Tuning with Evoscan
I have a question to those who use Evoscan to get data to tune their rom. Let us say that you have the following data:
RPM-----------AFR--------Load
3000-------12.054----173.75
3062.5-----11.9364---187.1875
3156.25----11.9658---200.9375
3187.5-----11.9952---221.25
3281.25----12.2157---165.625
3375-------12.2451---259.0625
3500-------11.9805---248.125
3562.5-----11.613---240.625
3656.25----11.3631---232.8125
3750-------11.6718---230.625
3875-------11.6718---228.125
3968.75---11.7012---229.6875
4031.25---11.5542---227.5378
The fuel map has three rpm rows for this data, 3000, 3500, and 4000 rpm. It also has the load points.
When you calculate the new AFR target number to enter in the cells do you only use the 3000 (12.054 afr) rpm data point, 3500 (11.9805 afr) rpm data point, and 4000 (11.5542 afr) rpm data point?
OR
Do you add the AFR points between 3062 and 3500, average them out and then calculate the new AFR target to enter in the load/rpm cell of your rom? For example, you add the 6 afr data points from 4062 to 3500 and divide by 6 and you get 12.05645. Then you calculate the target afr from that number.
If you use the latter method, then do you have some sort of a template to make your life easier since it would be a PITA to manually do this for all the logs that you have done.
RPM-----------AFR--------Load
3000-------12.054----173.75
3062.5-----11.9364---187.1875
3156.25----11.9658---200.9375
3187.5-----11.9952---221.25
3281.25----12.2157---165.625
3375-------12.2451---259.0625
3500-------11.9805---248.125
3562.5-----11.613---240.625
3656.25----11.3631---232.8125
3750-------11.6718---230.625
3875-------11.6718---228.125
3968.75---11.7012---229.6875
4031.25---11.5542---227.5378
The fuel map has three rpm rows for this data, 3000, 3500, and 4000 rpm. It also has the load points.
When you calculate the new AFR target number to enter in the cells do you only use the 3000 (12.054 afr) rpm data point, 3500 (11.9805 afr) rpm data point, and 4000 (11.5542 afr) rpm data point?
OR
Do you add the AFR points between 3062 and 3500, average them out and then calculate the new AFR target to enter in the load/rpm cell of your rom? For example, you add the 6 afr data points from 4062 to 3500 and divide by 6 and you get 12.05645. Then you calculate the target afr from that number.
If you use the latter method, then do you have some sort of a template to make your life easier since it would be a PITA to manually do this for all the logs that you have done.
use this
Suppose that the data points are , , ..., where is the independent variable and is the dependent variable. The fitting curve has the deviation (error) from each data point, i.e., , , ..., . According to the method of least squares, the best fitting curve has the property that:
Suppose that the data points are , , ..., where is the independent variable and is the dependent variable. The fitting curve has the deviation (error) from each data point, i.e., , , ..., . According to the method of least squares, the best fitting curve has the property that:
d1^2 + d2^2 +...+ dn^2 = Sum (y-f(x) ) ^2, from x=rpm low to rpm high
in your case it would be like this
(afr1-rpm1)^2 + (afr2-rpm2)^2 +(afr3-rpm3)^2.....
(12.054-3000)^2 + (11.9364-3062.5)^2 + (11.9658-3156.25)^2...
got it?
in your case it would be like this
(afr1-rpm1)^2 + (afr2-rpm2)^2 +(afr3-rpm3)^2.....
(12.054-3000)^2 + (11.9364-3062.5)^2 + (11.9658-3156.25)^2...
got it?
The fuel map uses a very linear interpolation method which makes it quite simple to tune. I usually take this interpolation in to account when tuning.
For instance, if it's lean at 3750/270, then the 4 data points around that value (3500, 4000, 260, 280) will effect that reading equally. I look in the map and log to see which value is the most logical to change.
For fine tuning, you can sometimes make very small adjustments by utilizing this interpolation. For instance, you can make fine adjustments to a load value of 265 by adjusting the 280 cell.
When it comes down to it, the ECU is only using the defined RPM and load points to calculate the fuel delivery, so we should think in the same manner while tuning. The ECU doesn't average, it interpolates, which is a little different ... so I'm not sure if averaging would be the best method.
For instance, if it's lean at 3750/270, then the 4 data points around that value (3500, 4000, 260, 280) will effect that reading equally. I look in the map and log to see which value is the most logical to change.
For fine tuning, you can sometimes make very small adjustments by utilizing this interpolation. For instance, you can make fine adjustments to a load value of 265 by adjusting the 280 cell.
When it comes down to it, the ECU is only using the defined RPM and load points to calculate the fuel delivery, so we should think in the same manner while tuning. The ECU doesn't average, it interpolates, which is a little different ... so I'm not sure if averaging would be the best method.
What does the "^" symbol mean in your formula?
Trending Topics
You are just a great help to me. Your tip worked. Now I can read Evoscan logs in Logworks and use the Logoworks chart function inorder to average the data points or use max/min/santard dev.A lot of people send me Evoscan logs, but Logworks has spoiled me in its data organization and presentation. Now I have the solution.
Thumbs up WAY UP to you
Ok, for those who are interested this how this works. I attached a basic LW template. If you have logged with LW, then you can create your own by chopping off the data and keeping the top portion of the variable description. Just open my template and you will understand what I am talking about.
Then you can open the Evoscan file and copy the data. PLEASE DO NOT copy the column name unless you have the same column name for both Logworks and Evoscan.
Paste the data into the template. For example, copy the data under the heading of "WidebandAF" to the column under the heading of "AFR" in my LW template.
Then save all the new data as a *.dif file.
Close the file and open it From Logworks. Then save it again as a *.log file. Now you can manipulate the data as you wish. Using the Logowrks chart tools that allow you to combine ALL logs into easy to understand tables that look the same as your rom.
All hail to Jorge for the tip
Then you can open the Evoscan file and copy the data. PLEASE DO NOT copy the column name unless you have the same column name for both Logworks and Evoscan.
Paste the data into the template. For example, copy the data under the heading of "WidebandAF" to the column under the heading of "AFR" in my LW template.
Then save all the new data as a *.dif file.
Close the file and open it From Logworks. Then save it again as a *.log file. Now you can manipulate the data as you wish. Using the Logowrks chart tools that allow you to combine ALL logs into easy to understand tables that look the same as your rom.
All hail to Jorge for the tip
Just giving back what I can from all I learn from here...I'm driking a beer.
I'm spoiled with logworks too!
I'm spoiled with logworks too!
Last edited by Jorge T; Aug 18, 2007 at 05:20 PM. Reason: slow connection
Awsome thank NJ. One last thing. I didn't see the jdm map in that file. I thought you used that too? If its not a hassle could you add that in there aswell. Thanks
Last edited by kreionic; Aug 18, 2007 at 05:43 PM.
while were on the subject of LW. I am having an issue with one of the charts. You know how with open port plug in you can view chart from ecuflash. How do you do that? I am lost when it says select the working rom. How do I do this. I suppose I would open ecuflash read image. (Do I leave ecuflash open while logging?) I am stuck when the manual says select "working rom"?







