Some observations regarding idle quality & power product

[Klaus]

Just spend a couple of weeks debugging some issues after the final 5-speed conversion. Car is a 81 E12 based E24 with an earlier M90 L-jet and a slightly modified transmission from an E30 M3. Engine has roughly 135k miles with a top end refresh about 25k miles ago. Engine management is done with the stock components of a 81 US 633CSi, but using the stock M90 distributor and an oversized Magnaflow converter to keep things pretty emission legal in the state of NV.

Anyway, idle was always and issue, checked for vacuum leaks, experimented with different valve clearances, timing set-ups and fuel pressures from 2.5 -3.0 bar. But it never occurred to me to actually check the Throttle Position Sensor for functionality past the initial check of clicking of idle and clicking again above 3/4 throttle. Imagine my surprise when I actually ohmed it and the switch would not go of the idle signal, meaning as far as the ECU was concerned the engine was always in idle. Looks like that might explain the more then sluggish acceleration past 3000 rpm, obviously more noticeable with the manual then the slush box.

Well, after replacing the TPS, idle was still crappy, but acceleration was getting better. So, let's try another AFM, after all I collected 7 over the years, some undisturbed with the seal still in them and the lid firmly glued on, some modified and tempered with following any procedure to refurbish and readjust a used AFM. Also have one refurbished and recertified by A-1 Cardone, nicely painted in spray can black for quality control. Improvements to idle quality and power production varied through the whole range, some with better idle, some with more top end power.
Then I ended up getting a decent deal on a brand new one, old ware house stock produced in 92-93 still in the original box.
Wow, what a difference! Idle is virtually perfect, or at least as good as it might get for any M30 or M90 and the power difference is spectacular through the whole range. That's without touching a single adjustment other then turning down the idle 200 rpm on the throttle body.

So the lesson of the story, while you might be able to squeeze some life out of your worn out AFM, you might have a hard time reproducing the factory calibration. Also, vendor rebuilt AFM's are as bad as trying to adjust them your self, unless you prefer some nasty black spray paint over a cleaned, natural casting AFM.
After looking around for a while, most AFM's are still available through BOSCH distributors, the trick is not to search by the BMW part number but to use the BOSCH part number. A lot of them are available through Porsche and Alfa sites, even rockauto.com lists quiet a few as available.

Hope this helps some body,

Klaus

[BillDe]

Klaus, is this AFM the same for the stock euro system, as for the US system you have on your motor? I have a 1980 euro 635, and I to have some extra used AFM and other parts in reserve, for just in case. Did you happen to notice in your research if they are different and/or available?

Your result sounds great, good work.

Bill

[Brucey]

Re AFMs; two points;

1) IIRC L-jet ones have a reverse reading to motronic ones, so that when the most common fault occurs (which is an intermittent reading on the wiper) the result is either a lean misfire or a rich misfire. (BTW you need the exact right part number in most cases and searching via Bosch number is definitely the way to do it...)

2) The spring calibration in the meter. For some years I have hoped to at least find definitive proof that AFMs go out of calibration because the spring tension varies. In the absence of that I have assumed that it is unlikely that the spring tension would vary, and indeed measurements of spring tension appear not to change dramatically. BUT... more recently it has been suggested to me that the spring is actually a bimetal spring (anyone?) and that as well as a static (cold) tension value it varies with temperature....

Anyone have any info on the latter point?

cheers

[Klaus]

Klaus, is this AFM the same for the stock euro system, as for the US system you have on your motor? I have a 1980 euro 635, and I to have some extra used AFM and other parts in reserve, for just in case. Did you happen to notice in your research if they are different and/or available?

Your result sounds great, good work.

Bill

Bill, it is my understanding that the individual AFM's are matched by part number to the matching ECU's, meaning that a particular ECU needs a particular AFM to work properly through the whole range.
Your 80 Euro 635CSi uses a BOSCH AFM 0280203002 or 0986280066. The application list for this AFM just includes Euro vehicles without Catalytic Converter. BMW uses this meter for E12 528i, 535i; e23 728i, 733i, 735i; E24 628CSi, 633CSi, 635CSi; E3 3.0Si, 3.3Li. I found several new ones for sale during a brief google search, all in the range of $350 to $450.

I have personally mixed and matched ECU's and AFM's over the years, with different results, but no catastrophic failures, and depending how picky you are, you might be able to recycle and readjust a similar meter and achieve good drivability.

Re AFMs; two points;

1) IIRC L-jet ones have a reverse reading to motronic ones, so that when the most common fault occurs (which is an intermittent reading on the wiper) the result is either a lean misfire or a rich misfire. (BTW you need the exact right part number in most cases and searching via Bosch number is definitely the way to do it...)

2) The spring calibration in the meter. For some years I have hoped to at least find definitive proof that AFMs go out of calibration because the spring tension varies. In the absence of that I have assumed that it is unlikely that the spring tension would vary, and indeed measurements of spring tension appear not to change dramatically. BUT... more recently it has been suggested to me that the spring is actually a bimetal spring (anyone?) and that as well as a static (cold) tension value it varies with temperature....

Anyone have any info on the latter point?

cheers

Brucey,

good point on the difference between L-Jet and Motronic AFM's, really no way to interchange them. As points of failure, I also had one AFM fail at the fuel pump switch between 36 & 39, and I had several stick open after high air volume runs, maybe due to a slightly bend shaft or residue.

As for spring tension, I would assume that the shaft gets harder to turn over the years and the spring looses tension. The bimetal spring theory sounds intriguing, but after looking at an open meter I had laying around under a magnifying glass, I don't see any indication. I rather think the meter uses the build in temperature sensor at 6 & 27 as reference for warm-up.

Klaus

[Brucey]

I have personally mixed and matched ECU's and AFM's over the years, with different results, but no catastrophic failures, and depending how picky you are, you might be able to recycle and readjust a similar meter and achieve good drivability.

Note also that in motronic AFMS there are several different variations; some give a linear output with flap angle and others are calibrated to give a linear output with airflow. It isn't obvious which is which (they might appear to be physically and electrically similar otherwise)

Re AFMs; two points;

1) IIRC L-jet ones have a reverse reading to motronic ones, so that when the most common fault occurs (which is an intermittent reading on the wiper) the result is either a lean misfire or a rich misfire. (BTW you need the exact right part number in most cases and searching via Bosch number is definitely the way to do it...)

2) The spring calibration in the meter. For some years I have hoped to at least find definitive proof that AFMs go out of calibration because the spring tension varies. In the absence of that I have assumed that it is unlikely that the spring tension would vary, and indeed measurements of spring tension appear not to change dramatically. BUT... more recently it has been suggested to me that the spring is actually a bimetal spring (anyone?) and that as well as a static (cold) tension value it varies with temperature....

Anyone have any info on the latter point?

cheers

Brucey,

good point on the difference between L-Jet and Motronic AFM's, really no way to interchange them. As points of failure, I also had one AFM fail at the fuel pump switch between 36 & 39, and I had several stick open after high air volume runs, maybe due to a slightly bend shaft or residue.

I've seen the same thing too. The pivot bushings wear and the casing (which is glued together) can deform over time or with heat. If there are any scuff marks inside the housing from the flap then the AFM is 'at risk'. You can often correct this by using wet and dry paper to dress the housing and the flap edges.

As for spring tension, I would assume that the shaft gets harder to turn over the years and the spring looses tension. The bimetal spring theory sounds intriguing, but after looking at an open meter I had laying around under a magnifying glass, I don't see any indication. I rather think the meter uses the build in temperature sensor at 6 & 27 as reference for warm-up.

Klaus

The flap should appear to open further, more easily, as time goes on if the spring weakens, and the result would then be rich running.

My assumption has always been that the air temperature sensor would be used for temperature correction but this doesn't exclude the possibility that there is an additional system that also contributes to this.

I intend to investigate the bimetal spring hypothesis further; I agree the springs do not look to be obviously bimetal but such materials are not always obvious by appearance alone.

cheers

[BillDe]

Thanks for the info Klaus!!