tune your afr and stop your MAF from maxing out for $100

[Trevor]

Originally Posted by Myxalplyx

Remind me again......

Why is it that the maf sensor is significant if you can push as much airflow through it as you want and add the amount of fuel needed by measuring your O2 sensor (or Wideband O2 sensor)?

I mean, if you can add the necessary fuel needed by measuring your exhaust O2 readings (Or EGTs which are my preference with O2 readings), you really would not have much to worry about. As long as the ECU isn't pulling timing based on the erratic readings. An Apex ITC could be used possibly but that is for another discussion.

OK...I'm scrolling back to read some more. Must've missed something.

The O2 reading is taken after the opportunity for adjustment has passed. The opportunity to adjust has long gone.

The requirement to adjust, must be ascertained in advance of the opportunity to adjust. Hence the MAF at the intake.

Time is waits for no one.

[ItsPeteReally]

The O2 reading is taken after the opportunity for adjustment has passed. The opportunity to adjust has long gone.

Is this the right room for a five-minute argument?

I don't want to seem to be teaching someone who knows all about negative feedback how to suck eggs, but that is exactly what the O2 sensors are there for - to correct for errors in air/fuel ratio estimates generated from the MAF output. The narrowband O2 sensor (more properly, in my estimation, called the lamda 'λ' sensor) detects any deviations from the stoichiometrically correct air/fuel ratio and feeds back the error signal to the ECU. Because of the limited transient response of the λ sensor and the shape of its transfer function the air/fuel ratio continually oscillates around the correct value.

The problems start when the input airflow drives the MAF beyond its measurement range.

The first problem encountered is that there are no longer any correct fuelling values in the ECU's look-up tables. The error feedback from the λ sensors then signal a weak mixture and 'encourage' the ECU to deliver more fuel, but sooner or later the ECU will be unable to increase the injector duty cycle timings any further.

And this is only the first of the problems

If the only constraint upon arriving at a workable solution is maintaining the correct air/fuel ratio then, as long as we have access to the λ sensor output and a way of feeding that output as a control signal to an ECU, we do not need to know how much air the engine is consuming.

[Trevor]

Originally Posted by ItsPeteReally

Is this the right room for a five-minute argument?

I don't want to seem to be teaching someone who knows all about negative feedback how to suck eggs, but that is exactly what the O2 sensors are there for - to correct for errors in air/fuel ratio estimates generated from the MAF output. The narrowband O2 sensor (more properly, in my estimation, called the lamda '?' sensor) detects any deviations from the stoichiometrically correct air/fuel ratio and feeds back the error signal to the ECU. Because of the limited transient response of the ? sensor and the shape of its transfer function the air/fuel ratio continually oscillates around the correct value.

The problems start when the input airflow drives the MAF beyond its measurement range.

The first problem encountered is that there are no longer any correct fuelling values in the ECU's look-up tables. The error feedback from the ? sensors then signal a weak mixture and 'encourage' the ECU to deliver more fuel, but sooner or later the ECU will be unable to increase the injector duty cycle timings any further.

And this is only the first of the problems

If the only constraint upon arriving at a workable solution is maintaining the correct air/fuel ratio then, as long as we have access to the ? sensor output and a way of feeding that output as a control signal to an ECU, we do not need to know how much air the engine is consuming.

In five minutes you are teaching me nothing. Practical feed back loops are bread and butter within the field of industrial control engineering.

A feedback loop is a corrective circuit not a means of control. As you say that is what they are there for and the affect is a continual oscillation as you put it. A feedback loop can not be successfully used as means of correcting an inbuilt error signal, as the “oscillation” will override the designed parameters.

What you are suggesting is that the feedback loop should be used as a way of correcting what would be an abnormal condition. Modifying the the feed back loop in order to correct an inherent anomaly, is putting the cart before the horse, or an ambulance at the bottom of the cliff. Certainly not good engineering and in any event not a feasible option.

You correctly define the problem ---”The problems start when the input airflow drives the MAF beyond its measurement range.” ---The correct approach is to correct this, the first and a major error in the chain.

You state:- “If the only constraint upon arriving at a workable solution is maintaining the correct air/fuel ratio then, as long as we have access to the ? sensor output and a way of feeding that output as a control signal to an ECU, we do not need to know how much air the engine is consuming.”

The measurement of airflow is paramount in the management of any internal combustion engine. You are discounting throttle control, which is the deciding instruction commencing the process. A massive variant is involved, which can not be accounted for by any loop back from an output measurement, even if time were not a constraint.

Consider the train of control without air flow sensing as you suggest. Throttle opened/closed, the engine responds, the changed result is measured after time has elapsed, running adjustments are made based on after affect, the engine then responds. The resulting feedback cycle would surely be interesting/destructive.

[ItsPeteReally]

In five minutes you are teaching me nothing.

Is that my problem or your problem?

[Myxalplyx]

The reason why I stated what I did is because I have done this before. On my Outback Sport, I ran 5psi of turbo boost along with a 60hp shot of nitrous. The only thing I relied upon religiously was the O2 in the exhaust and even more than that, my Exhaust Gas Temperature (EGT). The car ran fine without issues at all.....at WOT. I left all non-WOT measurements for the ECU to take control over. Only at WOT did I take control of my air/fuel ratios by adding or taking away fuel. My point of reference was where the EGTs were when I ran the car stock. That is what I 'shot' for while tuning. As stone age as that may sound, I never had a problem between 59k and 110k (when I removed the turbo and nitrous). I did not know how to 'tune' so that is what I did and it worked.

So my point is, despite what the MAF says, why not take control of your air/fuel ratios by keeping the EGT at stock temps by controlling the fuel and timing yourself? Take moderate steps if boosting (Like 1psi, tune at WOT, repeat at 2psi). This should be able to work unless I am missing something.

[Trevor]

Originally Posted by ItsPeteReally

Is that my problem or your problem?

The sarcasm is noted. Obviously the wink I included, sa as to represent a sense of humour, was beyond your comprehension.

[Myxalplyx]

Originally Posted by Trevor

The sarcasm is noted. Obviously the wink I included, sa as to represent a sense of humour, was beyond your comprehension.

I'm still confused. You stuck your tongue out at the end of that one. Are you still joking?

[Trevor]

Originally Posted by Myxalplyx

I'm still confused. You stuck your tongue out at the end of that one. Are you still joking?

As a result of your unexpected reply so am I. Ask the addressee --- ItsPeteReally.

[ItsPeteReally]

On the whole, I’d rather people thought for themselves. I like to think that most people are capable of that.

Consider the humble MAF in its box, sitting next to hundreds of other MAFs in their boxes.
Are they all the same?
What do you think?

Well they are all obviously pretty similar, but they are probably not laboratory standard. Let us hope that someone in the manufacturing plant samples at least a few percent of them so that we may have some degree of confidence that they are not too wildly out.

Let us try and retain a degree of confidence that the MAF gives us some indication, within a few percent, that a certain quantity of air is passing, because we can measure the cooling effect that the air has on a heated surface.
We can convince ourselves that despite the pulsations and turbulence of the airflow, we have positioned the MAF so well that we can conveniently forget about such problems.

We have done some elementary physics and believe that, providing we know the air temperature and the barometric pressure, we can pretty much translate that information into knowing the mass of air, and consequently derive the mass of oxygen.

But wait a minute, how accurately does the ECU know about the variations in air temperature and barometric pressure?
Actually, wait even longer, shouldn’t there be a factor for air humidity in this calculation too, doesn’t that have a bearing on the cooling abilities of the air?

Ask yourself how much faith you now have about the knowing the mass of oxygen entering the engine.

Ask yourself how much faith you have that equal quantities of oxygen enter through the intake valves of each bank of cylinders, in any position of the IRIS valve.

We won’t even try to wonder whether equal amounts of oxygen pass each inlet valve in each bank of cylinders.

Well, now we know how much oxygen is entering each cylinder, don’t we?

Let us consider how accurately we can inject a known mass of fuel into the combustion chamber to be burnt by this mass of oxygen.

We could try and go through the same process, trying to figure out how accurately we can measure the fuel.

It starts to get very difficult, as we don’t even know what the fuel’s chemical composition is, (how much carbon, how much hydrogen for a start), we have no idea what its viscosity is and how it flows, and we don’t even attempt to measure what its temperature is, so we have no idea about how its volume relates to its mass.

The ECU simply attempts to pulse the injector for a certain amount of time, and thereby deliver a not very certain volume of fuel. How much oxygen that amount of fuel actually requires for complete combustion is largely open to conjecture.

It appears that we have gone through a lot of time and trouble to get something that could probably be done just as well, perhaps better, by the humble S.U. carburettor (at least you can actually adjust the S.U. carburettor!).

Using such an air/fuel management system the engine will run, but all of the air/fuel ratio mappings have to err on the side of caution to preserve the health of the combustion chambers.

But, consider what happens when you close the feedback loop by knowing just how well the combustion process has been carried out.

It turns out that the application of negative feedback, via the lambda 'λ' sensor, turns this flawed control and measurement system into a thing that might just work.

The lambda 'λ' sensor continually sends too weak/ too rich error correction messages to the ECU, which adjusts the fuelling. Not only that, but there is a sensor for each bank of cylinders!

It’s the feedback that makes the system work properly.

In my earlier posting I alluded to the saying that “You shouldn’t try to teach your grandmother how to suck eggs”.

An interesting phrase, what does it mean?

Does it mean that your grandmother has so much knowledge and experience about the craft of egg sucking that only a fool would believe that he knew anything that could contribute to her skills?

Or does it mean that she is so inflexible and sure that she must be right that she will never listen to you?

I leave the answer up to you.

Because my understanding of the matter could very easily be wrong.

[Phast SVX]

Pete,


Thats all in good. You do realize that the oxygen sensors in our system are not 5 volt lamdba sensors, and our ECU is not a wideband EUGO controller/decoder?

The basic Oxygen sensors in our system output a voltage based on the amount of oxygen present to indicate a "rich" or "lean" condition, not by means of RBT or LBT, but by means of stoichiometry (14.7 parts air to 1 part fuel). This type of sensor is only, yes i repeat only, utilizied by the ECU in closed loop, or partial to low throttle cruising conditions where 14.7:1 is an optimal operating range. This is what the feedback loop 02 approach is designed for in our system, but open loop, high load conditions switch to the load/volume based maps that we are takling about altering.

A lambda sensor can be utilized well by many aftermarket ECU's, that will tune on the fly. That coupled with EGT's and you can tune to near perfection and safety. I have a sensor in my car, and monitor all the time while driving. I do not run it in "Lamda" values, a .00-1.00 system that willi give you true output values, unless I am running Ethanol E85.

[Myxalplyx]

Originally Posted by ItsPeteReally

In my earlier posting I alluded to the saying that “You shouldn’t try to teach your grandmother how to suck eggs”.

An interesting phrase, what does it mean?

Does it mean that your grandmother has so much knowledge and experience about the craft of egg sucking that only a fool would believe that he knew anything that could contribute to her skills?

Or does it mean that she is so inflexible and sure that she must be right that she will never listen to you?

I leave the answer up to you.

Because my understanding of the matter could very easily be wrong.

http://www.worldwidewords.org/qa/qa-tea1.htm
"don’t give needless assistance or presume to offer advice to an expert"

Too many experts on this board for me. I bow down!!

[Trevor]

Originally Posted by ItsPeteReally

On the whole, I’d rather people thought for themselves. I like to think that most people are capable of that.

Agreed.

Your verbose and rather patronising lecture, indirectly indicates that you have now grasped the reason for, as well as the exact function of a feedback loop. What is more important, that a feed back loop does not, and can not form a function of control, as was your earlier suggestion.

[ItsPeteReally]

Originally Posted by Trevor

What is more important, that a feed back loop does not, and can not form a function of control, as was your earlier suggestion.

Maybe it's me, perhaps I've got it all wrong.
Convince me.
Pick a simple feedback system and show me why output measurement and feedback does not form the essential part of the control system. In fact you go further than that, you say it cannot perform a function of control.

Perhaps we could consider the ballcock in a cistern.
Or the pendulum and escapement in a simple mechanical clock.
Or the thermostat in a domestic oven.
Or you can pick your own example.

Educate me. I'm not too old to learn, and I'd like to think that I'm not too stupid to learn either.

[Trevor]

Originally Posted by ItsPeteReally

Maybe it's me, perhaps I've got it all wrong.
Convince me.
Pick a simple feedback system and show me why output measurement and feedback does not form the essential part of the control system. In fact you go further than that, you say it cannot perform a function of control.

Perhaps we could consider the ballcock in a cistern.
Or the pendulum and escapement in a simple mechanical clock.
Or the thermostat in a domestic oven.
Or you can pick your own example.

Educate me. I'm not too old to learn, and I'd like to think that I'm not too stupid to learn either.

Carefully read again my words as below, exactly understand their meaning, do not add any suffix and be convinced.

“a feed back loop does not, and can not form a function of control,”

N.B. “a function of CONTROL” not “of the/a control System".

From the dictionary within my computer:-

control
noun
1 the power to influence or direct people's behaviour or the course of events : the whole operation is under the control of a production manager | the situation was slipping out of her control.
• the ability to manage a machine or other moving object : he lost control of his car | improve your ball control.


Your original theory was that altering the feedback signal could directly alter an instruction originated and dictated by the MAF . This is not possible. The loop can only have an affect AFTER control has been set and the result is displayed and measured, whereupon recurring adjustments are put in train. The basic instruction from the MAF remains valid and is not altered.

I quote and comment on your suggestions of similarity:-

“Perhaps we could consider the ballcock in a cistern.” ---

The position of the float controls the inlet of water and fixes the level and there is no form of ongoing adjustment in respect of original setting which remains fixed.

“Or the pendulum and escapement in a simple mechanical clock”

Fixes the speed of the movement as per set parameters, but makes no ongoing adjustment. High class clocks incorporate mechanical thermal components for this duty.

"Or the thermostat in a domestic oven."

Again the operation of a thermostat does involve feedback but the desired temperature i.e. control is fixed via the setting. The thermostat functions maintain the set parameters but can not correct an error resulting from an incorrect setting. The manually operated dial correlates with the MAF in any form of comparison.

My example:-

Automatic audio volume control, AVC as has been incorporated for many years within every domestic radio. Here the volume control is the controlling component (Think MAF). An electronic loop adjusts amplifier gain in spite of variations in the input signal, by taking/using measurement post volume control. N.B. There is no control or influence in respect of the original set parameters.

QED.

[Phast SVX]

Wow

12345