[moved from Six Cylinder Forum]

I know what a 100% failed MAP sensor does: black smoke and raw fuel out the exhaust, massive fuel consumption, might not even idle (just unplug it to see). But does a MAP sensor gradually or partially fail?

I had at least four isolated instances when the full-failure scenario above happened, and the car was undriveable, but the problem would go away after shutting it down for a while and restarting it. Suspecting a dodgy MAP sensor, I replaced it today with a brand new one. The car has suddenly gone very rich - dropping AFR by 2.5 (eg, from 13.5 to 11.0 at idle) - and running like a pig.

The 4.0 six is modified a fair bit (hotter cam, head work, higher compression, headers, oversize throttle body, etc), and I have been using the TI Performance J3 chip and full-house tuning pack to tune it. Also, I run a wide-band O2 sensor - for monitoring and logging - so never run closed loop. The ECU is always running from the fuel maps on the chip, without O2 sensor feedback, so should only be doing what the chip is commanding. Without the narrowband O2 feedback, it cannot learn around a change in the engine - I have to manually re-tune it. This is not a problem at all if you take the time to tune it properly after each modification - which I do.

I have tuned the chip many times. I noticed that it was gradually tending lean over time, for no apparent reason, but chalked it up to all the mods. I ended up maxing out the fuel injectors at WOT over 5000 rpm - no amount of commanded fuel would stop it leaning out - so had to upgrade the injectors. Now I'm wondering if the MAP sensor could have been failing gradually all along and that's what was causing the car to run lean. Is that even possible?

To get the car to run properly and to get the AFRs right with the new MAP sensor, I will be pulling 20-25% of the fuel out of the tune. My concern is: Is the new MAP sensor defective, and I'm tuning the engine around a dud sensor? Or has my old sensor been gradually failing for a long time, and I'm OK re-tuning it to the new sensor? Or are production tolerances so wide that my new sensor is that much different, and the engine would ordinarily just learn around it (or be need to be reset)?

I would really appreciate any help you guys and gals can provide.

MAP sensors seem to fail rarely I've found.
Try unplugging the J3 chip and refitting it.

Mine gives me the issue the car will run like crap - do the above and it comes good.
Probably happens once or twice a year.

They seem to be very sensitive with the connection - the ECU port isn't the best quality.
The pin pitch is exactly same as a computer floppy disc drive, so I've been thinking soldering
the ribbon cable to the ECU port and plug the J3 chip into the connector - with gold plated sockets
the connection shouldn't play up.

Yep, have had that happen before. But that's not the problem here. I have removed and replaced the chip well over 100 times (for tuning) - at least five times in the past week.

I am trying to figure out why swapping in a new MAP sensor would cause the engine to go instantly 20-25% rich. Is it because the old sensor had been gradually drifting lean over months or years, and the new one has brought it back into line? Or is it because the new sensor is faulty?

I was thinking the chip had been pulled a few times.
What fault codes are coming up ?
I find with mine with the chip in you get 511 (EEC fault) as default.
I suppose the best way to check out those MAP sensors is find someone with an EA-L or XG/XH
and swap them out.