Mitsubishi Lancer Evolution 4/5/6/7/8/9 EvoScan Project v0.8

[laakness]

the wideband did not work i have the LM-1.......

but i had a problem when i started data logging...... i was idleing and i started
to data log my rpms jumped up to around 2500rpm and i throw a code... i turned the car off and started it back up everything was fine no code then i started to datalog and the same thing happened..... what is wrong???

Are you supposed to have both of the connecters hooked up when your logging or just the OB2 port?

I use just the OBD port.

[Ludikraut]

the wideband did not work i have the LM-1.......

but i had a problem when i started data logging...... i was idleing and i started
to data log my rpms jumped up to around 2500rpm and i throw a code... i turned the car off and started it back up everything was fine no code then i started to datalog and the same thing happened..... what is wrong???

Are you supposed to have both of the connecters hooked up when your logging or just the OB2 port?

I have the exact same problem, except that my idle jumps to 3000 RPM. I tried it with and without the second connector of the tactrix cable hooked up. The idle jumps with and without the second connector hooked up. I'm thinking it is somehow affecting the ICSV setting, but I cannot confirm, since I also got several garbage log files (reading RPM of 7968.75 at idle and what look like max values across the board - see attached file).

LC-1 support does not work. It tries to access com1, but gets no data.

l8r)

[Evo4Mad]

Thanks for the datalog and feedback, looks like the LC-1 doesn't respond as quick as I thought, its hard when I don't have any widebands here on my desk

LC-1 support does not work. It tries to access com1, but gets no data.
l8r)

[laakness]

Laakness... Definition works good, only there are spaces appended to the code you posted, you probably should attach it as a file to avoid issues.. I was almostt as far as you got when I noticed a issue.. DLLs definition files dont allow combining of column data...

Hence the One glitch I found, Evoscan breaks out the time and seconds as separate columns, they need to be combined otherwise you can only have 60 seconds of data displayed when it goes back to zero it all drops off the end... Hopefully Evo4Mad can give us a cumulatve seconds from start of logging column, or combine the time and seconds columns together.. I havent found a workaround other than to use the samples (LogID) but the software doesnt maintain a specific sample rate, so none of the time calculations work right..

Hamish, ya think you can accomodate something for us so we can also use this tool? Maybe release sourcecode so I can create a simple logger for this?

It looks like if you change the [Column] definition for the seconds to data type 51 and then change the logid column to 63 (time column) the graphing shows more data. I think all the program wants for a time column is an incrementing column with unique values.

[MalibuJack]

Laakness... Definition works good, only there are spaces appended to the code you posted, you probably should attach it as a file to avoid issues.. I was almostt as far as you got when I noticed a issue.. DLLs definition files dont allow combining of column data...

Hence the One glitch I found, Evoscan breaks out the time and seconds as separate columns, they need to be combined otherwise you can only have 60 seconds of data displayed when it goes back to zero it all drops off the end... Hopefully Evo4Mad can give us a cumulatve seconds from start of logging column, or combine the time and seconds columns together.. I havent found a workaround other than to use the samples (LogID) but the software doesnt maintain a specific sample rate, so none of the time calculations work right..

Hamish, ya think you can accomodate something for us so we can also use this tool? Maybe release sourcecode so I can create a simple logger for this?

It looks like if you change the [Column] definition for the seconds to data type 51 and then change the logid column to 63 (time column) the graphing shows more data. I think all the program wants for a time column is an incrementing column with unique values.

Yep, your correct, however then since the sample rate varies, all of the horsepower, time, and speed calculations are wrong.. Last I checked my car isn't making 43000 horsepower and travelling 840mph.. My car is a figurative rocket.. Not a literal one..

[MalibuJack]

Thanks for the datalog and feedback, looks like the LC-1 doesn't respond as quick as I thought, its hard when I don't have any widebands here on my desk

LC-1 support does not work. It tries to access com1, but gets no data.
l8r)

Found another issue you probably already know about... the Airflow reading (presumably MAF HZ) never goes over 1603.95.. Kinda puts a hurting on my logging when its frequently hitting over 2000hz at 11psi.. Also the timestamp column thing I mentioned earlier.. the Time should include seconds as one field, not two..

[laakness]

LOL, correct. There is that problem I believe I hit 9 million horse on one of my logs with that setting. I'm a damn space shuttle.

Anyway, that's the only work around I can think of right now to get a correct graph, but other numbers are way off.

[MalibuJack]

LOL, correct. There is that problem I believe I hit 9 million horse on one of my logs with that setting. I'm a damn space shuttle.

Anyway, that's the only work around I can think of right now to get a correct graph, but other numbers are way off.

Best we can do is ask Mad4Evo to combine the time/second columns together in hh:mm:ss.sssss format...

[danl]

Ok this is a calculation that I have found that works, this will do until we work out more about Evo8 Roms. Basically exactly what was suggested above already:

The ecu is doing something similiar to this:

load = Airflow weighted for both temp and pressure = air mass/rev

Load = Air Velocity/Rev * Air temp * barometer (0.00486xBar) * 0.625 (128th bar) / 4

The more I read the more I realize how similar the EVO ecu is to the 1g and 2g dsm ecu’s. Makes sense as the 2g ecu can run the evo 8 just fine with just an ecu conversion harness!

Anyways, I hope this helps. The DSM’s ecu’s do all the fuel and timing lookups in grams/rev. So you are correct in that you are looking at airmass/rev and to answer your question the airmass is probably calculated in grams.

Also to help aide in determining the calculations the ecu makes for each logging parameter here is a list of how it is done on a DSM. EVO’s are more than likely the same. http://www.syndicate.evilsoft.org/proje ... tocol.html

Best we can do is ask Mad4Evo to combine the time/second columns together in hh:mm:ss.sssss format...

That is a must IMO. If you do it like that then people can mess with the data in Excel and do all kinds of wonderfull things!

Found another issue you probably already know about... the Airflow reading (presumably MAF HZ) never goes over 1603.95.. Kinda puts a hurting on my logging when its frequently hitting over 2000hz at 11psi.. Also the timestamp column thing I mentioned earlier.. the Time should include seconds as one field, not two..

Once again this is a problem in DSM's as well as was previously mentioned. Its not that the ecu can't measure above 1600Hz, its just that it can't spit it out. THe solutions for DSM's was to divide the Hz number in two and put the value in a different memory location. Then you set the logger up so that it multiplies teh number by two at the logger then displays it. You have to change some code in the ecu ROM to do this.

[danl]

Here is the way most DSM loggers organize the information. I've also included a graph made by just opening the data up in excel and creating a graph real quick. As you can see, if you organize it in this way people can simply open up their favorite spreadsheet and make their own graphs. No need for a seperate log viewer. Also, you can copy and paste the data on BBS and visually to a human even the raw data is extremly useful.

Also note that my injector PW was capping out as I raised the boost past 24psi. This is a problem in the DSM code that was meant as a saftey feture. Unfortunatley it will blow up a motor if not removed. This is something for the coders of ecuedit to notice!

[c-spec]

The next logical step I think is to add the ability to select which items to plot instead of having a finite list. Personally I like to look at Ignition Timing, TPS, O2 Voltage, and Knock Sum for tuning, and also graph the fuel trims and other sensors for review occasionally.

+1

[Evo4Mad]

Don't worry EvoScan will have this all built in, in the next release.

The next logical step I think is to add the ability to select which items to plot instead of having a finite list. Personally I like to look at Ignition Timing, TPS, O2 Voltage, and Knock Sum for tuning, and also graph the fuel trims and other sensors for review occasionally.

[DataLogLab]

Hi folks.

Just spotted the Data Log Lab definition files on this thread, and the request for changes to the timestamp. The least ambiguous format is a simple second count from the start of the log, with decimal places , e.g.

0.023
0.134
0.167
...
123.678
124.002
...
2476.123
2476.169

etc.

If anyone's having problems with the definition files, just fire them through to support@dataloglab.com and we'll take a look. The easiest way of checking and tweaking the files is to load them using the Load button in the File Locations and Power Graph tab of the Options window. Any error messages displayed should have the name of the column that there's a problem with.

When you load a data log, any columns that Data Log Lab doesn't recognise are recorded here:
C:\Program Files\DataLogLab\Program\Universal Load Unknown Columns.txt
(assuming standard install location)

Any help required, just ask away.

Glenn
Data Log Lab Support
www.dataloglab.com

[mad_viii]

for now you can awk columns 3 and 4 together to get hh:mm:ss.ssss format on the time.

gawk -F , -v OFS=, '{print $1,$2,$3":"$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16,$17,$18,$19,$20,$21,$22,$23,$24,$25,$26,$27,$28,$29,$30}' EvoScanDataLog_2006.07.07_09.28.55.csv >>EvoScanDataLog_2006.07.07_09.28.55.fixed.csv

I'm sure there is an easier way than listing out each of the fields but this works.

[MalibuJack]

for now you can awk columns 3 and 4 together to get hh:mm:ss.ssss format on the time.

gawk -F , -v OFS=, '{print $1,$2,$3":"$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16,$17,$18,$19,$20,$21,$22,$23,$24,$25,$26,$27,$28,$29,$30}' EvoScanDataLog_2006.07.07_09.28.55.csv >>EvoScanDataLog_2006.07.07_09.28.55.fixed.csv

I'm sure there is an easier way than listing out each of the fields but this works.

It doesn't quite work.. For some reason Data Log Lab doesn't handle the date format quite right.. However I came up with a fix...

Check out EvolutionM's EcuFlash subforum for the DataLogLab thread I started if you need more info...

I used Laaknesses Definition file for DLL, and created a little app that you drag and drop the file into, creates a fixed file that has the delta seconds from start.. should work fine..