Not sure where vacuum comes into the picture, except as it might affect closed loop. The ECM allows (or not) O/D engagement under certain conditions, including closed loop, i.e., the engine must be warmed up. Often this occurs at a specific temperature, as mine did, but that number changed with time.
The '84 logic is unique and requires other-than-first gear selected, closed loop, switch ON, 2nd gear below 50mph or 3rd gear below 60 (I think those speeds are accurate) and less than about 3/4 throttle. If the accelerator pedal is nearly 'floored', it will not reengage in 2nd or 3rd until 4th is selected. It may reset if you get back to 1st gear; I don't recall.
I once 'fixed' my O/D (it was cycling On/Off at normal speeds in steady cruise control) by changing spark plug wires. I swear, that is ALL I changed and the fix was immediate and lasting. My understanding is that the poor spark was allowing the O2 sensor to read a rich, or open loop state from my slightly modified engine.
Low vacuum (leak) may tell the ECM to add fuel, as it 'thinks' the engine is under load; that the throttle is open wider. I do not know why the TPS input would not override this, but similar issues arose with an aggressive cam grind, and low vacuum.
In either '87 or '88 the O/D was 'beefed up' with more springs in the clutch pack, etc. When I had my unit redone by the best experts, it was upgraded, after which, I was told that I would not break the it; that it is a tough unit. It now works fine behind my new engine, but as always, I unload it when engaging/disengaging, like when shifting the T-10.