Memory dump of ECU

[b3lha]

Originally Posted by Dessertrunner

When I got a ECUstage 1 from Lan he said that the Aus SVX had a different pressure sensor (i think it was) at that his chip would not work in our unit. I have a reader on the way so when I get the info from my unit we will find out what the map is like. I would suggest that the differences in the maps between different models is related to things like air tempreture, allatude and fuel Octain aviability. The possiable differences in performance is most likly not much between all the models.
One question I have is do we all understand what each map does and how the ECU makes each decision as to what map info to use. Here is a couple of questions as to what I mean,
How is "Load" determined?
Is the fuel injection on timing only basied on Air/Fuel or is that only during crusie?

The reason I ask this is that we need to be clear about how each map and sensor fit together to understand how to change things.
Tony

Hi Tony,

I don't think we know enough at this stage to start changing things. Having said that, Mike (UPnorth362) has apparently tried running modified roms on his car.

I believe "Load" is calculated from the MAF. But I don't know the formula yet. More work needs to be done.

The fuel map is a target air:fuel ratio, based on the load and rpm. The ECU uses this to calculate the fuel injector pulse width. ie. The amount of fuel to inject.

The timing map is the amount of spark advance based on the load and rpm. But I don't yet know how the formula to relate the numbers on the Z axis relate to degrees of advance or retard.

The knock map is (I think) the amount that the ECU is allowed to retard the timing if the knock sensors report knocking. Again, I don't know the formula to relate the numbers on the Z axis to degrees of retard.

Lots more work has to be done before I can confidently say I understand how it works.

I think you are right that the different maps are related to the fuel availability in different markets. In the UK we have both JDM and UK models running very different maps on the same fuel. Assuming the UK one is optimised for UK fuel, the JDM one must be out of tune on UK fuel. That's probably why my JDM cars run better on 98 octane "super" than 95 octane "regular".

[Dessertrunner]

I think it would be good if you could read this artical I have a hard copy but don't have a scanner. In the full artical provides a insite into how part of the ECU system works. The artical talks about measuring vacum pressure but I can't see any sensor in the SVX to read that, can anyone help.
I might be able to borrow the SVX anlyzier and from that we might get a bit more data to help.
If we can decide how some of these things should be calculated we would know what to look for. Great project.
Tony

[Dessertrunner]

Fogot the link
http://www.siliconchip.com.au/cms/A_108521/article.html

[b3lha]

Originally Posted by TomsSVX

No, I am going to dump the factory ROM for you from the upgraded USDM ECU's.

Cool. Thanks Tom.
I don't suppose you have a Linux computer? It would be great to dump the whole ROM to a file. But unfortunately we don't have a rom dump program for Windows at the moment

[Trevor]

As I see it, the load on an engine must be expressed by way of torque, and there are no measurements available from sensors which can be used to calculate torque.

However the engine can be managed in anticipation of a load, and here the TPS would become an essential part of the equation. Once the engine has settled against a load, adjustments can be made so as to achieve maximum efficiency, but a calculation or measurement of the load/torque is not involved as part of that equation, nor is it required.

[Dessertrunner]

I have a Brother inlaw in the computer game think he might have one. Is it that the software you have operates on Linux?

[b3lha]

Originally Posted by Trevor

As I see it, the load on an engine must be expressed by way of torque, and there are no measurements available from sensors which can be used to calculate torque.

However the engine can be managed in anticipation of a load, and here the TPS would become an essential part of the equation. Once the engine has settled against a load, adjustments can be made so as to achieve maximum efficiency, but a calculation or measurement of the load/torque is not involved as part of that equation, nor is it required.

The term "load" is somewhat confusing.

I too was originally thinking of it as the "amount of load" the engine is under, which as you say is in terms of torque.
But actually I think it means load as in the amount of air being "loaded" into the cylinder. I found a description of a Ford ecu which said the Load = k x (M/n). Where k is a constant, M is the reading from the MAF and n is the theoretical max airflow. In other words it's the current airflow as a proportion (percentage?) of the maximum possible airflow. The formula may well be different for a Subaru, but probably similar. I can't help thinking that the TPS reading should be involved somewhere. I'll only know for sure once I get back to analysing the program code.

[b3lha]

Originally Posted by Dessertrunner

I have a Brother inlaw in the computer game think he might have one. Is it that the software you have operates on Linux?

I wrote a Linux program that will download the ECU memory into a file. The original purpose of this program was for finding the variables for the Select Monitor Software. I was dumping the RAM to a file, unplugging sensors, dumping it again and then comparing the files to see what changed. Then I realised I could download the ROM to a file and disassemble the code.

If your brother in law does have Linux (or a linux boot disk like knoppix) then it would be great to have a look at your ECU rom.

[svxistentialist]

Originally Posted by b3lha

No. That is the whole point. A person can only be challenged for breaking the laws of the country they live in. Do we really have to discuss this in a Technical thread?

Well, unless there is a separate thread set up elsewhere to discuss the possibilities and ramifications of copyright infringement as it relates to the ECU memory dumps and downloads discussed here, then I think yes, it should be discussed in this thread.

The Network has no policy statement about this area, and I think it is wise to debate and acknowledge the legal issues and the uses to which the information is put. It at the very least defines the work as non-commercial and for private use which may keep the wolves from the door.

Joe

[Dessertrunner]

I will check it out with him, I agree with you explernation of load, I will get the Subaru monitor and see if we can get some sort of relationship between TPS & MAF it would help.
Tony

[Trevor]

Originally Posted by b3lha

The term "load" is somewhat confusing.

In other words it's the current airflow as a proportion (percentage?) of the maximum possible airflow. .

That which is not strictly correct is surely somewhat confusing. However I now understand the bending of language involved.

I don't think the TPS would be involved in the factor you now describe.

Don't bend your brain in all of this Phil.

Cheers, Trevor.

[b3lha]

Originally Posted by svxistentialist

Well, unless there is a separate thread set up elsewhere to discuss the possibilities and ramifications of copyright infringement as it relates to the ECU memory dumps and downloads discussed here, then I think yes, it should be discussed in this thread.

The Network has no policy statement about this area, and I think it is wise to debate and acknowledge the legal issues and the uses to which the information is put. It at the very least defines the work as non-commercial and for private use which may keep the wolves from the door.

Joe

OK. Fair enough. But the phrase "as it relates to the ECU memory dumps and downloads discussed here" should be emphasised to keep any discussion on track.

[b3lha]

Originally Posted by Dessertrunner

I will check it out with him, I agree with you explernation of load, I will get the Subaru monitor and see if we can get some sort of relationship between TPS & MAF it would help.
Tony

If you have access to a proper Subaru Select Monitor, there is one experiment you could try that would be really useful.
Connect the select monitor to the car, then connect the receive wire of your interface to the transmit wire of the Select Monitor. Connect the ground too, and the power if necessary. I can draw you a diagram because it's hard to explain.

Then you can run the hex comm tool, and log the commands that the Select Monitor sends to the ECU for each mode. That will give the definitive memory address for each parameter.

[mbtoloczko]

In its most simple formulation, load is proportional to MAF signal divided by RPM. Usually there is an air temperature compensation built into the MAF value or one is multiplied into the formula. About a month ago, I downloaded the SVX ROM code and the first thing I did was start tracking down the exact load formula, but got sidelined also by a new family addition as well.

Load generally follows the same trend with rpm as torque. They are not equivalent values though. At any given RPM, torque is more closely proportional to load * engine efficiency at that RPM.

[b3lha]

Originally Posted by mbtoloczko

In its most simple formulation, load is proportional to MAF signal divided by RPM.

OK. So simplistically, if the MAF is measuring the volume of air per minute and you are dividing that by the Revolutions Per Minute, then Load can therefore be thought of as the volume of air per revolution? Or even per combustion if the constant multipler takes into account the number of cylinders.