Reviving this thread because it’s the one that always comes up when I’m looking up AFM info. 
Figured I’d add my 2 cents while I’m in here.
When I first went through the AFM refreshing exercise, I measured using a 9V battery and the testing worked fine. I’m currently in the process of building out an “M20 simulator” to run a Motronic 1.3 ECU against. Still in the early stages, so I’m going through and getting a more concise understanding of the various sensor inputs I need to simulate. I’m on the AFM now, so found myself reading this thread (again).
Scott brought up the “circuit bridge” differentiation between the AFM and a typical “variable resistor”, but I figured I’d add details. Partially to inform those that may not know, partially to be corrected if my understanding is wrong.
The AFM is actually a potentiometer (or trim pot). This is the simplest type of voltage divider. To me, a good way to understand a potentiometer is like an adjustable resistor ladder. A simple resistor ladder is 2 resistors configured serially. The ratio of one resistor to the overall ratio dictates the voltage difference over that resistor.
Examples:
- If both resistors are equal – say both are 200 Ohm for a total resistance of 400 Ohm – then the voltage between the two resistors is the half of the input voltage.
- If the one resistor is 3x the other resistor – say 600 Ohm and 200 Ohm, for a total of 800 Ohm – then the voltage is 25% or 75% of the input voltage (depending on which side of the ladder the input voltage is).
Here’s the diagram of the M20/Motronic 1.3 AFM:
This is why simply measuring the resistance between 2 pins on the AFM isn’t sufficient. There are infinite possible resistor combinations that could result in the same output voltage. Based on that diagram, here are the important piece if you’re just going to look at resistance measurements:
- Measure the resistance between pins 3 (Vin) and 4 (GND). This is your overall resistance.
- With the flap closed, measure pins 2 (Vout) and 4 (GND). You should find very low resistance. Based on the numbers from the 944 page, with a closed flap being about .05 of Vin, then I’d expect a closed resistance of about 1/20 of your total resistance.
- With the flap fully open, measure pins 2 (Vout) and 3 (Vin). This should yield a number close to your overall resistance. Based on the numbers from the 944 page, with a closed flap being about .9 of Vin, then I’d expect an open flap resistance of about 9/10 of the overall resistance.
- Test intermediate flap angles to ensure linear change of resistance from closed-to-open.
I will test at my next opportunity, but I believe this test should yield identical results to the apply-voltage test. Of course, not saying applying a voltage is hard, but if all you have is a Fluke, it should work.
With that said, I’m still trying to get my head around the “laser calibration” and its purpose. The descriptions seem to all be around “fine tuning the resistance of the sweeper”, but it doesn’t really add up. Here’s the image from the 944 page:
Notice the irregularity in the gold sectionals around the sweep zone. This certainly seems intentional, but why? Could this be adding any kind of significant capacitance to the circuit? Also, if you look at the layout, we’re looking at a more complex bridge circuit. Sort of another resistance ladder running in parallel with the sweeper.
My totally post-midnight tired hunch here is that the surrounding area isn’t about tuning the resistance to make the sweeper’s resistance more precise. Instead, it’s about providing capacitance to prevent ringing/spikes/dips when the sweeper moves – making the voltage reading more precise. Effectively, capacitors work for electrons how water towers work for water pressure – providing localized stabilization of voltage (or water pressure). When the sweeper slides to a different section of the track, there’s probably a finite time it takes for the electrons to reach the equilibrium state needed for a valid measurement. Without the aid of the surround capacitance, these spikes could negatively impact what the ECU reads from the AFM. The capacitance, therefore, would be there to help fill/remove those electrons more quickly so the voltage snaps to its proper reading faster.
Since you have this long serial voltage ladder running parallel to the sweeper, the precision probably comes from the need to ensure every point of contact with the sweeper effectively matches the effective voltage ratio you’d expect at the touch point to the sweeper. If it’s off, you’d be affecting how linear the voltage response is to the flap angle.
The devil’s advocate in me wonders if the speed at which the sweeper moves could really be fast enough to induce the kinds of spikes that would necessitate a solution like the one I’ve described. If not, then I’m totally wrong and there’s another reason for it all.
Som
