Afm Testing

[Amcubayt]

Ok, I have a misfiring 89 325i and I’m trying my hardest not to just buy an afm and see if that solves it. I’ve tried TONS of stuff vacuum, iac, spark, all that. In this post, I want to discuss afm testing. Per the Bentley, it says to measure resistance at pins 7 and 8 on the 25 point ecu harness. I tried that and I didn’t get any resistance values.. At the time of me posting this, I’m realizing there are far more than 25 pins on my harness..

1) am I measuring from the wrong points?
2) is this measurement only able to be taken from the afm itself? (Doesn’t seem to make sense since the ecu would need this info)

When I measure at the afm itself and move the flap, the raises and lowers. It’s hard to tell if there are flat spots. I also checked the intake air temp leads on that same connector and it read 3k ohms and the Bentley says between 2200 and 2700 at 68f which it’s pretty close to.

any help? Maybe this can help someone else. I’m literally trying to same myself from throwing money at parts needlessly.

[E30Bastard]

Hi,

It would have to be set insanely lean to be the root of a misfire. Could be a bad injector, could be bad valves, could be a bad plug. I suggest pulling one plug at a time while idling to see if you can stop the behavior but the following is something I saved when researching and adjusting my AFM (it was not the original AFM) - it worked out well before I swapped to MS2. Does your AFM look like it was tampered with (scored or missing plug over the allen bolt, i believe it was colored blue). I also remember there being a mod for the older AFMs but at 1989 you should have the factory modification present. Will follow up in the next post with harness pinouts for your peace of mind:

Here is the procedure I followed.

First some background information. The narrow band O2 sensor in our cars is designed to keep the engine operating within a very narrow range around the ideal 14.5:1 air to fuel mixture. The O2 sensor is basically a little chemical generator that will vary in output voltage between 0 and 1 volt depending on how much oxygen is in the gases exiting the engine. If the engine is running too rich, the output voltage will rise to 0.7 volts or higher. If the engine is running lean, then the output will drop to 0.2 volts or less.

The Motronic ECU tries to keep the voltage output of the 02 sensor between 0.4 and 0.5 volts by increasing or decreasing the fuel injector pulse width compared to the internal reference map. A couple times a second, the ECU checks the 02 sensor output voltage, the throttle position, the flow rate through the AFM, the idle bypass setting (telling it how much air is bypassing the AFM), the air temperature, engine speed and engine temperature and looks up the fuel injector pulse width and timing advance to use. The ECU then adds or subtracts a correction factor based on the 02 sensor reading. The problem is the ECU can only add or subtract so much from the reference value. Therefore, if there are any significant errors in any of the input readings the car will run too rich or too lean. That is why you need a new ECU chip if you swap the fuel injectors or make any other significant changes to the car's engine.

O.K. now back to what I did.

I took a piece of thin insulated wire and stripped about 2 inches off the end. I disconnected the O2 sensor connector and placed the wire into the connector for the output line (black on a Bosch sensor). I then carefully reattached the connector. I ran this wire over to my multimeter and another wire to the car's ground.

I then fired up the engine and let the car fully warm up (coolant gauge at the halfway mark). Once the car warmed up, the sensor was putting out an almost constant 0.7 volts indicating that the engine idle mixture was too rich and the ECU was not able to compensate. I then turned the adjustment screw on the AFM counter clockwise a half turn and then blipped the throttle. I kept doing this until the voltage output dropped down to below 0.6 volts and the output started bouncing around between 0.25 and 0.7 indicating that the ECU was now varying the mixture to keep it at the ideal output.

Now hold the car at various rpm levels (1K, 2K, 3K) and watched the voltmeter to see if the readings continued to bounce around the .4 to .5 level. If the results look good, then you have the idle bypass adjusted roughly correct.

[E30Bastard]

Here is the correct ECU pinout:

[E30Bastard]

Here's another article on the subject that I used for finetuning:

https://www.r3vlimited.com/board/forum/e30-technical-forums/engine-drivetrain/m20/363254-adjust-afm-spring-tension-and-bypass-screw-without-guess-work

[E30Bastard]

P.S. Your moderator said I'd never be able to do a burnout in his MR2 without responding to this post.

[2mAn]

P.S. Your moderator said I'd never be able to do a burnout in his MR2 without responding to this post.

Dammit... lol

[cheffy30]

buy a refurbished one. all these cars have 30 years old parts. well worth the money spent... sent from hell using Tapatalk

[E30Bastard]

Was there ever a fix to this or am I getting a free mid-engine rwd burnout?

*edit because

[Amcubayt]

Haha no solution yet. The afm is testing well.. I did however find metal shavings in my distributor cap and a piece of my rotor was bent.. weird.

[cheffy30]

did you open the afm ? the issue is inside there... I wouldn't trust the diagnostics too much... sent from hell using Tapatalk

[Gregs///M]

As crazy as this sounds but the Bentley repair manual AFM testing section is wrong. IN the manual, they use a multimeter to test the resistance of the resistor substrate. To use a multimeter is completely wrong and I will explain why.

A multi meter takes measurements and displays an average. So when you are measuring the resistance of your AFM, you are actually viewing the "average" of the resistance. As you can now be aware, this is completely useless.

The AFM should be tested on the car, if running, using only an oscilloscope. The o-scope can view the AFM output signal in real time with a micro or pico or nano second resolution. This means it can capture failures, or voltage drops on the micro second scale, something a meter can never do. A meter should never be used to test the output of the AFM.

If you test the AFM on a bench, only a 5v regulated power supply can be used, not a 9v battery or other power source. There are many articles and posts on the internet that will be different than what I am explaining but it would be wise to trust my knowledge as I am have been working in the electrical engineering industry for 15 years and have been a pro mechanic for close to 25 years. IN addition, I have refurbished many thousands of AFM's for classic BMW owners across the world.

In the end, and ultimately, if your AFM is 30 years ago, assume its bad and get a professionally refurbished one. Well worth the money, I have seen cars that run perfectly fine gain 10+ HP with one of our AFM's, better idle, more torque and pass smog easier. The AFM is ultra important because it tells your DME how much and what temperature the air is so it can determine how much fuel to give. See my article for more info and pics of my testing graphs. Email me if you have more questions. https://www.r3vlimited.com/board/for...afm-s-for-sale

[TobyB]

A multi meter takes measurements and displays an average. So when you are measuring the resistance of your AFM, you are actually viewing the "average" of the resistance. As you can now be aware, this is completely useless.

You are referring to sampling rate. The refresh on a DIGITAL meter is often quite slow. An ANALOG meter of decent quality that has a relatively little damping can show all sorts
of things that the DMM does, indeed, try to time- average.
Flukes and other GOOD DMMs try to compensate for that by providing an instantaneous 'digital bar graph' at the bottom of the display, and if you've had some experience
reading that AND lock the damned thing into one range, you CAN see an obviously bad AFM on a DMM. But yes, it takes experience.

In the end, and ultimately, if your AFM is 30 years ago, assume its bad and get a professionally refurbished one

IF it's misbehaving. Do keep in mind that the 'professionals' can't get new resistive traces either, and just bend, modify, or replace the wiper
to hit a clean part of the old trace. If you're lucky, they then check to make sure that the start and end points are within 'spec'

'Professional' success rate is marginally higher than 'intelligent car owner' but neither is anywheres close to 100% success. As dozens
and dozens of Pro3 drivers will attest.

Haha no solution yet. The afm is testing well.. I did however find metal shavings in my distributor cap and a piece of my rotor was bent.. weird.

check your motor mounts.

t
older than alpha-n