and magnification. might be worth a close look at the band switching section if you havent already, especially if the radio's seen any thermal cycling or been dropped at some point. those hairline cracks are annoying to find but easy to fix once you see them.