A much needed update. After weeks of probing in the wrong place, I got this solved. Sadly, as you will see below, if I had sent out my boards for repair, they would have (or should have) all been returned back as good. The few repairs I did were harmless, and possibly unnecessary other than to ensure the corrosion wasn't at play.
Totally pissed. However, in this "fun" process I became an expert at schematics, logic chips, circuitry, flow, and could probably fix any System 11B board - seriously.
So to recap, this Whirlwind board had two primary issues that I thought were related,
1.) Sound Overlay Solenoids all locked on at startup or never worked at all
2.) The music/speech didn't work, but the second test for the sounds all worked.
This board had prior battery corrosion, but it was neutralized, the traces looked OK, and all pinned out OK with continuity.
I replaced U42 (unnecessarily) -- although was on the right track.
I replaced the blanking circuit -- probably not related, but necessary.
I replaced SRC 1 and 3 (possibly necessary in the corroded area to clean up the board).
I took all the boards up to the arcade except the MPU. Everything swapped into another Whirlwind worked perfectly, meaning the MPU was at fault.
I purchased Leon's TEST ROM and everything checked out perfectly at the PIAs down to the output pins. Sigh.
Last night I was using the sound solenoid overlay board to try knocking out the bizzarre solenoid lock on or failure to work issue. I knew the problem was with the MPU, but I figured I'd work backwards. The U1 74LS374 chip was ending up in a seemingly "locked" with different results from when the solenoids were locked on vs. not working at all. The clock signal was HI, but after reading the data sheet on this flip-flop chip, it kept talking about the clock going low to high in transition, which got me thinking this morning that maybe it's supposed to be pulsing.
MD0-4 coming off pins 3-7 on 1J21 on the MPU are the solenoids, while MD5,6 and 7 (pins 8,9,10 on 1J21) drive connect via ribbon cable to 5J2 on the sound solenoid overlay board. MD5/6/7 all control both the clock on U1 and are also somehow used at U20 to work with the sound board.
Oddly enough, when I probed all the pins on 5J2 last night, I made an annotation next to pin10 (MD7) because the probe read HI, but it was "fuzzy" sounding as if there was a subtle pulse in there. AHA!!! I checked it again just now and sure enough, it's bizzare.
OK, back to the MPU. MD7 comes from U42, pin 17, which is PULSING CLEANLY!! I then check SRC pack 1, pin 9 next in line (which I had replaced last week). CLEAN. I check 1J21 pin 10. CLEAN. I check the other end of the ribbon cable at pin 10. HI and scratchy. ARE YOU F--KING KIDDING ME?!!
The problem: A bad ribbon cable. I jumped 1J21 pin 10 to 5J2 pin 10 and now all the music speech works perfectly and the solenoids behave perfectly!!
Lesson Learned: Logic probe signals should be clean - LO, HI, PULSE. I'm not sure how I could have saved any time here. None of my other games used a 20-pin ribbon cable to borrow from. In this case the test ROM passed, no errors with the game ROM, which despite my thinking the issue was at the PIA , pointed to something else. One would have easily suspected the corroded area, and despite confirmations of a strong continuity signal across the MPU traces, I replaced parts that looked suspect (probably good to get rid of the corrosion anyways). After that I used the game manual to understand the flow to the solenoids, and noticed that 74LS374 appeared "locked" in the turn on state, which is not how a flip-flop chip should behave. I noticed that the MD7 signal seemed off, and the key to all of this was the logic probe. After noticing the LS374 clock and the music were joined at that MD7 signal, it was then finding out how I was going from strong pulse to "scratchy hi".