Modifying Motronic 2.3.2 ECU hardware and software.

[prj]

Scratch that, divider is indeed R200 and R201, but R202 is also in play.

I had it worked out somewhere… Let me look it up.

[prj]

R1 = R200 = 4.86k
R2 = R201 = 5k6 or 6k2

For AAN you get 6.2/(4.86+6.2) = 0.561x
For ADU you get 5.6/(4.86+5.6) = 0.536x


To me everything checks out. I think you had resistor values between ADU and AAN mixed up.

[-ice-]

Yes, i think so too... i didn't measure R201 by myself, read it somewhere...


R202 doesn't play a role, because its connected to an operational amplifier or something like that (very high impedance input). There only flows very little current and so there won't be a remarkable voltage over R202.
I think R202 is used in an RC configuration and acts as a lowpass-filter.

So if you run RS2 components and Software with AAN Resistor the ecu will "see" less airflow than it actually is and inject to little fuel. Dangerous under WOT i think.

Bye

[prj]

It's the other way around, RS2 software + AAN resistor = too much fuel, as signal is higher than it should be.
Until the point where you max out the VFC that is (which is why this resistor modification was done on ADU in the first place).

[-ice-]

correct, i'm still mixing ADU/AAN resistor

Makes sense to use the whole input voltage range of the vco to get the best resolution and of course it must not max out.


Bye

[prj]

It's connected through S201 (op-amp) to S200 (VFC).

[-ice-]

I searched a long time to find a datasheet of the 91535v2 VFC chip from Harris cooperation. But i couldn't find anything. Would be interessting, because it's adjusted by many external resistors. So i'm satisfied with the determination of the voltage to frequency factor at the moment.

Bye

[prj]

I have all that data (voltage to frequency curve) etc, from when I made the ECU work to accommodate the HFM5.

Suffice to say, it takes voltage and outputs frequency and the frequency output is non-linear.

If you want to know more, then you will have to scope the output on the bench… although it is not really interesting...

[-ice-]

Ok, good to know that it's even non linear. But therefore we have the linearisation curves, which are consequently different between aan and adu.
In the pulsewidth calculation, when the linearisation factor is get from the curve, there are some constants used. One is added and the other one used as a multiplier. Do they have a name in the real world? Or is it just meant as gain and offset?

Regards

[Acki]

For my V8 I got an MAF sensor curve from Bosch.
Maybe you can also get this for both MAF sensors and you can better image what happens?!

[-ice-]

Hi,

I'm not sure, but I think ABH and AAN Maf are totally different. is there a curve for 034133471K MAF?

Bye

[Acki]

That's why you should ask bosch for your sensors.... ;-)

[-ice-]

I was hunting TVUB today...

I think it's a map (2D 5x1 voltage dependant)

I have seen such a map (f.e. 0x8ce4 for 551aa), which would fit and is also different between aan and adu. But i can't find where it is accessed in the code (no ram75/76 77/78 r2 combination fits...)

Anyone can help me?

Regards

[kymi]

TVUB is there just before FGAT0 (second map read before FGAT0). Goes to RAM_6F after the map access routine. At least in 551C adu.

[-ice-]

Thanks for the hint

I have forgotten a thing which I knew already...R2 is incremented with a map access. So it reads the assumed map with R2=1 after the first map access.

Bye