Cold Start Stalling and/or Idle Speed "Hunting" In 16V Engines

This idle hunt at start up deal is usually NOT the AIC valve. If you have a car that isn't doing it and one that is, try swapping the air mass meters if they have the same Bosch numbers and see if the problem doesn't go away. This is assuming that the car has no vacuum leaks, has a clean AIC and throttle plate and that the base idle is adjusted properly.

The ECU depends heavily on the AMM until the oxygen sensor comes on line, taking a few minutes to reach its operating temp of 600&. I believe the thermistor in the AMM for ambient temperature is what causes the cold start blues, but I have found no conclusive test for it on a bad meter. Factory manuals list pin tests for power to the meter, a ground check, a burn off function check, and continuity to the LH ECU pins, but give no test that pinpoints this symptom. I have compared readings on meters I know where causing the problem to ones that did not have the problem. It seems that the feedback voltage on the bad ones is close to 2.0 V DC while the good ones show 3.0 V DC, but I cannot use this test for certainty. Sometimes a bad AMM will differ from a good one only by 0.2V, and that's cutting it too close to call just from a voltage reading.

Substitution is a easy test. If there are NO  vacuum leaks and the ignition timing and base idle are good, the AMM usually fixes the cold start hunters and stallers. After the meter gets worse, it will run rich, lean, cause black smoke or spark knocking, or cause hesitation and low power. Hard to distinguish between a bad AMM and oxygen sensor when the car exhibits symptoms at operating temperature without testing. I test when I have these problems more to eliminate the oxygen sensor than to condemn the air mass meter. Before beginning any tests for this problem, check ALL the basics first. Make sure the fuel filter is reasonably new. Make sure you have good fuel. Spark plug gap should be within specs. Check the secondary ignition; plug wires, distributor cap, and rotor. Clean the throttle plate and AIC valve shutter wheel. Set the base idle. Check the throttle position sensor at closed throttle. Check the ignition timing. On a warm engine it should be 14 degrees before top dead center (BTDC) on non turbos and 16 degrees BTDC on turbos at idle (850 rpm).

Look for any and all possible vacuum leaks. Cracked vacuum lines are common on older cars. On turbos, the boost can blow weak lines off their nipples. Check the intake tubes for cracks in the rubber or in the plastic tubes of turbos. Even the aluminum tubes on turbos can't be overlooked, the overheat switch in the top radiator hose has clamps which can wear a hole in the aluminum pipe from the turbo to the intercooler. If you find this, loosen the clamp on the top hose and reposition it so it has clearance from the tube. Epoxy will patch the damaged tube. A piece of old radiator hose will make a good insulator on the tube to prevent any future rubbing. Locate the vacuum line going to the charcoal canister's vacuum actuator ( on top of the canister) and make sure the actuator holds vacuum. This is a common point for a leak. On turbos, check the turbo bypass valve (hooter valve). It lives in the intake tube from the air mass meter (air filter) to the turbo. Its vacuum line when disconnected from the intake nipple should hold vacuum. If the valve doesn't hold vacuum, you may also be experiencing stalling when decelerating after being in boost. Check the vacuum advance on the distributor. Though this is ported vacuum, I find that when the diaphragm there ruptures, the advance in the distributor can release further and cause incorrect base timing.

An oxygen sensor will read between .2 and .8 V DC at idle when the car is at operating temp ( after the cooling fan cycles a time or two) if all is well. It will average just under .5V DC and should cross .5V DC at least 8X per minute. You take these readings on its single wire connector with a DVOM. Peel back the boot on the single wire connector where the connection to the engine harness is and connect the positive lead of the meter, but leave the sensor and the harness CONNECTED, just jump the DVOM lead to it. Attach the negative lead of the DVOM to a good ground, nothing better for this than the negative battery cable. If you disconnect the single wire lead, the computer no longer receives a signal from the sensor and falls back to a allowed voltage of .5V for the sensor. This may lead you to believe from the readings out of the sensor that the sensor is bad when it is not. I have been told at least one after-market manual lists a test for the sensor calling for unplugging it and testing straight off the disconnected lead. This will give false readings and is not indicative of the actual conditions when the sensor is plugged up and feeding back to the computer the mixture status. It will usually give you a idea if the sensor is alive at all, but will not reflect the state of the true air/fuel mixture in the engine.

Check the preheater circuit to the sensor. Near the same place the single wire lead from the sensor joins the harness, the preheater circuit for it does as well. Disconnect it and with the engine running, check for battery voltage across its leads from the harness side, not the sensor side. If there is no voltage there, check the fuse for the sensor. On early 900s it is in a inline fuse holder on the right side fender, near the evaporator box. Later 900s have it in the fuse box. Early 9000s have it under the left plastic cover near the windshield under the wiper transmission. Later 9000s have it in the fuse box as well. The inline fuses are the ceramic type. The fuses in the boxes are blade fuses. You can check the heater circuit in the sensor by unplugging the two wire lead to it and measuring across the pins on the sensor side of the harness with an ohm meter. Look for continuity in the 8 ohm range for a good circuit. If the preheater circuit is faulty, it may take the sensor longer to reach its operating temperature and could cause it to foul easier. If you have had head gasket problems where coolant went into the exhaust, the silicates from the coolant could have coated and contaminated the sensor. A engine that has run rich for a long time may also have a contaminated sensor, fouled from the excess fuel just like a spark plug. An engine that uses oil can foul the sensor the same way. If the sensor becomes fouled, there is usually no fix for it other than replacing it.

If you are getting a reading that fluctuates out of the sensor, pop off a vacuum line and see if the reading out of the sensor changes, it should drop closer to .2V DC. Revving the engine and then letting the throttle shut should bring the reading up toward the .8V DC side. Removing and plugging the vacuum line to the fuel pressure regulator will cause an increase in the voltage as well as the engine is initially driven richer by the increased fuel pressure until the ECU reacts to it and leans the mix. If the sensor responds to these two tests, it is probably ok and that leaves either the air mass meter or the ECU to blame, if you are certain there are no vacuum leaks. If the problem seems to be nothing more than an incorrect mixture, the AMM is more likely to be the culprit than the ECU.

 PDF File of factory manual pages describing 2.2 LH CO adjustment. Note that 380 Ohms is NOT the preferred setting, but simply a baseline from where to begin to adjust to the proper level. Without the proper equipment, the CO screw on the AMM is best left alone. Some manuals erroneously state 380 ohms is the set point. Refer to the PDF files to see the error in this assumption.

RETURN TO TECHNICAL INDEX