Geschrieben von Paul Axford aus Neuseeland. Thanks, Paul.
The AFM is a 80's relic from the first generation of electronic fuel injection. This massive sensor sitting dead-center on top of the engine does two things - measure the flow rate and temperature of the incoming air. The flow rate is determined by measuring the angle of a spring-loaded vane in the path of the air flow. There is a potentiometer in the base which translates that angle to a voltage divider. As primitive as it is, this is a very pricey item so don't break it.
I checked the operation of the AFM by pushing the vane open while still in the car (measure 0 - 1000 ohms between pins 2 and 4, counting from the nearby rubber mount.) As the output was not as stable as I would have expected, I removed the AFM and carefully removed the base cover - not an easy task as it's firmly glued in place with silicone rubber. (Be aware, if you remove the screws around the connector you will damage the AFM.) After removing the cover I cleaned off all the loose RTV from both sides of the joint.
The potentiometer wiper contacts and track clearly needed to be cleaned. I used a small sliver of paper wetted with isopropyl alcohol, slid under the wiper. I then cleaned the track with alcohol-wetted tissue paper. I didn't see any point in tampering with the wiper spring force or mess with the calibration, as other people have done.
Next, I shifted the track base very slightly by prying at the edges with a screwdriver. The idea was to move the wiper tracking path to a fresh area. If I could have gotten the three screws loose I might have considered drilling out the holes slightly to gain a bit more movement. Even after 140,000 km of operation the track was not badly worn. A quick test with the ohm meter showed that the output was now much more stable.
As the potentiometer housing appears to be completely sealed it clearly needs to contain a dry environment otherwise moisture will condense out and corrode the electrical parts. I procured a small bag of silica gel from a local camera shop, dried it in the oven (150° C for 30 min) and jammed it in the hole indicated in the photo below before re-installing the cover with fresh RTV silicone rubber adhesive.
As a side note, my AFM had evidence of design revisions, as clearly the intent of the original designer did not work. (He must be related to the cylinder head guy.) Note the wire jumpers that parallel spring loaded contacts from the connector base to the potentiometer wiper and track. Clearly the contacts did not work well after a bit of corrosion took place. This is why including the silica gel desiccant is so important.
