If I unplug the vacuum line at the vacuum switch I hear the air escape (i.e. good strong vacuum) and then the piston in the vacuum motor will return, the cable will release, the wheel will latch and the car will then start. So what releases the vacuum?
Can you clarify the statement above? There is no "vacuum switch". The two vacuum related components in the passenger compartment are (1) the vacuum release valve (aka solenoid valve) and (2) the vacuum release motor. Did you unplug the source hose into the vacuum release valve - item (1), or either end of the short hose between items (1) and (2)?
If you enter the vehicle, close the door, and pull the steering wheel to normal position the aforementioned item (1) vacuum release valve will be de-energized, and bleed vacuum off of the connection to the item (2) vacuum release motor, and the wheel should remain in place. If you followed these steps and the wheel did not latch, the first thing to do is to check the vacuum system, which is what Richard was alluding to in the first reply.
Go ahead and re-establish the conditions you originally had - a fully charged vacuum canister, engine recently shut off, door open, wheel away. Check the voltage on the wire feeding the item (1) vacuum release valve. It should be 12 volts, because this is what energizes the vacuum release valve which ports vacuum to the vacuum release motor. Step 1 will be to make sure that the electrical system is working properly. When a voltmeter still on that circuit, close the door. This should result in the door jamb switch being engaged, opening the contact to the interposing relay coil, and interposing relay contacts opening, de-energizing the vacuum release valve. If that is not the case, we need to drill down further into this electrical circuit. Please advise.
If the electrical circuit works as described, then I go back to your statement describing "good vacuum" "escape" followed by being able to latch. I would suspect that the item 1 vacuum release valve, which is essentially a solenoid valve, is malfunctioning. Not enough leakage, mind you, to release the latch under normal conditions; but enough leak-by that prevents proper bleeding and subsequent latching. The actual bleeding is done at the item 2 vacuum release motor. In other words, I don't think that the solenoid is a three way valve design...but I have not taken one apart to be certain.
Based on what you described, I don't think there is any problem downstream from here, such as with linkage, cables, etc. After all, when you released vacuum from the vacuum release motor everything worked as expected, which tells me that all of the mechanical wizardry is fine.