Author Topic: Cheap Maf sensor conversion (Read 17456 times)

ApocolypseAutoEngineering · « **on:** March 15, 2008, 04:37:15 AM »

Hey guys I'm sorry I haven't been on in a while I've been in a bind between school work. Anyways since now I'm taking more time on my build I'd like to do some research and think this thread would do help for a bargain. I want to do a maf conversion for the m42 in my e30 but found that the only kit that still runs with the sttock ecu is supersprint, and that's a grand!! I know that the ecu is a motronic as well as an older stang's. They have real nice MAF sensors for the 302's that are alot cheaper than the Split second kit. Any suggestions?:confused:
Perhaps using the Ford motorsports maf sensor with a piggyback?

Thanks.
Louie

kowalski · « **Reply #1 on:** March 15, 2008, 11:38:01 AM »

if you migrate away from the stock ecu you can do it easily.

ApocolypseAutoEngineering · « **Reply #2 on:** March 16, 2008, 02:58:04 AM »

My problem is that I want the markd 93 chip and like the flow chart he has on it. If I can allow more air to flow in the motor with the maf sensor, it should prove to be a nice power increase. What I'm tryin to figure out is to make the maf sensor imitate to flow signals of the afm without the use of a new computer. I'm good with electronics so diodes, resistors and (hopfully not cuz its a pain) a board would not be a problem. I could possibly even create a diagram on how to do it if I could figure out the specs of the signal but can't find it anywhere without buying Alldata, and thats expensive.

1991 E30 M42 · « **Reply #3 on:** March 16, 2008, 11:38:09 AM »

I have access to Alldata pro through the dealership that I work at. Let me know what you need to know and I can probably find it.

ApocolypseAutoEngineering · « **Reply #4 on:** March 16, 2008, 06:13:13 PM »

Quote from: 1991 E30 M42;44879

I have access to Alldata pro through the dealership that I work at. Let me know what you need to know and I can probably find it.

Awesome!! What I need is the Ohms signals that get sent to the ecu from the afm. The Afm works on sending resistance measures and would like to find out what the flow in ohms is at idle, 1/2 throt, and wot. That can get me started on at least finding a MAF sensor of similar calibration and resistance band.

vatos_p · « **Reply #5 on:** March 17, 2008, 01:02:01 AM »

split second also has a maf convertion kit for bimmers...

kowalski · « **Reply #6 on:** March 17, 2008, 01:28:55 AM »

search around the site, bmwman91 has done a lot of research into this, and even made one for himself. it was A LOT of work.

ApocolypseAutoEngineering · « **Reply #7 on:** March 17, 2008, 03:26:08 PM »

Does anyone know if the afm signals of an m20 and an m42 are similar? I just found a sight that sells a MAF conversion for the m20 that is rather inexpensive (http://http://millerperformancecars.com/catalogue/67.html) and might be able to modify it for the m42. What I'm hoping is that the chip it comes with doesn't change how it reads signals.

kowalski · « **Reply #8 on:** March 17, 2008, 03:56:53 PM »

Quote from: ApocolypseAutoEngineering;44943

Does anyone know if the afm signals of an m20 and an m42 are similar? I just found a sight that sells a MAF conversion for the m20 that is rather inexpensive (http://http://millerperformancecars.com/catalogue/67.html) and might be able to modify it for the m42. What I'm hoping is that the chip it comes with doesn't change how it reads signals.

Miller performance is about 30 min away from me. they tried to do an m42 maf conversion, when i say tried, i mean TRIED. they failed, they had sold numerous pre-release units, which all had to be refunded... and they scrapped the project.

ApocolypseAutoEngineering · « **Reply #9 on:** March 17, 2008, 07:06:52 PM »

You know, not two second before I opened the forum back up I got this email.
"Hi Luis,

We tried an M42 MAF conversion this time last year. There were issues
with converting to MAF such as low air flow on idle. This caused issues
with drivability and we couldn't make a proper solution. Sorry for the
inconvenience but we do not offer a MAF system for your engine.

Thank you for you interest is our company though."
How ironic!!

bmwman91 · « **Reply #10 on:** March 18, 2008, 08:54:26 AM »

It depends on what you define as "cheap." Honestly, the EASIEST way to get the benefits you want are to go with a stand-alone ECU if you are after a DIY solution. If you are good at writing microcontroller app's and have access to a flow bench then you can probably do your own MAF conversion. EVENTUALLY I will have a version I will be comfortable selling. Presently, there is a little tuning left to do and until I replace my wide-band O2 sensor I cannot finish (more reliable than narrow-band...pffffft).

That is not to say that it cannot be done, but some working knowledge of electronics and programming are essential. The M42 and M20 AFM's outputs are not the same at all, so that wouldn't work. The main problem is not even that nobody makes a sensor that is mapped to the original M42 AFM. The issue is that the M42's lack of intake stroke overlap creates some nasty resonance in the intake track. MAF sensors are extremely sensitive, and these high-amplitude pulses (if not dealt with properly via software) will lead to the Motronic freaking out and cutting fuel. Check out this data log I took. The magenta line is the MAF's raw output. Each spike is an intake valve opening (range ~2000-2300RPM):

For the background on the first few versions of the MAF converter I developed, see below.
http://bmw.e30tuner.com/articles_mafcon.php

The present one is what I would call "Version 5" and is completely different than the latest one listed on the site as far as how it handles the signal (had a big breakthrough as to why I was having fuel-leaning issues). Eventually I will let one out for sale. However, I definitely think that I will need to purchase an M44 MAF just to see if some resonance issues can be worked out with it. Oddly, the original 5.0 Mustang MAF worked the best of them all, but physically fitting it in was ridiculously clunky and I scrapped it for the VW one. The VW one has, more or less, caused nothing but problems as far as getting this thing working quickly. It works fine, but getting it to do so has been a bear.

ApocolypseAutoEngineering · « **Reply #11 on:** March 18, 2008, 12:51:35 PM »

So in essence its reading pressure and vacuum pulses that would freak out the computer? Plus the pulses are happening for every valve opening? Is that throughout the rev band or at lower RPMs?

bmwman91 · « **Reply #12 on:** March 18, 2008, 03:45:37 PM »

It is not reading the pressure per-say, but the pulses as masses of air are traveling past the sensing element. Every time a valve opens, a discrete pulse travels back up the intake track. Engine designers spend a lot of effort tuning the system to be optimized for a specific RPM band on a given motor. When the intake pulses are aligned with the resonant frequency of the intake system properly, you get a supercharging effect (usually where the motor's peak power is made). Conversely, it will do the opposite when the pulses are 180 degrees out of phase with the natural frequency. Being a dynamic system, the intake pulses are only optimized for a narrow RPM range. A lot of newer 4 cylinders have a double intake track system with 2 sets of "pipes" of different lengths. This allows for 2 different target performance peaks. The E36 M42 uses one. Some fancy new BMW motors in the 7 series have an infinitely variable length system to keep the pulses optimized at all engine speeds.

The M42's (most 4 cylinders actually) have resonance issues because each intake pulse happens independently. On a 6 or more cylinder car, the intake pulses overlap causing a little bit of a damping effect. The issue with the MAF conversion is that the MAF is ridiculously sensitive so these effects and needs to be compensated for properly. The stock AFM is not nearly as responsive, which sucks for driving experience, but is a really simple way to bypass these issues.

These pulses are easily detectable throughout the entire RPM band, but only cause interference between 2500 and 3000RPM. Everything runs well outside of that band, but when the throttle is opened more than 80% in this range the car bogs. Actually, if you ease onto the gas in that range (nobody really floors it there anyway) it accelerates just fine. I just want to have it worked out before selling it to avoid having to tell people to take it easy on the gas there. Then again, it may be completely unavoidable. Most people say this is an acceptable "limitation" (easing onto the gas at low RPM's versus stomping it...seems reasonable).

ApocolypseAutoEngineering · « **Reply #13 on:** March 18, 2008, 06:26:48 PM »

Do you think it would work if the signal being sent to the ecu were dampened a bit? Not a lot but enough within the ecu's limits. Kind of like a capaciter smoothes power pulses sucked by the amplifier but backwards, it would prevent the signal to fluxuate as radicaly.

bmwman91 · « **Reply #14 on:** March 18, 2008, 08:50:07 PM »

Quote from: ApocolypseAutoEngineering;45046

Do you think it would work if the signal being sent to the ecu were dampened a bit? Not a lot but enough within the ecu's limits. Kind of like a capaciter smoothes power pulses sucked by the amplifier but backwards, it would prevent the signal to fluxuate as radicaly.

Unfortunately, no. That was actually the first thing I tried, but due to the non-linear nature of the MAF output this will make a badly leaned "average." The averaging had to be done in software with some custom algorithms I came up with to provide more damping (with about 40% of the phase delay of the original RC filter, yay).

Also, you will have to contend with the ambient air temperature. My original design "fooled" the ECU into thinking it was a constant 40C. All conversions from the mass flow rate to the corresponding volume flow rate were done using this constant value. This is great when the outside air is about this temperature, but I suspect the ECU retards the timing a little at high ambient temperatures. That's why I picked 40C initially...I didn't want knocking when it was summer time.

The new one uses a thermistor that is matched to the stock one *****ded in the AFM and it sends the real-time IAT signal to the ECU. My converter also picks up the IAT value and computes density on-the-fly and runs all the calc's for the mass-volume conversion rather than some lame look-up tables. So, the motronic is happy because I am not sending any fake signals to it, and I am happy because I don't have air meters with moving parts under the hood.

Sorry if I am being vague about the real meat in the converter. This is comething I have been planning to sell for a couple years now, but working a full time job makes development slow lol.