[Gmecm] IAC control??

[David Allen]

  What kind of voltage do you have when this problem is showing up? Is the
alternator taking a long time to excite?
  Extremely low voltage: ECM will hold the IAC at its current position.  Not
enough voltage to move the stepper motor safely without letting it get out
of step.
  A little bit low voltage: ECM will increase idle speed a few RPM to make
sure the alternator is doing all it can to restore the battery. Not all the
way to 2000 RPM however.
  My 86 Century with a 2.8 had a similar problem.  It would idle at a very
high speed from time to time. (read- ripped out motor mounts going into
gear!)  I could do just like you said and extend the pintle, start the
engine, then reconnect the IAC and it would run fine.  The IAC would move
with the engine off but when it was working against vacuum it would stick.
Then it would be out of step with how many counts the ECM thought it has
extended.  The ECM takes for granted that the pintle actually moved when it
steps the IAC.  If the IAC is sticking even every once in a while, the ECM
will loose track of where it really is.
  Turned out I had an old-style unreliable IAC motor. Replaced it and
haven't had any trouble since.

> Also, the
> temperature GUAGE works fine, but the temp indicator
> light is always on.  If the ECM uses the same
> (faulty?) sender as the indicator light, shouldn't it
> pull a code?

   If the temp sensor is bad, my exprience has been it will cause the engine
to run "pig-rich" if it runs at all - and it may or may not set a code.  The
ECM only sets a code if it thinks the sensor is shorted or open-circuit.
The sensor can give a wildly false reading without setting a code.
  While charging the air conditioner on my car, the engine gradually revved
up to about 1200 RPM then faultered, smoked black, and died.  Wouldn't
re-start and had white gasoline vapor out the exhaust.  No codes.  Guessing,
I unplugged the temp sensor and the pins were all green and corroded.  I
left the temp sensor unplugged and after several seconds of cranking in
"clear flood" mode the engine started. At first it smoked horribly, then ran
fine.  A new temp sensor (and connector) fixed the problem.
   Some ECMs control the overheat lamp based on the reported value from the
temp sensor.  Look at your wiring schematic- the overheat lamp will either
go to the ECM or to a temperature switch on the engine.
  They all have a temp sensor for the ECM (2-wire with sealed metripack
connector). If you have a gauge it will have a temp sender of its own. If
the ECM does not control the overheat warning lamp, it will have its own
temp switch.   Sometime the gauge sender and overheat lamp switch are in the
same housing which looks like a 2-wire sensor.

As a general rule to tell them apart, measure resistance between the
terminals, and between each terminal and ground (the body of the sensor).
The following apply to undamaged parts at room temperature.  Resistance
values vary...

ECM coolant temp sensor: Resistance between two terminals and no continuity
from any terminal to ground. Always has sealed connector.
Gauge sender: Resistance from terminal to ground.
Warning lamp switch: No continuity (unless switch is heated above warning
threshold)
Dual sender/ switch: No continuity between terminals; one terminal has no
continuity to ground (lamp); other terminal has resistance to ground
(gauge).

Have fun!
David