3.2 AFM Inspection and re-set

[ant7]

My car is a 1985 white gold 3.2 Carrera, with around 103,000 on the clock, I have owned it for over 10 years now, when I purchased it, it wasn't in the best of condition, and needed quite a bit of mechanical repairs and fettling, which I have carried out over that time, the latest item on my list for examination was the Air flow meter [AFM] which I know from experience can give rise to bad running and poor idle etc. I wouldn't say mine was particularly bad at either, but the issues well known to be associated with the AFM were there to an extent, so I decided to remove the AFM and take a look, luckily the AFM on my car seemed to be original in that it didn't look as if it had been apart before which was a good start! Anyway, after removing the cover, I set about doing the usual checks of the resistance track and wiper, the track itself although showing the tell-tale two lines from the wiper arm weren't too bad considering, but I decided that since it was on the bench I would look to see what I could do to alter the path of the wiper. Looking at the screws that hold the resistance board in place I decided to first try and loosen those and see if there was enough movement in the screw holes to move the board enough to allow the wiper arm to ride on a new arc of the track, it seems I was in luck. The next job was to thoroughly clean all the contact areas and surfaces with a good switch cleaner [being an electronics engineer I always have some to hand] once I had cleaned everything, and checked all the contacts for good continuity, I set about checking the resistance of the track, and then setting up a voltage supply across the the track so that I could measure the voltage difference on the wiper in order to see if there was any drop outs or weird readings as I moved the wiper [barn door] slowly through its full travel, I was relieved to see that there didn't seem to be any issues i.e. the voltage increased very smooth and consistent across the full travel etc. After re-fitting the now cleaned and re-set AFM to the car, I decided to check all the inlet manifold nuts as I had read that these can become loose over the years, and thus cause air leaks etc. Anyway, I was glad I checked as most were in need of a good tweak. First start up and I have to say I was pleasantly surprised that the car idled very smooth from cold for the first time since I had had it, the car never had the usual up and down hunting issue, but always felt like it had a partial miss-fire at idle, all that now seems to have gone, and after setting the CO on my gas analyzer, I decided to take her for a spin, the car seems smoother, with no intermittent bucking on part throttle as could sometimes be an issue, all in all; for little to no expense, just a bit of time and patience the over-all improvements are very worthwhile.

[911obgyn]

Nice write up. I found gaskets needed on my high mileage intake manifold, while its off is a good time to do fuel lines too.

[ant7]

Yes, I did wonder if there might be an issue with some of the gaskets after finding how loose some of the nuts were, but fingers crossed, time will tell I guess. A...

Originally Posted by 911obgyn

Nice write up. I found gaskets needed on my high mileage intake manifold, while its off is a good time to do fuel lines too.

[wazzz]

Torque needs to be checked again after a while, but only if you have changed gaskets.

When I changed mine car was 30 years and most gaskets on head side were cooked and stuck to head inlets so I had to carefully scrape and clean surfaces. Those on the intake manifold side were ok but I changed them anyway, as well as plastic shims in between.

I also did the AFM job you describe. Nice write up by the way.

[ant7]

Originally Posted by wazzz

Torque needs to be checked again after a while, but only if you have changed gaskets.

When I changed mine car was 30 years and most gaskets on head side were cooked and stuck to head inlets so I had to carefully scrape and clean surfaces. Those on the intake manifold side were ok but I changed them anyway, as well as plastic shims in between.

I also did the AFM job you describe. Nice write up by the way.

[TDHarris]

Hi Ant7,

Very nice write-up. I have a question about the resistances you measured across terminals 2 and 3 of the AFM. The Bentley Manual says the resistance should vary without interruption as the sensor vane is moved, but gives no numbers. I measured the following on mine (in kilo-ohms): 0.56 with vane closed, increasing to a high of 1.09, then decreasing to 0.75 with the vane full open. The changes were smooth, suggesting no breaks in the resistance board. But I was surprised to see the numbers reverse direction in the middle. Is this normal? How do my numbers compare to what you found?

Thanks in advance.

[wazzz]

Yes the reverse direction in the middle is normal. Can't tell about your numbers but to test linearity, you need to power the AFM with a 5V source and measure voltage between signal terminal and ground terminal. Measuring resistance while moving the flap is only good to make sure there is not sudden jump.

[ant7]

Hi Tom,
As wazz explained, it is a bit confusing when you measure the track with a multimeter, you get a better understanding of the condition of the track if you use a voltage source accross the full resistance, [ a small 9v battery is fine] and measure between wiper and end of track, while moving the wiper slowly accross the whole track area, meanwhile; here is another thread explaining the resistance measurements this OP came up with on his AFM, hope it helps, I can't remember the exact measurements of my own unit. AFM.http://forums.pelicanparts.com/porsche-911-technical-forum/256109-84-89-carrera-motronic-afm.html
Reading the resistance of the track can be confusing because if you look at the resistance board you will see sevral oblong sections which are tagged on to the resistor track at multiple intervals, which is why its better to use the battery method to see if you have a smooth uninterupted variable voltage accross the whole track.
Another thing to look at is the position of the toothed wheel, the AFM in the pick is as standard, again hope this helps.

Originally Posted by TDHarris

Hi Ant7,

Very nice write-up. I have a question about the resistances you measured across terminals 2 and 3 of the AFM. The Bentley Manual says the resistance should vary without interruption as the sensor vane is moved, but gives no numbers. I measured the following on mine (in kilo-ohms): 0.56 with vane closed, increasing to a high of 1.09, then decreasing to 0.75 with the vane full open. The changes were smooth, suggesting no breaks in the resistance board. But I was surprised to see the numbers reverse direction in the middle. Is this normal? How do my numbers compare to what you found?

Thanks in advance.

[Flat Six]

IIRC, AFM signal works on the basis of ratios: output Vdc / input Vdc. 3.2 Carrera and later (85.5+) 944 AFMs use a ~5 Vdc input signal. If you want to test your AFM, use a 5 Vdc signal on the input and measure Vdc output. A pretty good description is here:

https://forums.pelicanparts.com/porsche-911-technical-forum/1057714-afm-ohm-reading.html

FrankM shared the values he got by measuring Vdc at 11 different, equally-spaced points of wiper travel; this is much more accurate than trying to eyeball a "smooth rise". What you want to see is a linear progression of the ratio between output Vdc as a percentage of input Vdc.

Here are the values from my AFM measured at the same 11 points FrankM used:



You can see a very smooth progression across all 11 measured positions (e.g., a straight line):

[FrankM_]

Indeed ! The exact ohms do not matter it is the Vout vs Vin that gets measured in the ECU (the ecu provides a stable +5Vin).
It gives the ecu an indication on how ‘open’ the flap is and that combined with the temperature sensor in the channel allow the ecu to calculate how much air mass in entering the engine.


Sent from my iPhone using Tapatalk

[TDHarris]

Great information everyone. Thanks again.

[ant7]

As Frank and Dale have pointed out, the normal voltage accross the AFM from the DME is 5v, however; my reasoning behind using a small 9v battery was out of convenience, ie; not many folk have a 5v supply to hand.
Good luck with your tests, and let us know the results.

[wazzz]

Anthony, I used an old phone charger. SMPS supplying exactly 5V DC.
Any USB charger will also do the job.

[ant7]

Originally Posted by wazzz

Anthony, I used an old phone charger. SMPS supplying exactly 5V DC.
Any USB charger will also do the job.

[famoroso]

Heidi (of Heidi & Franny's Garage) shows how to test AFM using an oscilloscope at the ~3:45 mark...

https://youtu.be/L1jJpLzshP8

[ant7]

I was going to sugest using a scope as another method, but again; not everyone who would like to do their own AFM check have a scope to hand, or would know how to use one for this purpose.

Originally Posted by famoroso

Heidi (of Heidi & Franny's Garage) shows how to test AFM using an oscilloscope at the ~3:45 mark...

https://youtu.be/L1jJpLzshP8

[FrankM_]

Originally Posted by ant7

I was going to sugest using a scope as another method, but again; not everyone who would like to do their own AFM check have a scope to hand, or would know how to use one for this purpose.

It still is the best way however, my AFM had a discontinuity just above idle that was hard to see without a scope and between 2 measurement points (see table above). So it looked lineair but it wasn’t in reality and was the cause of a lot of jerking at low speeds.


Sent from my iPhone using Tapatalk

[ant7]

Indeed!
My own method;
Using a signal generator and a scope.

This is the same method I would use to test a standard audio potentiometer for drop outs.
This would be as follows:
Since you already have the unit out of the car, you will need a signal generator capable of producing a sign wave, for this test 1khz would be fine, feed the signal generator into the full resistance of the AFM track i.e.; ground on one end, signal on the other, 5 volts peak to peak should be fine, set your scope up to receive the sign wave, take your scope probe and look for the sign wave at the wiper contact, move the barn door/wiper contact and you should see the sign wave amplitude rise and fall as you move the barn door across its full travel and back.
Using this method it should be easy to see any drop outs or bad contact areas across the potentiometer track.

Of course; an analogue multimeter is much better than a digital meter at detecting real time voltage anomiles in this paticular situation too.

Originally Posted by FrankM_

It still is the best way however, my AFM had a discontinuity just above idle that was hard to see without a scope and between 2 measurement points (see table above). So it looked lineair but it wasn’t in reality and was the cause of a lot of jerking at low speeds.


Sent from my iPhone using Tapatalk