Spent some time this morning looking at the wiring diagrams and it still isn't making sense to me.
One of the tests I did last night was to bypass the ignition switch again and ensure that there was power to +X, +15, and +54. I didn't also include +S because that is supposed to only be for the key and seatbelt warnings. According to the FSM, that should represent the key in the "drive" position:
I also verified that when the key is turned to "drive", I get battery voltage on pins 15, 54, and X at the switch. I didn't attempt a voltage drop test on those, but everything behaves the same way when I have the wires clamped together with a hemostat, so I'm assuming that the switch is working okay there.
When I do that, I get the fuel sender tickling and the low fuel light on, the coolant temp needle moves a bit, and the lights come on for Check Engine, oil level, charging, parking brake, and ABS. The Brake Fluid light doesn't come on and the pump doesn't run, even though the pedal is hard as a rock.
If I either pull the +15 wire from the clamped mess or turn the key partway between "Parked" and "Drive", the cluster is still and dark except for the charge light and the Brake Fluid light. The pump will run and will then shut itself off. If I leave it like that and pump the brakes a few times, the pump will kick back on and then turn itself off. So it seems like the power supply and the relay feeding the brake pump is working correctly at that point and the pressure switch is working correctly. But, if you look at the wiring diagram, the brake fluid light - and the pressure switch - should be getting their power from the +15 circuit which is not showing any voltage at the ignition switch:
#294 is the pressure switch and #293 is the relay for the pump motor. The control circuit pin 86 comes from +54 off the ignition switch and is grounded through the pressure switch which should be getting it's +12 through the Brake Fluid light off of +15.
This is where my lack of understanding about how electrons move about the cabin makes me start to stare out the window aimlessly and drool a bit. If +15 goes straight from the ignition switch to the gauge cluster - which it appears to do - then that light shouldn't be on when +15 is unplugged.
You can see 47F in that diagram off on the mid-right. The path goes through the ignition switch to fuse #22 then through a distribution terminal and then to the bulb. But it also goes to the parking brake bulb and that isn't lit.
So that means that "something" is getting voltage to the pressure switch, pump relay, and/or the bulb when +15 is disconnected, but when +15 is connected, that voltage is getting interrupted somewhere.
I have absolutely no idea how or if this system worked before I started monkeying with it. All I did to the system was bridge the relay between ports 86 and 87 which would have run the pump on the 10amp +54 circuit instead of the 30 amp feed direct from the battery while bypassing the pressure switch. When I got it, it had the wrong relay installed in the socket which would have meant that the pump was getting its +12 from the 10amp fuse fed by +54 from the ignition switch while the control coil was getting its +12 from the 30 amp fuse fed by the constant +12 from the battery. That would have meant that if the ABS computer or the pressure switch could provide a path to ground with the ignition off, the control coil would be constantly energized. And before I re-pinned the connector to accommodate the standard Bosch relay pinout and soldered in new diodes it would click when the battery was hooked up with the ignition turned off. And I'd get the really dim gauge lights.
With the new diodes and the re-pin of the socket, though, those lights are out completely with the ignition off.
The wiring diagram shows two diodes in the gauge cluster setup, though, and I'm wondering if those should be tested next. One is between the Brake Fluid light and the charge light and another is between the charge light and the alternator grounding point. I am also suspicious of the ABS computer itself. I cannot get it to flash any trouble codes for me despite the light being on, and every story I see on the internet about that problem is fixed with a new ABS computer. But if there's something janky with the wiring that cooked the ABS computer, I'd really rather not just fry another one.