How the MAF CO screw does not matter in closed loop operation

Folks,

I may be late to the party but wanted to propose a correction to previous assumptions on the MAF function.

PREMISE: The ECU has several fueling maps stored that ere switched to and from depending on engine temperature and other operation modifying parameters. The standard fueling map takes into account a number of variables, one of which is the incoming airflow mass, to then generate the injector pulse, which is assumed to deliver x amount of fuel (x = the injector delivery rate per second at fuel rail pressure Y divided by the switching interval) so to achieve a resultant air fuel mixture (AFM) of just rich of 14.7 (stoichiometric) / lambda 1. If for any given regime, any of the sensor values deviates more than 20%, the ECU replaces the values with the median value hard coded in the EEPROM. Other sensor values used in the M2.5 to compute the fueling interval are the Throttle position sensor, Crank position sensor, etc. etc. The Motronic 2.5 system can work in open and closed loop modes, where it either calculates the AFM based on the table and corrects it by a lambda factor (closed loop) or just calculates it without correction (open loop). In open loop it converts an analog value (CO screw on the MAF) to a factor and applies it to the injector timing as if it were the lambda system.

OBSERVED FACT 1: It has been reported here than the MAF CO screw was wired in series with the injectors, according to th Motronic 2.5 schematic I got from Bosch, this is NOT the case.

PROOF 1: You can test this by measuring (use a scope) the voltage applied to the injectors while turning the CO screw and you will find it to remain constant, the only thing that changes it the switching tim.

OBSERVED FACT 2: The CO screw on the MAF has NO effect while the closed loop system is in effect (O2 Sensor). I tested this by rewiring my stock non cat Motronic 2.5 (EURO) '89 car to use the O2 sensor and set the MAF to 0 ohms. In effect this would put it to maximum mixture. With the O2 sensor connected, the gas analysis was within Lambda 1.0 (14.7 AFR) and swinging +- .2 AFR as it cycles normally "chasing its own tail".

TEST 2: If you disconnected the O2, within a few seconds, the HC values shot straight up and you could tell the mixture went excessively rich. This change is only seen when the throttle position sensor (Motronic 2.5 only has 3 positions) is at idle or partial load. When at WOT, then the O2 reading is ignored (as is the MAF one if the O2 is disconnected) and a standard value of 1 volt is replaced as the ECU goes to stored tables by default. Basically this is a case where, at idle and partial load "pottering about" - emission and economy know better, when at wide open throttle - Ferrari knows better ;-)

CONCLUSION: So the notion that you need to set MAF screws for closed loop management of the AF mixture seems to be somewhat incorrect. The resistance set by MAF screw, AKA CO screw, is only used when in open loop mode. CLEARLY having it in a "neutral" value range is a good idea, see below...

POTENTIAL VALUE ADD: This is why all the folks who had a working 02 system could have hugely varying MAF CO screw settings and still get away with it. And also why some people with a failing O2 don't notice as much as others... If your MAF is set to around 340-380 ohms, and the O2 sensor fails, when the ECU switches to open loop, it finds a relatively neutral MAF value which delivers a relatively benign HC concentration to the cats, and without telltale CELs you'll never really know. If your MAF settings are way off (eg very low or very high) the ECU does a value replacement "limp-home" activity which is by default over rich to prevent damage through excessive leaning, which leads to slow downs and cat problems.

Note 1: to add O2 to a non cat M2.5 In a stock Mondial t F119ALD engine with 25K miles, the MAF value to achieve a CO of .7 and HC below 350 ppm, was around 340 ohm.
Note 2: to wire up an O2 when the car does not have an O2 or cats, you need to ground Pin 31 on your Motronic (signaling this is a cat car), wire in the signal wire of the O2 sensor (3 wire type) into Pin 28, wire one (either) of the heater wire into the fuel pump supply wire coming from the relay with a 10 amp inline fuse for protection wire the other into a good ground position. Weld a sensor bung in your straight pipe, screw in the O2 sensor and connect everything up.
Note 3: to test your closed loop system, set the MAF to about 340 (on the rich side of stoichiometry) start up the car and let it warm up to operating temp. Stick a gas analyzer in the test port of the individual bank collector pipe BEFORE the common muffler. You should read an AFR of very close to 14.7. Disconnect the O2 sensor with the engine running, and you will see a very quick enrichment of the mixture - reconnect the O2 sensor and it goes back to normal.

 

So on the M2.5 cars here's some news...

The CO screw is only considered when in open loop. Open loop is determined if the 02 values are more than 20% off the expected value for a given MAF curve point and RPM, OR if they are disconnected and Pin 31 on the motronic plug earthed. Furthermore, the CO screw is only considered when the three way throttle position sensor reports idle or partial load. At WOT it goes straight to tables completely ignoring MAF/CO/O2's This is a fact and the TEST was on a rolling road, where at WOT the gas composition values all remain the same regardless of CO setting, Open or closed loop etc.

I spoke to Robert Bosch in the UK and they confirmed this mode of operation.

On the M2.7, the same is true at 72% throttle opening or more (WOT).

The tech at Robert Bosch clearly stated the CO screw is practically useless at all ranges providing the O2 sensors are fine, and were there for the vry few applicaitons of M2.5/7 cars that were not intended for US markets or European markets after 1993, where catalysts became mandatory.

Not arguing here, just reporting what I have been told..

Marco

 

Right, MAF is not ignored, CO screw is. Funny test, wire a switch onto the MAF resistance circuitry on a Dyno (open loop), go to WOT and switch off the signal (don't disconnect as it will trip the fuel pumps). You'll find the gas mixtures -> thus fueling curves are exactly the same as if the MAF resistance circuit is connected - that's because at WOT, it goes to standard values (tables) which Ferrari tested to death in their dyno/test labs, and overrides the MAF coefficient and O2 signal (if it were closed loop).

I have confirmed all these things. Really, the MAF is only used to optimize fuel burn at partial/idle. Think of it this way, when the throttle is at WOT, air is no longer a variable, but becomes a constant (max intake capacity). With that amount of air the fuel is also a constant, with the only variable being RPM. Ferrari tested this, determined which were the best settings for a vehicle that needs to be somewhat reliable given servicing interval variations, user, fuel quality and octane etc.

When at partial load then things become more complex (hence the reason that the M2.7 with a potentiometer is so superior to M2.5 with 3 logical settings, idle, partial, WOT) to ensure there is the adequate fueling for the airmass. With O2 the loop is closed and this becomes a function of the amount of O2 present in the exhaust (mixture always chasing emissions), while in open loop it is a function of a setting based on th CO output at idle.

Idling, just partial load at the lowest setting. The idle position switch on the 2.5 and (I believe 2.7) is only there to cut fuel supply, if the throttle is closed and the RPM is above a certain threshold, to ensure no fuel is supplied while the engine decelerates. Once it gets to a set RPM, fuel is then re-introduced at a rate determined by the CO screw in open or the O2 in closed loop operation.

We agree on all points, and now after testing the MAF at WOT (because I could not believe the Bosch guy either) I am glad to confirm, that at WOT "Ferrari knows best"

Marco

 

Which manuals are they - I'm desperate for more tech info on the M2.5/7 system. I'm 90% of the way to making new looms and being able to update the F119 M2.5 engine to a 2.7 or 5.2 system for less than 2500 bucks. As the m2.5 is one of the sources of reliability problems, this can pay itself in a single year's maintenance costs. DYI it will come down to parts an labor at around 1500 bucks US...

If they are PDF's can you PM me?

Marco