Since you retrieved the error codes you must have a code reader. If the code reader is relatively good, it will have a feature on it which allows you to observe both the short and long term fuel trims in real time. Long term trims are the running average of the short term trims. Unfortunately, when you cleared the codes from the ECU, you wipe the long term trim from memory so the value starts out at zero after each reset. In a perfect world at sea level on perfect gas on a stock engine that was 100% perfect the long term trim values would be around 0% and the short term trim values would be small single digit values that continually bounce between + and - values (averaging out at 0). If you can, when the engine is up to temperature and running poorly, connect up the code reader and have a look at the fuel trim values. If they are large numbers, that indicates that the ECU is trying to do a lot of fuel mixture correction indicating that you have a fuel problem. If the fuel trim values are all really small values, then the fuel mix is likely OK and I reserve my previous comments about your problem being fuel related and perhaps you need to go back to examining the ignition system. Really large long term trims will normally cause a fuel mixture error code which you did not report; however, when you reset the ECU you wipe that value and it may take a little operating time to accumulate long term trim to the point that it generates an error code. Monitoring the trims may be easier than trying to log the AFRs because all the values are available at the same time.
Some other thoughts:
I assume you are running two wideband sensors and controllers with a single display and switch between the two controller outputs? Did you retain the original pre cat narrow band sensors for use with the ECU or are you generating a synthetic narrow band signal from the widebands for use with the ECU? If the latter, check to make sure the firm ware set up for the narrowband simulation is correct (if it is configurable), although if the car ran OK at one time that is not likely the problem.
What does SOS kit use for a fuel pressure regulator? If it retains the original FPR, that is something that could go 'off' all of a sudden because of age. Relatively easy to check by doing a fuel pressure test. Make sure that the manifold pressure reference line for the FPR has not come loose. That would cause the engine to run rich at manifold pressures below 100 kPa which might be tolerable when the engine is cold; but, lead to poor operation when hot.
NSX cam / crank position sensors frequently look like crap with the potting compound oozing out of the housing and look like they should be sent to the garbage can. They generally continue to operate just fine. The sensor is a dumb-ass simple device consisting of four coils of wire wound on 4 magnetic bobbins. There are no electronic components in the device. You can test the device checking for continuity in the four coils and the resistance of the four coils (I recall the values are set out in the service manual). If you have continuity and the resistances are OK, then the sensor is OK. Besides, if the sensor failed because of a bad connection the engine will not start. Save your money unless you are offended by the site of goo drooling down the side of a nice clean engine :smile:.
Checking the injector pigtails is a good no cost thing to do; but, if the engine truly runs great cold, but, poorly once it comes up to temperature I am having a hard time visualizing the connections causing that type of problem.
I know pretty much zero about the AEM FIC. Does it have a temperature compensation for ignition advance - some engines need a little more ignition advance when cold and the controller dials back this advance as the engine warms up. If the AEM has this feature perhaps there is a little too much temperature compensation? If your tuner checked the idle ignition timing on a hot engine and it is as per the service manual, then ignore this thought.
You made the comment in your later post about running a little bit fat. I am assuming that you mean running AFRs that are significantly less than 14.7. That is OK if you are talking about doing that at wide open or close to wide open throttle. Even the OEM ECU runs less than 14.7 at large throttle openings. It gets away with this by switching from closed loop to open loop at large throttle and running directly off of the fuel map with no O2 correction. Where that occurs is probably known only to Honda and perhaps sr5guy. Main point is that you can set up the AEM to run AFRs below 14.7 as long as the ECU is running in open loop fuel control. If you try to 'fatten up' the fuel mix in those areas of the fuel map where the ECU has switched back into closed loop operation you are going to have a fight on your hands because once the ECU is in closed loop it is always trying to force the engine to operate at 14.7. When the engine is hot. if you drive down a level road at about 50 km/hr with steady throttle (no acceleration / de acceleration) you should be in closed loop reading AFRs right around 14.7 (probably moving +/- a percentage or so. If you are way off 14.7, that is confirmation that you have some kind of fuel problem or a set up conflict between the FIC and the ECU. Keep in mind that I said I know zip about the AEM FIC so ........!