I have been having an awful time troubleshooting a stumble/loss of power when the car first gets going. The local Pontiac dealer is no help (he can't find the problem either) and I have replaced about everything from ECM to ignition to any and all sensors.

What happens is when the car (88 GT w/ Auto) is first started, and after some back-street (slow) driving I start to accelerate, after a few seconds it acts like it's ruinning out of gas for maybe three seconds... then it perks up and all is fine. The dealer checked fuel pressure (okay) and everything else HE could think of.

So, I had monitored the ALDL and finally discovered something that coincides with the stumble...Injector pulse width drops. Why? The ECM has just switched to "closed-loop" operation! This happenes at about 100°F coolant temp.

So, I ask: Since the car has FOCOA headers; and the O2 sensor is mounted in what I might think is a "cooler" location in that it's a long ways away from the engine and would take a longer time to heat up, could this be related to the problem... that the ECM *thinks* the sensor is warm, and operating correctly, and it really *isn't* warmed up yet?

It's possible. I'm not sure what conditions have to be met before the ecm will go into close loop. I know the coolant temp is one but don't know the temp that it does happen I know on my 88 it seems to go into close loop like your does fairly quickly.
On newer cars close loop will not happen until the ecm is seeing good signal from the O2. Maybe our earlier ecms do not check that.

I seem to remember others talking about wanting to putting in a heated O2 to resolve a similar problem.

You maybe on the right track. While in open loop you run off a set cold engine table with higher FA ratio and as soon as you go into CL would start to use the table that depends on Map/RPM and O2. Which would be leaner.

Know I have not really answered your question.

I am not an expert on this but 100F seems really low for the cpu to go into closed loop. Given that the air temperature right now is proably in the 80 a 20 degree hike seems too short.

I would look into the temperature at which the cpu should actually go into closed loop, it might have been changed by previous owners of the car

good luck

tim

I'm not sure of the temp either, but I do know when I've logged data on my 88 I noticed that it did go into close loop surprisingly quickly. I just don't know if it is due to temp. or that the O2 signal looks good already. My 88 is completely stock.

[This message has been edited by Dodgerunner (edited 07-24-2006).]

I guess the obvious , what is the 02 sensor reading when this happens?

Your problem is not the headers. Many others run them wihtout problems. It may take a little (and I mean little) more to heat but that should be minor. When you switch to CL is when the problem is evident because that's when all the sensors come into play. Of course your injector pulse will drop if the engine stumbles but that doesn't mean anything specific to them. ECM will manage injector pulse based on several sensors input. If it only ocurs on cold engine then it is temperature related. If you recently installed the headers/engine work then check all your connections. Heat makes things expand so a marginal sensor may be making contact once temp is up. ALDL could help but then also it provides so much info that it could throw you to many places. Good luck.

Usually a stumble or hesitation on accelaration would lead to the tps sensor. Other sensors, such as the coolant to ecm sensor, mat, ect , ect deal with fuel/air mix.

Thanks, folks,

A couple of quick answers...
1). the O2 sensor reads all over the place; which I understand isn't all that important, as long as the 'cross counts' keep incrementing... it can read as low as 100mV, or up to 800mV.
2). The TPS signal, MAP and temp signals read steady from ALDL
3). Block Learn stays at 128
4). Rich/Lean flag alternates at (as close as I can tell) a 50% rate
5). Headers have been on for maybe 5 years.

Incidentally, from one of my previous posts, this particular symptom is relatively new; I had a stalling problem (usually twice or more before she'd stay running) UNTIL I fixed a monkeyed-with idle screw per instructions. Then the problem 'migrated'.

The stop idle screw has been tampered with? I`ll post the proceedure below, but in expereince from past endeavous with my Fieros, it hasn`t worked fully 100% , I fixed mine by buying a used/new t.b. with the cap still installed...

If everything is working properly, no exhaust leaks, no EGR leaks, and
the engine is properly tuned, you can go through the IAC system.

Idle is controled by the IAC valve (idle air control) via the computer and
is not adjustable.

You can go over the system if your not happy with where it is idleing.
Proper idle for the V6 is roughly 900 RPM.

The set screw in the throttle body is for minimum air flow through the
throttle plate which is often misunderstood as the idle screw. Any tweaking
of the throttle stop screw will give the ECM fits. The ECM wants to be in
control of the idle and is not happy when it can't control it. If the
normal
learn limits of the ECM are exceeded, they will be reset to nominal,
causing
an erratic idle.

The IAC and associated passages need to be clean to work right. Remove the
IAC carefully. You can clean it (the nipple) using carb cleaner and a small
brass brush or rag. At this time also clean the throttle palte. Once clean,
install the IAC back in the throttle body and reconnect the IAC wires.

For the ECM to properly control idle, the throttle stop screw must be set
for "minimum air". This is a process that sets the idle with the IAC
fully
extended. To fully extend the IAC, jumper ALDL pins A and B together (just
like when you check codes) and turn the key on, but do not start the car.
With the key on, not running, and in diags mode, the ECM will keep
trying to
fully extend the IAC. After 30 seconds or so, pull the IAC connector
off the
IAC *before* doing anything else. This will capture the IAC fully extended.

Now pull out the jumper in the ALDL, and start the car. Typically the
"minimum air" idle speed is in the 500 RPM range. I find the car can
bearly
run at 550. So as long as you can get it to idle on its own between 600 &
700 your good. Set the idle using the throttle stop screw. (The engine
should be fully warm to do this.) Now shut the engine off and reconnect the
IAC wires. The ECM does not know where the IAC present position is, so pull
the ECM fuse (or disconnect the battery) for 20-30 seconds. (This will
cause
a complete ECM reset of all learned parameters, including the learned IAC
ones. Then reinstall the ECM fuse.

Turn the key on, wait 10 seconds or so, and turn the key back off. This
will
now reset the IAC to a known key-off "park" position. Now start the
car. The
engine should idle properly under control of the ECM. There are some
learned
values, such as an IAC offset for A/C, etc that need to be learned, but
this
will happen under normal driving conditions. I suggest driving the car
right
away under all conditions. Stop & go, steady cruising over 45 mph, full
throttle, and so on. Pull over a few times and turn the car off, then
restart it. The IAC can only learn X amount of counts with each run
position. If everything else on the engine is in good condition and
operating properly it should be around 900 RPM after coming to a complete
stop with slight variations and improve over time.

Thanks, superfast; those were the very instructions I used!