Again, thank you!
You're making some good points, and it got me thinking. I believe the tension is ok. I did the 1/4 turn test as you mentioned, and according to that, I'm fine. I watched the pulley closely as I accelerate and decelerate the engine, and it doesn't move or seem to be making noise, but I did not specifically check it while it was apart, so definitely a possibility.
Side story - I had a truck once that would randomly throw the serpentine belt for no apparent reason. That thing drove me crazy with checking pulley alignments, bearings, etc. Until a guy on a forum, just like this, told me to replace the tensioner. I cried "BS" on him because I was young and thought I knew everything. Guess what... as soon as I replaced that tensioner, the belt stayed on. I, of course, had to go back and eat crow, and admit he was right. So, your comments about the tensioner certainly carry weight with me!
Anyhow, back to this - I believe the tension is ok, and I believe it's more or less consistent throughout the running cycle, you got me thinking... every engine has bit of timing lash between the crank and the cams as the engine is running. I work on mostly diesels (which almost all have gears, so the lash is fairly minimal, and no flex at all), but this engine has a timing belt. Honestly, I don't work on very many timing belt engines. I have to believe there is some lash, and possibly some flex as the engine is accelerating and decelerating. In other words, the valve timing seems like it would wander a bit. Possibly only by a fraction of a degree on either side, but some wandering. Assuming that's true, as the engine is accelerating, and the intake side of that belt is under tension, and possibly stretching a little, the valves retard just a bit. Conversely, as the engine is decelerating, the exhaust side of the belt is under tension, and the timing advances a bit. Again, possibly by only fractions of a degree. If I have the valve timing too far advanced in my static timing, it is possible that while the engine is accelerating, the belt is allowing the valves to go into the proper (or closer to proper) timing. When the engine decelerates, the belt then allows the valves to go too far in advance, and we are hearing piston interference! That correlates a bit with some observations I made. One being that the noise seemed to correspond with crankshaft speed, not camshaft. The other being the sound itself sounded like a valve slamming shut. It was probably the piston slamming the valve shut.
I checked my timing marks one more time, and I found something. The timing indicator down at the crank has 3 "teeth" on it. When I took the old belt off, the timing aligned with the middle tooth, so that's the one I used with the new belt. According to my manual, I am to use the top tooth. So, if that's true, and everything I wrote above is true, my valve timing is too far advanced, and would cause severe engine damage if I took it out for a drive. I'm going to reset the timing to that top tooth, and report back. This is fascinating stuff. I can't wait to test these theories.