UPDATE- Furhter below are all the things I tried to cure a repeated leaness. I finally know that this lean running was caused by the 302 firing order I am using, instead of a 5.0/351.
The problem occurs because the computer can adjust the injectors left/right seperately, based on the O2s. Let's say it reads the left side as lean. Now it richens those injectors. However some of them have been switched to the right side. So now the right side is rich, so it tries to lean it out, but some of those injectors are on the left side. So it keeps getting leaner and leaner. Unplugging the computer was the only way to start from scratch.
I had clues all along but didn't think it through. When I did the conversion I left off one O2 exhaust fitting, and the engine ran good. When I added the 2nd the trouble started, but I knew the O2 was good, so I looked elsewhere. I guess with only one O2 it can't ping-pong back and forth. Running one O2 may be one cure.
About a week before I was totally clued in, I had disconnected both O2s and the computer defaulted to the backup fuel charts. These are a little rich for a stock engine, but with my mods the engine ran very well at all times. It still richens up when cold, but is a little lean off-idle. I don't know about the timing.
  I made an adjustable fuel regulator and used it to fine tune the fuel, just like it was a carb. This is another fix. I backed off until it started to ping, then went richer 1/4 turn. Around town I am getting 22 mpg, so I happy for now.
  I have now experimented with joining both O2s together on the bench. They still read the normal voltages and also seem to even out the signal. I don't know if the computer can detect this or not. I'm going to adapt a second harness and test it on the car. I'll use one of those connectors to tie both O2 wires together at the harness plug.
LTD and Tbird computers have the right firing orders but use 14 lb injectors. Maybe they could adapt. I was told tweecers can change the firing order in the software.
So, the real cure is to have a cam with the 5.0/351 firing order. ( people with 5.0s who plug the O2 harness in the wrong sides also get the same problems.)

 Once my EFI conversion was completed final tuning had to be done. It ran better than a carb, but it still had some irritating quirks. Mainly it ran very lean while cruising. Full throttle was good. In fact it had more power and response than before. However, while cruising it had the hiccups, and on light acceleration it sometimes had a light backfire! If it was a carb I would richen the main jets and go to a bigger pump squirter, but how to do that with the efi?
Various timing settings were tried with no result. 2 plug wires and the coil wire had very high resistance and were replaced. ( A website said they should have under 4000 ohms per foot.) The engine ran smooother but no fix.
 Reading various forums only showed that other people had the same problems and both Speed Density and Mass Air systems could have the same symptoms. The solutions offered were suspect. Several people said that an exhaust leak caused backfiring, because fresh air was sucked in and mixed with the exhaust and was ignited. Well, if there was a leak there would be pressure at the hole, not vacuum, and also a factory air pump puts air into the exhaust and that doesn't create backfires either.
 Others said the O2s must be bad because the computer uses them to adapt to the perfect mixture. I had tested several O2s before installation and the best ones passed. One was excellent. The out of car test I used was;

1 Hook the wiring to a voltmeter.
2 Heat with a propane torch.
3 It should show .6V in under 20 seconds.
4 Remove heat, and it should show 0V in 4 seconds.
5 Return heat and it should stay steady for 2 minutes.

 Next was the TPS: Throttle Position Sensor. If it was catching then the computer may not be reading the right throttle position. I bought a new one. It was very pricey but didn't fix the problem. It did improve the response anyway.
 I began to think of Tolerance Stacking, which is when lots of little things stack up to a big hinderance to operation. I would test everything I could and fix as much as could be fixed.
 The sensors were tested according to the specs on several websites. ( I didn't pay attention to which, and now several of them seem to be down.One was http://www.nettally.com/silly34/main.htm) To make testing easier plug-ins were cut from cars at the wreckers. However all the temp sensors, the TFI, and the EGR stuff checked out okay, and several copies of each as well. This was surprising considering that it was all from the wrecking yard.
  One improvement was in the ground wires. I had added several but later found the old engine ground was very rusty. I cleaned and coated it with metallic anti-sieze.
 The problem was being eaten away around the edges, or perhaps the EEC was compensating faster.
 Next, I knew home computers can be affected by dust buildup causing static, so I cleaned the EEC with compressed air and alcohol. It seemed to help.
 I made an adapter to plug the engine temp into the air temp, and put a second one in the intake. It cured the stutter, but then it turned cold out and the sensor never warmed up, so the idle went to 1800rpm. The idea might work well in the summer.
 I had 2 voltmeters and hooked them up to watch various sensors while driving. The first problem was that my 'excellent' O2 sensor gave no signal. The wiring checked okay, but not the O2. The problem was a clear coat had plugged the vent to the outside air. Oops. I dremelled a new opening in the side and it worked great again. Now the EEC could read it as well as myself. This in itself helped.
  While cold the O2 read about .8v, which is slightly rich, but when hot it was about .3v, which is very lean! On light throttle it went to .1v! I was right after all.
 As the EGR opens it sends a growing voltage signal. It opens in the midrange and in cruise. I found that after flooring it the O2 would read .8v and the egr would stay off for up to a minute. However there was also a point in light acceleration that the egr would shut off, but the O2 stayed lean. So this meant that the unburnt fuel in the egr flow was removed but the 'accelerator pump' didn't kick in, so it was leaner right when you wanted it to be richer! The computer wasn't coordinating things properly, even if it thought it was.
 Next was the alternator. Modern alternators with internal voltage regulators use a wire to test battery voltage to self regulate. However, this wire can be placed in different spots. I knew this before but was reminded by http://www.madelectrical.com/ . I tested the voltage at the wire to the EEC and injectors. It was 13.7V with the car running instead of the preferred 14.5. The sensing wire was moved to the EEC line, and the voltage went up.
 This finally pushed it over the hump. The engine only had a very slight stutter now.

 It is now running like a fuel injection should. Up next would be a look at richening the mixture slightly by putting a 2000 to 3000 ohm resistor in the engine temp circuit or altering the fuel pressure. The site http://www.pro-flow.com/techinfo/tech_support.htm states that speed density can be fooled by altered fuel pressure, while mass air will adapt. I'll leave it for now though. I'm sick of testing wires and purging the computer.

UPDATE   It took about a month but the engine slowly crept up on leaness again as the weather became colder. There was some surging and backfire pops during light acceleration again. I plugged the ECT harness back into the intake sensor and it was instantly cured. However the idle also went back to 1600 rpms. I also noted from the voltmeters that the EGR didn't open anymore, or only very slightly. So, it appears that the EGR causes, or increases, lean running. If the ECT temps could be altered just right it should fix the problem. Maybe.
A forum thread revealed the fact that the Idle Air Controller lets air in at all rpms. I assumed it would only be at idle. I unplugged the IAC and it then cruised fine. In fact it seemed a little rich. However there was now no cold start rev-up and it stalled out. The answer might be to sleeve down the air entrance into the IAC.
 I closed the small air bleed in the throttle body with brass and epoxy and it helped. The idle screw had to be turned in one turn and the voltage reset.

Somewhere in the middle of all this I pulled the vacuum to the fuel regulator as a test. Leaness solved! I was laying rubber everywhere, for a while. Then the borderline clutch gave up. I could idle around but that's it.
When the 10 inch disc came out it was torn to shreds, so I thought about upgrading. I knew the 86 EFI Mustang still used a 10 inch, but the 10 1/2 from the 87up seemed to be the better idea. The problem was my older engine had a different balance. Forum sites said 'any shop can rebalance a flywheel' but when I phoned around no one knew what I was talking about. Finally one shop did, and did it often. So, I bought a used flywheel and had them balance it. Too bad for me that they doubled their quote, and with other expenses, it cost more than buying a new steel flywheel from the aftermarket!
At least they balanced it right. The replacement Valeo clutch works better than the old one ever did.

UPDATE 2   I finally put in an adjustable fuel pressure regulator and the leaness seems to be cured. I built my own from a CFI regulator that sits on top of the throttle body. It has an adjuster that is covered with a plug at the factory. Pop out the plug and turn the socket head bolt.
 This regulator also needs to have a vacuum line added. A hole can be drilled on the edge or the side and a tube soldered in. Remember that there is a spring under the cover.
 After installing it I turned the adjuster about 1/8 of a turn per day until it cruised without jerkiness.

 It is a long learning curve but now I know a few things. The big change is that with carbs the thinking is mechanical while EFI has an electric heart. It assumes certain things and makes decisions based on that. Kind of like us.