Air Flow Sensor behavior
 

In the ongoing saga of my 83's very rough idle and rich running investigation , I thought I might have narrowed it to the AFS. I re-tested the AFS and I noticed that if I unplug the AFS it will purr like a kitten for about 5 secs then die. Twice I have tested the AFS per Clark's and even the internal temp sensor I suspected checks out per spec as do all the other tests. I even re-tracked it , and it looks brand new inside. It is getting 10.8 volts from the DME which according to Clark's is ok too.

So even though all the AFS tests are in spec, is it normal ops for it to run better with it unplugged or does this indicate that I do have a bad AFS ?

Well, when the DME loses the return signal from the AFM, it injects next to no fuel, so when you unplugged it, the motor probably ran on the extra fuel wetting down the ports and combustion chambers from running too rich. Does it still run badly if you run it with the o2 sensor unplugged? How is the fuel pressure? You may want to try some seafoam or BG44K in the fuel to clean possibly dirty injectors and restore the correct fine mist spray or to unstick a stuck open or leaking injector. Does it run too rich all the time or only when warm or cold, possibly a bad DME coolant temp sensor. You may want to test things out at the DME connector to see if any bad wiring is causing problems. If the DME coolant temp sensor or air temp sensor voltage is low by the time it gets to the DME, it will signal for more fuel than is needed since both sensors read higher resistance with decreasing temperature. With the AFM, higher air flow means larger flap angle, more voltage returned to the DME, and consequently more fuel injected. Usually problems with the AFM result in bad contacts causing undervolting of the return signal and a lean mixture, sometimes in only in spots on the power delivery.

Does the early 944 have any sort of cold start valve? I can't remember.

DME Connector 35-pin

AFM and Open loop/ limp mode, bad sensor
 

:eek:If the AFM return is out of range Typical at idle ~0.40 to 0.48v or intermittent > 2sec. the DME will go into open loop / limp mode. It will inject base fuel and timing programmed from the eprom.

I posted an AFM on ebay 4 Sale check it out...

I hope this reply was helpful.

Later,

Just want to clarify the DME models a little bit:

85-89 2.5 & 2.7L NA's 24pin/28pin Eprom Motronic DME ML 1.2 LM Single knock sensor
87-90 2.5, 3.0L NA 16v S, S2 28pin Eprom Motronic ML.2.1 LM Dual knock sensors
86-90 2.5 Turbo /Turbo S 24pin/28pin Eprom Motronic ML 3.1 LM Single Knock Sensor
91-95 3.0L 16v Vario-Cam 28pin Eprom Motronic ML 2.2 LH Dual knock sensors

Just be extra careful....

I hope this post was helpful, Later...

2.5L 8v motors do not have knock sensors. Only on the 16v and the turbo motors and there are two of them, IIRC.

Thanks for the replies guys - that's more than I got from several posts on Rennlist. Ok, a little background....
It runs rich all the time. Starts great ,but hardly idles without help and smokes .
After 15 mins it may run only a fraction better. Had to back idle screw out 4 turns for it to idle without my help.

Before I bought the car recently the PO had replaced about 1 yr prior:
-fuel pump
-fuel filter
- new Beru wires
-new plugs
- new injectors
- new dist cap, rotor
- new AFS

I replaced:
- battery
- all vac hoses and j-boot
- intake gaskets
- new timing belts
- new coil
- Aux Air Valve (like the cold start valve that someone mentioned)
- both ref sensors
- DME (tried a spare)
- plugs
- DME temp sensor

Tested good according to Clark's tests:
- TPS
- throttle body (looks new)
- AFS (looks new) I re-tracked anyway
- plug wires
- all engine bay grounds
- DME harness
- fuel pressure at key on, idle , and leak down
- FPR no leaks or fuel smell
- injector leads with noid light and ohm meter
- compression equal and good, 169
- removed cat to see if plugged

o2 sensor seemed to be the same plugged in or not.
I might try more injector cleaner , add the redundant wire on the AFS arm, and check the AFS input voltage at the DME again. The chance of two bad DMEs are small but I am trying to find someone that has a know good one as my spare was condition unknown.

In the 3 mths I have owned this , I have yet to feel what a 944 really drives or sounds like.
Thanks

Wow...that's a big list.

Yeah, add the extra ground strap on the AFM. Is there any difference in AFM feed and return voltages to indicate bad wires when tested at their sources and then at the ends of the wiring harness If you catch my meaning.

If it were me, I would pull the fuel rail off with the fuel injectors still attatched and run the fuel pump to check for leaking injectors. Then crank the motor to visually check for good spray patterns. Disconnect the primary voltage supply at the igniton coil before doing this.

Have you checked the connection to the DME & block temp sensors (located on the block directly under the throttle body). I'd also recommend testing the resistance of the sensors themselves. I had a similar issue and found that the DME and one wire temp sensors in the block were very corroded. I ended up cleaning up the connectors for both and replaced the single wire connector. The car ran fantastic after.

See excerpt from Volume 1 of the 944 manual.

http://mysite.verizon.net/vzeylh08/s...tures/RIch.jpg

Here is a better picture of the block temperature sensors (linked to Clarks-garage) that I've seen corrosion issues with. Your computer uses these to adjust your mixture.

http://www.clarks-garage.com/graphic...-locations.jpg

Are you sure you don't mean the DME temp sensor would cause the issue? My DME sensor was corroded when I replaced it last month ,but I was not aware the regular temp sensor would come in play here. I thought it only drove the gauge ? I tried to remove it but that bugger is in there, so I will have to remove the j-hose and t-body to get a better approach. I spent most my time trying to verify the resistance at the AFS plug is the same a DME harness.

Here is my confusion. When I test the sensor pins 7,9, or 22 using pin 6 (the ground) all is well. When I tested the corresponding AFS pins on the DME side I got nothing. I thought I was on to something. But then I decided to test the sensor plug using a body ground instead of pin 6. When I did this , even using the battery ground itself, I got nothing ! How can this be, where is the sensor getting it's ground for pin 6 especially when it is not plugged into the DME ? So until I can swap a ground point at the sensor side I do not feel my failure to get a ohm reading at the DME plug is valid yet. For ground on the DME side I tried all pins listed as grounds as well as the door bolt that has worked as a good ground for other tests.

Pin 6 is the ground specifically for the AFM and is not linked to any other ground, I believe. The AFM is isolated from any normal point of grounding and so must have it's own in the wiring harness, although I believe it's only needed for the air temp sensor. If you get continuity between pins 22 and 6 at the DME connector with the AFM plugged in, then both the ground wire and the air temp sensor signal wire in the harness is working. The resistance value you measure will be the air temp sensor resistance plus whatever the extra resistance is from the connectors, wires, etc. Measure between pins 6 and 22 on the AFM and compare the two to ensure the wiring isn't bad and the sensor is within spec.

The way to test the AFM flap signals and the effects on them from the 7 and 9 wires is a little trickier since you must measure voltage and not resistance. You should technically be able to get resistance values between 6 and 9, 6 and 7, and 7 and 9, but they are meaningless. The voltage leaves the DME and goes to the AFM through wire 9 and the adjusted signal returns to the DME through wire 7. The DME must be plugged in in order to get signal readings since it is the source of the supply voltage for the AFM flap sensor.

The DME connector does come apart. There is a little screw that holds the back housing on. Remove the screw, pull back the rubber boot, remove the cable tie holding the two halves together if there is one, and the cover will slide off. Plug the DME connector into the DME, stick test leads througn the back of the connector block at terminals 6 and 9, turn the key on making sure the AFM is plugged in, and note the voltage (8 - battery voltage on early, 5v on late). Move the test lead on pin 9 on the DME to pin 9 on the AFM and compare readings. There shouldn't be any significant voltage drop between the AFM readings and the DME connector readings. Next, test the return voltage at the AFM, pin 7, flap wide open, with the other test lead still gounded at 6 in the DME plug. Then move the test lead from pin 7 on the AFM to pin 7 on the DME (flap still open) and compare readings. No significant voltage drop at pin 7 at the DME compared to the AFM is a good thing.

Once all of that check out, connect a voltmeter between pins 6 and 7 on the DME connector with the DME and AFM still plugged in and the key on and slowly move the flap open and closed. This is the final signal voltage the DME is seeing from the AFM and should respond smoothly and increase steadily for increasing flap openning angle. An analog voltmeter is useful for this if one is available.

HondaDustR, thanks as that was great info that is not directly apparent in the manual. I like this method better than with the AFS off the car. I was not aware the connector came apart so that made it easy to test with the DME plugged in.

When comparing the resistance on 6 and 22 , I got the same (1.8) at the sensor and DME , so air temp sensor seems good.

Input voltage: pins 6 and 9, I got 9.6v on both ends. Battery was somewhat drained so this still seems ok.

Output voltage: pins 6 and 7, flap 100% I got 9.6v at the sensor and 8.67v at the DME. Flap at 0% measured .481v. Not sure if 8.67v is too much a delta or not.

It appears the hunt may still be on.

With the test leads at pin 6 on the DME and pin 7 on the AFM, where you said you measured 9.6v flap fully open, without moving the leads, do you get decreasing voltage at the same point when the flap closes, just to confirm that you are measuring the AFM output and not the input from the DME? If that is the case and you still get ~1v less at the same pin 7 on the DME connector, then there must be a problem somewhere in that #7 wire, or in the AFM connector. At least on the late, it is normal to get slightly less coming out than going in at 100% open, but that "slightly less" should measure the same both at the AFM and at the DME. It's always good to double check, as keeping track of what's what with this thing is a brainful...I had to read over my post several times to try to make sure I wasn't getting mixed up myself. I believe you can pull back the rubber on the AFM to expose the wire connections to the connector and be able to test before and after the connector to see if the connector is bad at pin 7. Also check for possibly a cracked connection internal to the AFM between the circuit board and the connector pins, depending on exactly where you measured 9.6v at the AFM.

On a side note, an interesting way to know if the AFM is a source of a running problem is to go WOT while driving. If the engine clears up and runs normally, then the AFM could definitely be a problem. WOT sends the DME into open loop mode off of seperate 1 dimensional fuel maps and disregards the AFM signal. I guess the Porsche engineers figured they already know the maximum volume the motor can pull in at any given rpm with open throttle.

Well, you were dead on about keeping track of what is what. When I rechecked tonight the voltage at the sensor and DME were the same so who knows what I did last night although I swear I tested methodically. This car is really trying to kill me. All connectors seem fine and show the same as probing the wires. I did get decreasing volts when closing the flap so good there too. As for running it wide open ...since it runs like it is on 3 cylinders it is not very easy to drive.

In trying to step back and think, the only test I have not done is pull the fuel rail to look at the injector flow. Since a prior shop that serviced the car installed new injectors in hopes of curing this problem , I left that one out. I did test fuel flow , FPR , etc. Since I am rich I guess a stuck injector could cause this but I applied voltage to each injector and could hear a click when it fired so I don't think this could be it. I also confirmed signal to each from DME with noid lights. Maybe a retest of the O2 but it runs the same unplugged or not.

Hmmm...running out of ideas. Did you try unplugging one injector at a time while running, and same thing with the plugs, as extra confirmation they are all working. If anything doesn't make it run worse, or not much worse, that would help pinpoint the problem. Running one cylinder with a working injector and no spark will make it seem to run rich as well as very rough.

Is the FQS (fuel quality switch) in the DME set all the way counterclockwise to the default setting? Does it change for better or worse when the A/C is turned on or the cooling fans come on? What 12v supply voltages do you get at the DME connector when the car is running?

If you haven't pulled the fuel rail, it would also be a good thing to check, maybe even jumper the fuel fump on and manually pulse each injector to check the spray and that it instantly pulses on and off with no sticking or leaking.

Ok, I checked all of this again and I am throwing the towel. I hate to do it but I plan on taking to a indy Porsche shop and seeing what they can do in 1-2 hrs.