Rolling Road Graph - give it to me straight....

[Kitchski]

Max_Torque said:

Max_Torque ages ago said:

the first thing i'd be worried about is the massive lean excursion at 3000rpm!

Think there is some kind of echo in here............. ;-)

I know, funny how it all works out

[Kitchski]

Mr2Mike said:

The old Motoronic systems only ever used a narrow band sensor, so WOT would be open loop. Why is there a target AFR table for WOT, and how is it used?

Lambda? ML4.1 doesn't use a Lambda sensor. As far as I can tell, mine simply uses:

Throttle position
Air flow
Crank sensor

Later cars used a knock sensor on a 3-row ECU (ML3.1 I believe) and subesquent later cars after that had a cat fitted (ML1.3 I think) which used a narrow band sensor.

[Mr2Mike]

Kitchski said:

Lambda? ML4.1 doesn't use a Lambda sensor. As far as I can tell, mine simply uses:

Throttle position
Air flow
Crank sensor

Later cars used a knock sensor on a 3-row ECU (ML3.1 I believe) and subesquent later cars after that had a cat fitted (ML1.3 I think) which used a narrow band sensor.

Which makes the "WOT target lambda table" even more unlikely to actually exist.

[Kitchski]

Mr2Mike said:

Kitchski said:

Lambda? ML4.1 doesn't use a Lambda sensor. As far as I can tell, mine simply uses:

Throttle position
Air flow
Crank sensor

Later cars used a knock sensor on a 3-row ECU (ML3.1 I believe) and subesquent later cars after that had a cat fitted (ML1.3 I think) which used a narrow band sensor.

Which makes the "WOT target lambda table" even more unlikely to actually exist.

I understand it as this:

clevermotronicguruinOZ said:

The ML4.1 uses an AFM (usually a VAF)to determine the amount of air entering the engine, and uses that to calculate an injector pulse time that will theoretically deliver a 1:1 lambda.
It then consults the fuel maps (idle, part throttle or WOT map depending on TPS position) and sort of divides the injector pulse by the desired lambda.
So if it has calculated an injector pulse of say 14ms for a given air flow to give 1:1 lambda (14.7:1 AFR) but the fuel map dictates a 1.1 lambda at that rpm/throttle/load/whatever set of conditions, it divides 14ms by 1.1 to work out the required injector pulse time to achieve 1.1 lambda, which would be 12.7ms.
So you can see that if 14ms gives 14.7:1 AFR, 12.7ms would deliver about 10% less fuel, so leaner mix...

That's a very crude and simplified description of how the target AFR/lambda calculation works, but essentially the ECU always starts by calculating fuel for 1:1 lambda, then uses that value to determine the required fuel delivery to achieve the desired AFR that the map specifies.

[Mr2Mike]

Ultimately it's just an open loop fuel map, even it's it worked out in a roundabout way. Since there is no long term correction (if there is no lambda), you could have the stoich fuel demand wrong and the target AFR wrong and actually end up with the correct fueling!