Billy Bad Assr© wrote:

Apparently he's not the stupid one here.  Fuel pumps are not supposed to
run continuously UNLESS the engine is also running. Its a safety issue
in case of an accident that ruptures fuel lines- you do NOT want the
electric fuel pump continuing to pump out fuel if you're unconscious and
the car is smashed but the ignition is still on. That's what
auto-shutdown relays are for, and that's where I'd start looking for the
problem.

Well, I have another silly question...
If he turns the switch on, the pump would normally run until some minimum
acceptable pressure level were obtained, wouldn't it, even if it takes
longer
than the default time?  Is 42 psi about right for this car?

Would the fuel pressure regulator bypass to tank at this pressure?

I wouldn't think it's a separate circuit controlling the fuel pump
priming time when the key is switched to the on position.  More likely
would be the CPU in the fuel computer (which I think is still an EEC-IV
for the smaller engine Mustangs of that year, but I could be wrong as
my familiarity is pretty much limited to the V-8s) has actual programming
to turn the fuel pump on and off.  That makes more sense from a computer
hardware/software design standpoint, since the computer likely needs
fuel relay power access and inserting a one second cycle time in code
is an easy task in any programming language, even assembly.  

Agreed, although in an accident there's an inertially activated switch
at the back of the Mustang that will trip and cut power to the fuel pump
whether the engine control computer wants it turned off or not.  Still,
that problem needs fixing.

I'm not near my factory manuals to check if it's a valid theory or not,
but going on the assumption that the CPU in the engine control computer
is what's keeping the fuel relay circuit energized, maybe there's some
sort of external input feeding the computer that is triggering some
predefined condition to keep the circuit energized?  As an example I know
to be true, I have a 1993 Chevy truck that, if it is reading oil pressure
even when the engine isn't running, will turn on the fuel pump.  Seems
strange, but I found that one out the hard way once.  I had the key set
at the on position so the gauges would work, and was spinning the oil
pump with a drill to prime the system and see what pressure level was
reached.  As soon as it came up above a couple of PSI, the fuel pump
kicked in and spooged gas all over my hand, which was inconveniently
located next to the fuel line which had been disconnected from the
throttle body.  So in the case of that vehicle, erroneous data from the
oil pressure sensor could activate the fuel pump even with the engine
turned off.  I'm not saying that's the actual case in your situation,
but I'm wondering if it's a possibility.  Do you have any other strange
behavior in your car?  Bad gauge readings?  Possible shorted wires?