OH @#$&(#!!!!!!!! I have WiFi problems you wouldn't believe in addition to it being barely above dial up level (damn cheapskates) and severely overloaded even before adding a few med carts that bogs it down even worse. "They" changed several channels bungling it more, now I get disconnected several times a day and struggle to get reconnected, meanwhile I lose what I was doing, this time my nearly completed reply.

Trying to pick up a few of the pieces, you're on your way for the most part, so I suggest whipping it into shape before you even THINK of a 160M mod that for one thing isn't needed, 100W is quite enough to get you where you want to go. Working under supervision not the least bit needed except for legal reasons I worked from 160M up through 6M with 100W, less above that, and worked the world. Remember that 160M is MW like AM broadcast (HF begins at 3MHz) and is clear channel bottom to top and coast to coast. 100W is like BS in business, a little spread thinly goes a long way. Just for fun I worked Georgia, USSR on 10M AM with a re-tuned CB set, no BS.

You're running those 4-400s like they're supposed to be run with a proper power supply, and better when you bias them for Class AB1 or Class AB2 linear and Class C CW operation like the factory specs call for. I hope you downloaded a pdf page from the RCA Transmitting and Industrial Tube Manual I had once upon a time, or the sheet from Eimac that will guide you on the path that you must follow. I noticed a couple of transformers that need attention, T103 LV & bias is connected for 110V mains because that was nominal when the amp was made, put it on the 120V lead because that's today's nominal. Then there is only one way of dealing with T102 powering the 4-400 filaments that require 5V +/- .25V because they're very fussy about that. Being run out of spec shortens the life of a couple of very expensive tubes, price them, you'll flip! I suspect voltage to be a bit on the high side with higher mains nominal now, so if it is first remove the tubes and meter one of the secondaries no load for reference. Then try wiring the unused secondary winding in series with the primary and measure again to see which way it goes, up or down. If it goes up reverse the winding, in phase boosts, opposite phase bucks. The idea is to buck out some of the primary input to reduce secondary output, with a bit of luck you'll wind up with filament voltage under load within specs. If not it's time to get out the big guns, but we'll hold that idea for later only if needed.

Why do you think you need doorknobs to replace C10 that maybe shorted and put the full output of the bias through L8 and blew it open? All you need there is a new 2.5mH 4 pie RFC with or without Cosmoline (;->) and a .001uF 600V ceramic disc cap. Looking at the 4-400 spec sheet the maximum voltage to be expected at that point assuming under load in AB1 plate 3500 and screen 750 is -145, and AB2 plate 3500 and screen 500 is -75V so the 600V cap should be sufficient. With those parts replaced and before waking the sleeping dragon it's a good idea to check for shorts, something else may be providing a DC path to ground.

"I believe all these caps that are shot probably contributed to the issue I saw as I was trying to load on 40m just before the choke blew."
AHAAA, so it was YOU who blew the choke! Then maybe it's not because C10 shorted, but no harm in replacing it anyway, now it's looking tube related, maybe a grid short. That's easy enough, with voltages all wrong grid dissipation shoots skyward, they can overheat and sag against other electrodes, fizz, zap, BANG! BTW there's a mistake on your schematic, that dot at the filament common point needs a ground symbol.

"There was for sure not enough capacitance coupling"
Possibly correct, it could also be a parasitic VHF oscillation problem.
"and C10 being almost invisible, probably does something bad for the tubes output plate capacitance by the looks of it."
No, it's a bypass cap that grounds the cold end of the grid tank for RF while leaving it float for DC bias voltage. At the end of the day you'll probably have to neutralize any parasitics by adjusting C2, you'll need the manual for that. Parasitic oscillation can cause tube overheating, and some really weird tuning and loading effects, like maximum output on one side of the dip in plate current for one thing.

Until next week, same bat time, same bat channel. Don't try it, only bats can sleep upside down.

Bats in Rudder Park Kempsey 2011.jpg