I signed up for a 72 hour access at helminc.com and found info on where the floor/panel actuator is located and how to remove it. I took a whole lotta dash apart which in the end wasn't necessary. I found the actuator and learned that it's a three position actuator with two vacuum line connections. One line pulls the diaphragm halfway down when connected to vacuum to enable the panel/floor setting. When the second line is also connected to vacuum, the diaphragm gets pulled down the rest of the way. Anyway, both worked when vacuum was manually applied so it isn't the actuator. I also found the door itself was fine when moved manually. I guess there's a problem with the blend door that is actuated with a stepper motor and driven into its mechanical stops whenever it self calibrates. Luckily that hasn't failed for me.
Next I wanted to test the lines and the back of the EATC is a convenient place. I used a tee and two unions from the hardware store along with some fish air-pump tubing to jump black (vacuum source) to blue and yellow. I was able to control both. This meant that the EATC was the culprit - not applying sufficient vacuum. I'd seen another post about o-rings that go bad and I gave it a shot. Sure enough it worked. While I had the EATC open, I also tested the solenoid resistance and found them all at about 50 Ohms. Must be the o-rings.
I used 5/16" OD, 1/16" material. The post below suggests a 9/32" OD and didn't specify the material size. The original o-rings didn't look damaged and the material diameter was much smaller than 1/16". I first tried just spraying some silicone on the components but that didn't help. As soon as I put the new (slightly larger) o-rings in, the problem was solved. I can control floor, panel/floor and no-floor correctly. Yay!
See the o-ring replacement post here. A lot cheaper than a new EATC.
EATC is Rebuildable
BTW, the EATC passed the self diagnostic even when it had the o-ring failure.