Don't let my post put you off. Experimenting is a lot of fun. I toyed around with a MAF conversion since 2005 and basically built up a pretty hefty set of EE skills in the process (background is in ME). I am down to provide some input along the way, because there are certain complications that you'll run into as you progress with the project. Getting the car to RUN with a MAF or MAP sensor isn't all that hard. Getting it to run as well as stock is the real challenge, and trying to make it run better is a feat of its own. The main advantage of a MAF or MAP is that you get slightly improved throttle response on a stock engine. I spent the time and money at a dyno to prove that you get no power gain with a MAF on a stock M42. If yours is turbo'ed as your username indicates, then there is some room for power gains, probably. Honestly, you should ditch the stock ECU in that case.
Really, the big bottleneck is the ECU. It is designed for the old, relatively slow AFM, and super fast new sensors are sort of wasted on it. In fact, the new sensors' fast response rates can and will give inputs to the ECU that it cannot properly handle and you'll end up running really lean, to the point of fuel-cut. There is no simple analog way to deal with that in the case of a MAF which has a non-linear output curve. If the MAP sensor you select has linear output, then you can get by with an opamp and some resistors and capacitors to filter the output (if it doesn't have filtering built in). Basically, the ECU seems to sample the air intake signal at around 180Hz based on the oscilloscope captures I observed and the RPMs that caused issues. That means that you really can't have large signal oscillations above 90Hz (Nyquist sampling theorem), and in practice anything higher than 70Hz is trouble for this ECU. 70Hz equates to 2100RPM on a 4 cylinder. New MAFs are so sensitive that they can resolve individual air pulses as intake valves open, and 4 cylinders are really difficult because there is no meaningful intake valve overlap to smooth the pulses out. A MAP sensor will also be very very sensitive to this, probably moreso. So it should be fairly obvious that you need to filter the signal. For a non-linear sensor like a MAF, you need its transfer function and then to apply a filter to the actual air flow values it puts out, rather than its voltage. You could either measure the MAF voltage, convert it to air flow in software, linearize it, kick it out via DAC and then apply an analog filter, or just apply a digitial filter to the air flow value in software and then kick it out after converting that value to its corresponding one in the stock AFM's transfer function. This all of course ignores that the stock setup uses the AFM to measure both volumetric flow rate and intake air temperature (IAT) to determine the mass flow rate, whereas the MAF does not output IAT since that is inherently accounted for so you'll need to figure out what to tell the ECU there. The MAP output is correlated to volumetric flow rate, so you could essentially yank out the IAT sensor and reconnect it. I have also found an equivalent thermistor to the stock one on Digikey if you end up needing one. Anyway, MAPs are awesome since they are simple and produce no obstruction, but the big drawback is that you need to re-tune every time there is any physical change to the intake system since their relationship to air flow basically follows the engine's volumetric efficiency curve.
Sorry about the wall of text. The point is that the conversion is not a simple matter. If you read all of that, get some/most/all of it and still want to work on this then I am happy to give advice. Messing around with this stuff is a lot of fun, but I want to give you a fair warning that you won't come out if it with any meaningful performance improvement aside from better throttle response if your engine is mostly stock. It took me YEARS to learn this and accept it, and it was disappointing. Truly, the best way to take full advantage of a newer air flow sensor is to use a stand-alone ECU with it.
Topic: better afm position (Read 5323 times)