Thread: How to completely get rid of MAF on de-EGR'd TD5

Originally Posted by OffTrack

It's definitely not direct, or even obvious, probably via engine torque.
Parameters like engine torque are derived from multiple sources of data and incorporate adjustments for various factors.
MAF will be part of the mix.

Thanks Paul, thought it was a function of several inputs. You've torn the TD5 ECU to pieces so knew you'd have it.

Originally Posted by Tombie

Thanks Paul, thought it was a function of several inputs. You've torn the TD5 ECU to pieces so knew you'd have it.

Tombie, I wish I could say with 100% but decoding the Auto code hasn't been a top priority so it's on a long list of todos.

Paul

I've wasted a couple of hours proving to myself what I already knew so I might as well share.

NOTE: The gory details are in the attached PDF, including plots of the MAF and MAP readings from the Nanocom log.

The way I've done this is to:
1. Calculate the MAP output voltage from pressure.
2. Use MATLAB to fit a curve to the ECU MAF calibration
3. Calculate the output of the MAF running from the MAP voltage.


1. MAP voltage is calculated using simple algebra. The slope of the sensor curve is determined from the minimum and maximum points.
In the case of the stock MAP these are 20kPa//200mV and 250kPa//4650mV.
Plugging this into the slope-intercept equation gives the voltage output for the corresponding pressure.

2. The Td5 ECU stores it's MAF calibration in a lookup table which is used to convert voltages -> MAF readings.
I've taken the voltage and airflow figures from the table and used MATLAB's Curve fitting toolbox to create a model.
This returns the Airflow for a specific input voltage.
Note: The sum of squared errors for this model is 0, which indicates a perfect fit.

3. The MAP output voltage and MAF interpolation model are used to calculate the MAF reading for the MAP output.



From these calcs it can be seen that the MAPvoltage ->MAF has a maximum theoretical reading of around 580.
In practice the stock boost limiter will trip well before this, so the maximum MAF value that can be produced by this method is 508 kgHr at full boost.


This is the final plot showing the MAF and MAP readings from a 2003 Td5 Disco (Auto) running a stock map with brand new VDO MAF and air filter.
This data was taken from a Nanocom log of a dyno run by a leading UK Td5 tuner done a couple of days ago.
His assessment was that the vehicle was running well.




The red diagonal line indicates what would be a perfect match between the MAF reading, and the MAF value calculated from the MAP.
What this shows is that the MAF driven from the MAP has significant under-read errors.

If you are running a EU2 Manual none of this is going to make a huge amount of difference, so knock yourself out.
However, I'd suggest there is no real point in doing so except to avoid the 20 second MAF timeout.

For Auto's and EU3 motors - this is worse than a stuffed MAF.

I'm not going to bother with the throttle pedal because its going to be even worse.

I've already recognised earlyer in this thread that using the TP input is not the best ideea so no need to convince me further with "put out of misery" or "knock yourself out" kind'a expressions.

If you want to have a constructive discussion better explain what do you mean by that:

Originally Posted by OffTrack

....I've been working on a recalibration for the stock MAF which extends the range quite a bit.
It should read virtually the same as stock to 680, but theoretically extends usable range to 840ish.....

cos as i've said i made various bench tests with voltage measurements on the MAF's output and there are cases when reality beats all kind of graphic simulations so maybe we can work together on this rather than having a debate....i'm in the process of buying an anemometer to make more accurate readings for that and be able to see how it works in reality

i'm searching for some cheaper version of this kind of tool Kane DAFM3 Digital Air Flow Meter , if anybody knows one please give me the link

Originally Posted by sierrafery

I've already recognised earlyer in this thread that using the TP input is not the best ideea so no need to convince me further with "put out of misery" or "knock yourself out" kind'a expressions.

"put out of misery": idiomatic english, used in the sense of: to end someone's worry or something's suffering. "I thought I'd call her with the test results today and put her out of her misery. Business was so bad, they considered bankruptcy to put the company out of its misery."

"knock yourself out" idiomatic english. It means roughly "Do whatever you want, it doesn't bother me."

Not exactly sure why you are getting your knickers in a twist about them...

Sierra made a comment a couple of posts ago which came across as rather dismissive what he called graphical simulations.

Lest anyone take away the impression that I'm making it up I thought I should clarify...

Over the past few years I've done a lot of work reverse engineering the Td5 ECU.

In the process I've had to trace the circuits between ECU plug and the inputs on the MC68336 micro controller chip used. This has also meant working out how the Analog to Digital Convertor module is configured, and how the output of ADC is processed to give the sensor readings you see on a Nanocom.

The Nanocom omits a number of less obvious data results. For instance the MAF sensor voltage in addition to the well known MAF Airflow reading.

This shows the MAF voltage a run where the MAF limited at 4950mV.



And the corresponding MAF Airflow log with the MAF reading limiting.


Applying the MAF curve model I mention in the earlier post to the MAF voltage log, gives a very similar result to the ECU MAF Airflow data.


Plotting the actual MAF reading vs Calculated from mV shows an over-read in the calculated values that increase are flow increases. I didn't show the linear fit because two outliers occur at 700,0 where the ECU MAF reading has clipped.



The end result is probably 95% rather than 100%, but usefully close to the real world.

There are other factors in play. Each log entry is roughly 400ms apart, but there is also a small time skew between individual items in each entry ue to the overheads in the diagnostic requests. The time skew has a tendency to smear results calculated for individual readings. The perils of log analysis...

cheers
Paul

About the idioms thing, i thought you meant like it's traduced at nr 1. (cos that's how i felt, sorry):
http://idioms.thefreedictionary.com/put+out+of+misery as you mentioned a back of the farm shed too
or ''usage notes" http://idioms.thefreedictionary.com/knock+yourself+out

as i said, i'm thinking in other language and writing in english

I would never insinuate that you are making up things believe me, english is my third language maybe that's why some of my statements are misunderstood... i really respect your work and that's why i asked about that MAF recalibration thing cos that could be very interesting.... still waiting for the answer

Originally Posted by sierrafery

About the idioms thing, i thought you meant like it's traduced at nr 1. (cos that's how i felt, sorry):
Put out of misery - Idioms by The Free Dictionary as you mentioned a back of the farm shed too
or ''usage notes" Knock yourself out - Idioms by The Free Dictionary

as i said, i'm thinking in other language and writing in english

I would never insinuate that you are making up things believe me, english is my third language maybe that's why some of my statements are misunderstood... i really respect your work and that's why i asked about that MAF recalibration thing cos that could be very interesting.... still waiting for the answer

Sierra,
Any member making actual or implied threats of violence against another would be banned from the forum immediately. You are reading what I said literally, when it was intended figuratively. I'm sorry you have misinterpreted that.

The post was not directed at you personally - just the fact that there is so much wrong information in the thread. Dougal for instance assumed that the Td5 operated in the same way as his Izuzu diesel and as a result his comments on the effect of disconnecting the MAF are incorrect, yet the argument on that point spans pages & pages.

cheers
Paul