Ok this will be long, but I want it to be detailed, hopefully that helps
someone recognize this problem.

'96 F150  5.0  dual tanks, xlt supercab, 235000 miles
owned since Feb, problems sinc Apr.

I bought this trunk from a friend of a friend who was moving and couldn't
take it with him.  He didn't want to put it in the paper because of the
issues it had.

When I bought it, both fuel tanks worked, but the rear fuel pump was a
little iffy.(it's dead now)  You could switch once it was running, but it
usually wouldn't start on the rear tank. After a couple weeks, it died
altogether.  It made a couple horrible groaning sounds before it went and
hasn't run on the rear tank since.

He thought it needed a fuel filter, and that was why it wouldn't start on
the back tank.  He gave me the filter and the tool with the truck, but he
wasn't able to change it.  I changed the filter since, and that didn't do
it.  I have the schematics for the pump, but havent had a chance to test it
since I got them.

Besides the back fuel pump going, the truck was great for about 3 months.
Then it mysteriously died as I was getting off the highway, from a 20minute
drive.  It was very sneaky, it ran right until I stopped at the light, then
stalled.  It wouldn't start back up for several minutes, but finaly after
about ten, it started like I had only imagined it, and there was really
nothing wrong at all.  It continued doing this for several months.  I
eventually had to replace the 5 year old battery b/c it couldn't take the
stress.  I also replaced the battery to starter cable at the same time.  It
was pretty ate up from road salt.

It's gotten to where it does it nearly every trip, which on the positive
side, gives me lots of chances to try to isolate it.  I tried pulling fuses
and relays trying to get it to reset early, and eventually stumbled onto the
fact that if I pulled the ECU relay at around 1-2 minutes it would start
right up (most of the time) cutting the down time from 10 minutes to 2-3.
This was a big step.

I finally made it around to changing the fuel filter about a month ago.  I
was excited the first day after.  I made it all the way to work without it
dieing on me.  It couldn't last too long though, on the way home it stalled
again.  Same thing.

I'm sure someone is dying to know.  It did throw a couple codes.  Few and
far between though.  The first one was EGR low flow (402?).  Which after
discussing and reading up on, I decided was probably not the real code.
That tends to cause detonation at WOT.  which it did too, but was not the
main problem.  and I didn't read anything else about EGR causing problems
similar to this

It seemed obvious to me about then that the engine was a little neglected
about then.

So after that I bought a little spark tester light and left plugged into the
distributor, when it was running, it stayed lit.  Whenever it died, no
light.  Voilla, ignition system problem right.  But where at.

In the last week  it threw codes p0320 and p 1115?, engine speed sensor
(pip) and low bank2 blah blah  O2 sensor error. Which was exciting again.

I took off the distributor cap, cleaned and inspected the pip sensor.  Which
looked fine, a little dusty though. Cleaned it.  Put in a new rotor, cap,
and plug wires (the plugs were replace immediately after buying it.)  The
pip seems like it should be pretty bulletproof, the magnet was in fine shape
and the window is just a piece of metal.  No scratches or marks.  The next
day the truck felt great.  There was noticeably more power, 15 to 20HP by
the seat of my pants.  It was great, until it stalled.

Since I had read up on the TFIV ignition system, I took voltage readings at
the PIP/distributor connecter, ICM, and coil, with it in stalled mode, and
all had power available. I took resistance reading across the pri/sec coil
and they are fine.  I tend to think a coil is more of a good/bad thing, the
resistance will not change back and forth.  The ICM, mostly a high speed
switch could be a different story.  Once a switch starts to go, it usually
doesn't last too long though.  Which leaves the PIP sensor and the ECM.

Reading through the book, the overview is:
the PIP(hall effect sensor) distributor sends the PIP (position) signal (or
spark source) to the IgnitionConrolModule and the EmissionControlModule
(main computer).  In start mode the ICM takes the raw PIP and uses it to
generate the spark signal, in run mode it takes the signal pumped through
the ECM which is called spout.  This is how the computer controls spark
advance.  The ICM is supposed to failsafe to the PIP input if the ECM's
spout signal is lost.  Which should mean that even if the computer was
messing up the truck would run, just poorly (at least above 2500rpm.)

Now we are to the 'ewly'.  If you've never heard of ewly, or oulie, it's a
stumper question that people make up to make sure no one gets a perfect
score and test or quiz.  Sometimes after I arrive, if I try to start the
truck back up after half hour or so, it acts like its stalled again, ie no
ignition system.  Being clever, I thought I would pull the 'spout' jumper
from between the ICM and ECM which forces the ICM to run on the raw PIP
signal.  I did, and the truck started right up.  So far two of three times.
I left the connector out, and it didn't stall all the way home.  Which is a
given that it will.  The truck runs like crap above 2500 rpm and doesn't
shift right, but it doesn't stall on the highway.  Which is a pretty good
trade off.

So,
A quick review.  It stalls on the highway mostly, sometimes getting off.
There is no spark when it stalls. Pulling the ECM fuse will reset, although
it sometimes takes a couple times.  Pulling the spout connector plug lets it
start right back up too. I did notice that the fuel pump reprimes when I
pull the spout plug with keyonEngineOff (stalled), which is odd to me. It
doesn't seem like this should happen per the schematics.

I'm sure its the ignition system. But I'm not sure where.  Working
backwards, the plugs are good, the wires are good, the cap is good, the
rotor is good, the coil is good.

That leaves the PIP sensor, the ICM, and the ECM.  I have avoided testing
the PIP output because it is intermittant and don't expect to be able to see
it.  I know not the best technique, but realistic.

If anyone has any ideas or has seen this, please let know.  Thanks.