GM TBI idle to 2000 rpm instabiel a runs lean

[Subic Jeep]

Vortec 350 V8 1997-2002

I having some problems with the TBI from idle to 2000 rpm range above the 2000 rpm she runs good AFR 14.7 but under 2000 rpm the she runs lean and wants to overheat. (Haunting revolutions +/- 1-200 rpm wildly fluctuating AFR 15-20) Cleaned the IAC motor and set the minimum idle speed and correct voltage on the TPS. In the drive way she was running good but the IAC count is still high. Observing the injectors spray pattern when the haunting is going on there is a visible lack of fuel when accelerating the spray pattern is directly in relation to bad running. After the last adjustment of IAC the injector spray pattern increased dramatically and followed the RPM increase. Amazing how much fuel the injectors are capable of spraying.

Run to the city about 7 km and she was running fine. On the return trip she started haunting again and hesitates on acceleration. But at home she run normal again the idle AFR 16.5-17. Looking at the injectors they do not come up to full spray pattern when accelerating as she did the other day.

Fuel filter is ok checked

TBI pressure regulator is ok checked all new filters and gaskets.

Fuel pump sounds normal. Feed 12V directly from battery no change.

IAC cleaned and adjusted (old unit).

My theory, IAC motor giving too much air causing lean run, that is AFR 16.5-17.0 possible hanging or slow reaction

As she is running good above 2000 rpm the fuel pump must be giving enough pressure.

Any Ideas?

[Subic Jeep]

Exhaust air leak

Found intermittent air leak in the exhaust manifold on drivers side two loose bolts, causing the O2 sensor to get a lean reading at idle. The loose bolts where covered by the AC compressor brackets, difficult to reach and I missed them when checking the bolts last time, the engine was hot and could not tighten them and then forgot to do it when the engine was cooled don. The AC compressor is mounted to the brackets on the exhaust manifold. No visible or sound from this leak probably the manifold got slightly wrapped when fully heated, assisted by the pull from the AC compressor.