|09-07-2004 03:55 PM|
MAF sensor data *graphs*
In response to the "what's that" thread I thought people might be interested in the data on the MAF sensor off the T7 Saabs.
It may be overkill for those who just want basic info on it, but it could be very useful for anyone attempting to use this MAF unit on a custom ECU, or attempting to tune their own Saab ECU. It could also be useful when swapping MAF sensor's provided your new sensor has the same data.
Some quotes on it:
"The Mass AirFlow sensor contains two PTC resistors (RH) connected in parrallel in the airstream and electrically heated to 220C above air temperature.
Air temperature is measured with a special PTC resistor (RS).
As the flow increases, a higher voltage is required to maintain the temperature difference at a constant value. The voltage required is converted to ground pulses, the frequency of which increases with air mass flow....
... The control module converts the frequency to grams/second and then, assisted by the value from the crankshaft position sensor, to milligrams air per combustion. The unit is written mg/c (milligrams per combustion) and constitutes the principle value for the fuel injection. Normally, 1 mg of fuel is consumed at 14.7 mg/c. The value is also a good indication of the engine torque or load."
I only have data from 1-140 g/s. 140 g/s (grams per second) is about 20 lbs/min. Roughly 210 bhp of flow.
I used the data to make a regression which shows the parabolic nature of the plotted points. The reason for showing that is the show that the data should be able to be extrapolated further up in the power-range. So even though it's only plotted to about 210 hp, you could use the formula to find points (with reasonable accuracy) well past 210 hp.
After trying to use just one regression I realized that there was a fundamental flow change that caused a single regression to be inaccurate. So instead I used two regressions, one based on the 0-30 g/s data, and another on the 20-140 g/s data. The lower regression would be good for idle and slow cruise intepretation, and the other could be extended to high power interpretation. The Hz/flow is clearly parabolic.
The data goes as follows:
Hz = Flowrate in g/s
1045 = 1
1312 = 1.5
1527 = 2
2012 = 3.5
2355 = 5
2788 = 7.5
3134 = 10
3648 = 15
4075 = 20
4764 = 30
5289 = 40
5730 = 50
6114 = 60
6753 = 80
7286 = 100
7743 = 120
8158 = 140
Regression for flow 1-30 g/s: (2.092*10^-6)x^2 - .004537x + 3.838
Regression for flow 20-140 g/s: (4.757*10^-6)x^2 - .0291x + 60.267
The next two images show both regression curves plotted in two different frames. You can see that the curve which fits the data at low flow does not fit at high flow, and vice versa. 140 g/s is, again, roughly 20 lbs/min airflow.
Apologies for the poor nature of the graphs. I used a calculator to do the regressions and it was just easier to take a picture than try to mess around with Excell on the other computer, and then trasfer it to my laptop.