so this has been driving me crazy for a few weeks now. my 7300 runs fine on rx and even on low power tx, but if i push it past like 50 watts on 40m for any length of time it eventually throws the PA protection and drops to basically nothing. sometimes it recovers after a minute, sometimes i have to cycle power. other bands seem fine, or at least i havent been able to reproduce it on 20m or 17m. SWR looks clean so its not a load issue, ive got a good dummy load and tested that way too.

pulled the covers and nothing looks obviously scorched or swollen. been poking around with a DMM but honestly not sure what im looking for exactly. i read somewhere the final transistors on these can develop a soft failure mode where they only act up when hot but i dont know how to actually confirm that without just replacing them and hoping. anybody dealt with this before or have a systematic way to chase this down?

had almost the exact same thing happen on mine a couple years back, though mine was 80m not 40. ended up being one of the RD70HVF1 finals that was marginal — not shorted, not open, just leaky enough when it got warm to pull the bias out of spec and trigger the protection circuit. the way i confirmed it was with a curve tracer but most people dont have one of those sitting around. what you can do is pull the board and measure Idq on each device individually at room temp, then hit the board with a heat gun carefully and watch if one of them changes more than the others. its not perfect but it gives you a clue which one is drifting.

also worth checking the thermal paste between the finals and the heatsink while you have it open. icom doesnt always put a ton on there from the factory and if its dried out the thermal resistance goes up and the protection kicks in earlier than it should. might not be the root cause but its worth doing anyway while you have it apart.

is the fan spinning up normally when this happens? i ask because i had a weird one on a friends rig where the fan header had a cold solder joint and the fan wasnt getting full voltage, so it was barely moving air. everything looked fine externally but the finals were just cooking. the radio would eventually fault out and once it cooled down it was fine again which sounds kind of like what your describing. easy thing to rule out before you start pulling transistors.