Page 2 of 2 FirstFirst 12
Results 21 to 35 of 35

Thread: Saab 2.8T tuners, lets chat!

  1. #21
    Advanced Tuner
    Join Date
    Jul 2018
    Location
    Texas
    Posts
    214
    GM Look up tables. Base 2 - Eight 0's and 1's= 8 bits= 1 byte with 255 combinations/counts. 0 - 256 counts. 5.0 volt scale with 255 increments = @ 0.02 volts per count. 1 count = 0.02 volts , 128 counts = 2.50 volts. 256 counts = 5.0 volts. This can be applied to any GM sensor. Ex: 02 sensor 0- 1,000 mv - 128 counts = 14.7 AFR = 450 mv. 0 counts = 0 mv , 128 counts = 450 mv, 256 counts = 1,000 mv. MAP, MAF, BARO, etc. all based off of 5.0 volt scales. 0.0 - 5.0 volts. 0 - 256 counts. If your into computers. Ex: RAM 0,2,4,8,16,32,64,128,256,512,1024,2048 …..
    Last edited by DGS; 02-03-2019 at 02:14 AM.

  2. #22
    Tuner in Training
    Join Date
    Nov 2016
    Posts
    13
    Cobalt. Thanks man.
    Couple of things.
    The Torque to Load table doesn't limit me to a value of 255, It's set (Torque % request) X (RPM) And I think I can go to like 1500 some odd for the cell value, so it lets you look for crazy high numbers.
    Torque Max load value.PNG

    So I do have a larger MAF housing that I'm waiting to find out how to scale it properly without ability to read Hz just voltage

    Commute in this morning.
    commute in 2-1-19.hpl

    This was from yesterday prior to zeroing out the stock correction.
    maf fail 2.hpl
    Last edited by 2jzgte; 02-03-2019 at 08:53 AM.

  3. #23
    Advanced Tuner
    Join Date
    Jul 2018
    Location
    Texas
    Posts
    214
    Something like this could help you out. Build graph using MAF voltage with 5.0 volt scale. You could put in more increments to match the MAF/voltage lb/hr in the tune. Use wideband in place of stft+ltft.
    Attached Images Attached Images
    Last edited by DGS; 02-03-2019 at 11:45 AM.

  4. #24
    Senior Tuner cobaltssoverbooster's Avatar
    Join Date
    Dec 2008
    Location
    Nevada
    Posts
    4,452
    Quote Originally Posted by DGS View Post
    GM Look up tables. Base 2 - Eight 0's and 1's= 8 bits= 1 byte with 255 combinations/counts. 0 - 256 counts. 5.0 volt scale with 255 increments = @ 0.02 volts per count. 1 count = 0.02 volts , 128 counts = 2.50 volts. 256 counts = 5.0 volts. This can be applied to any GM sensor. Ex: 02 sensor 0- 1,000 mv - 128 counts = 14.7 AFR = 450 mv. 0 counts = 0 mv , 128 counts = 450 mv, 256 counts = 1,000 mv. MAP, MAF, BARO, etc. all based off of 5.0 volt scales. 0.0 - 5.0 volts. 0 - 256 counts. If your into computers. Ex: RAM 0,2,4,8,16,32,64,128,256,512,1024,2048 ?..
    I understand the 256 format but i didnt know it was being applied in this way. I only know just enough to understand there is gold there, but honestly im totally lost once i look at it. Thanks for that info, ill try to remember it.

    Quote Originally Posted by 2jzgte View Post
    Cobalt. Thanks man.
    Couple of things.
    The Torque to Load table doesn't limit me to a value of 255, It's set (Torque % request) X (RPM) And I think I can go to like 1500 some odd for the cell value, so it lets you look for crazy high numbers.
    Torque Max load value.PNG

    So I do have a larger MAF housing that I'm waiting to find out how to scale it properly without ability to read Hz just voltage

    Commute in this morning.
    commute in 2-1-19.hpl

    This was from yesterday prior to zeroing out the stock correction.
    maf fail 2.hpl
    So one thing with large turbo applications that causes huge issues is location. Your mafs best location is about 12-8" in front of the TB in a special charge pipe that expands to about 3". In the LNF its common to have 2.5" CP but the last cold side pipe will have a 3" section modification to house the maf sensor. Typically the atmospheric bov causes massive issues and so the maf has to be last in the chain so it reads only air getting to the engine. I have seen recirculation valves cause issue with maf sensors depending on location of ports and amount of air getting recirculated.
    Tuning the maf curve is simple. Go into your scanner and make a histogram with maf signal voltage as the x-axis. Plot the same editor values for the scanner. Now plot your afr error against this...or from what i can see you likely have factory wideband sensors so a simple stft error will work fine there too. *** Check for trim error still working when in pe, if it does then you have oem widebands.

    Yes i know the T2L table can go higher than 255. I am trying to say that initially we thought it was a map sensor related thing because the oem computer was limited to 255 kpa in the E69 platform. This is not the case. You can exceed this value but it can make the calibration sensitive. All im telling you is to make runs and verify your not getting performance fluctuations past 255. This stability limit at 255 likely has more to do with coding in reference to a bin that is registered to the limit of a byte (256) which has a 256 value range that starts with 0. this means you can plot from 0 - 255. Ill be honest ive gone over 255 before and didnt see a huge problem. it just made it harder for me to control boost. it kept waving around. i was still able to get over 30 psi of boost with it set to 255 and using wastegate controls to set the pressure level. This is why i would get people to drop into feature request and ask for gate controls to be accessible. Having gate controls allows you to keep the air load request down which allows for smoother tables. It just so happens that the ecu tends to operate smoother overall when these values are kept at the lowest request possible to attain pressure. Since you dont have the right features its hard to explain.

    Due to lack of gate control, the only thing you can do is say ok if the gate duty cycle is fixed due to limited access, then the only way to get pressure is to increase the air load demand.
    2000 Ford Mustang - Top Sportsman

  5. #25
    Tuner in Training
    Join Date
    Nov 2016
    Posts
    13
    Yes the vehicle has very fast acting wideband actually.
    Ok, I read many things that stated you shouldn't/couldn't use MAF volts, which I thought was weird. But good to know I can use the standard corrections against MAF volts for coarse adjustment. Then use the MAF correction to fine tune.

    So since the car has an accurate wideband, what method should I use? the ST+LT seems kind of odd, doesn't that essentially increase the number by which I'm modifying the table?
    Should I not just use one or the other? Or like you said the percent error?

    So the way the intake is made, is that the piping is very similar to the diameter of the stock piping, and it is using the factory plastic MAF housing. It's in a near identical placement to stock as well. The only thing is that I don't know what to do. The turbo I'm sure is pretty violent in it's recirculation, because when I INITIALLY tried starting the car once the work was done, I put the MAF on a 3"to 4" reducer right to the turbo and the car wouldn't stay running. It literally fell on its face. I had no idea what was wrong, since the car ran PERFECTLY with no MAF. It wasn't until I put the father away that I had less issues with it.

    I get MAF performance range errors up top somewhere around 38# per minute most times on the old intake setup. The new setup is the same piping just with a much larger more appropriate air filter.

    And fixed a small boost leak.
    Now the car feels like it's making a TON more power, spooling as fast as I ever expected such a large turbo to spool on a tiny engine with huge long tube headers.
    But now MAF is off (Reads high) 15% pretty much everywhere except at idle where it reads very lean... I don't really get it. That sounds like post throttle body leak to me, but, no leak found... The system is crazy tight.

    I just can't imagine a MAF needing such obnoxiously high adjustment. But you guys tell me, I dunno.
    I'm afraid if something goes back to normal, and I've pulled 15-16 or more percent of trim out of the MAF table that number 1 it lowers what the MAF will actually read... Reducing it's ability to fuel the car and I'm out of MAF already.
    And 2 of course that spontaneous lean cuisine if something goes back to normal. :Shrug:


    So, Im not sure what to do. I have a recirc in the most OE possible format. Right after the turbo outlet right before the turbo inlet with a fast responsive GoFastBits model.

  6. #26
    Senior Tuner cobaltssoverbooster's Avatar
    Join Date
    Dec 2008
    Location
    Nevada
    Posts
    4,452
    the maf can change wildly based on location and tubing diameter. turbulent influences dont help the case of making a maf more consistent.
    i would say that the offset isnt so much an issue. as you tune just make sure the maf calibration keeps its general stock shape. you dont want it to get out of shape, should be smooth.
    i tune with stft only on the E69. This is due to the ltft not functioning in a desired way.

    if the sensor fails then the car will run like ass. none of the ones i tuned ran good when something failed out. it doesnt default into race home mode....its limp home.

    i honestly believe you are having a location issue or a sensor issue. the only way to know is to test both.


    in regards to being told not to adjust the voltage....what do you think maf hz is? Its still a calibration axis. If the car was stock then i probably would leave the volts alone because the sensor is calibrated to an airflow setting to match the stock maf housing. the base table would be enough on its own to make adjustments. By moving the housing in size and location you change the airflow model through the sensor. at this point its worth modifying the voltage scale so the ecu knows the exact air density through the maf right off the bat. how much air is entering the engine per maf volt.

    Advanced Rant:
    Take this into consideration or thoughts. The oldest methods of tuning are linked back to the very base of testing a sensors calibration. If a sensor is calibrated by setting in a known environment and then adjusted until reading properly then we can simulate that with the vehicle. The sensor is installed in an unknown environment, but in our case we know that per application the airflow model at steady state can be considered constant. Because of this, you can say that the calibration of the sensor can be calibrated based on the error of the combustion burn process. This isnt a test cell so there will be inherent error, but it will be close enough that 1) we can let the computer handle it with its trim systems and 2) the event is over in ms of time and if the error is small enough then we dont exactly care because we dont currently have a way to predict every possible event and outcome.
    I use this method to set fuel injector flwo data per application. first i disable all the multipliers and modifiers so i am left with only the ecu being run on base injector flow rate data. if i tune this in to just about dead nuts and then force command 10% more fuel then i should see a -10% error where the ecu wants to remove exactly what i just forced commanded. If i dont get an error equal to my force commanded then i can assume the base injector flow rate is in error and re-calibrate this data until the returned error matches the forced error. Here with your maf sensor the same thing applies. if i take away all of the fueling modifications and the base data is known good solid data, then i can say any error induced due to the maf can be calibrated on the maf.

    Now what do i do in reality? I take the stock maf and record fuel error. i make adjustments using copy and paste multiply % to get my new test data. before running it i tell the software to smooth the entire range. it is of upmost importance to keep this data smooth.
    once i have this as close as the error will let me come in while retaining the smoothed shape, i switch over to the base table and take up any 4th dimensional errors with this table. I never really focused on making this table super smooth but solid state testing will get it damn close. The maf calibration is a 1-D table (flat). In addition to this you get a maf base table that is 3-D (multi-elevation model). adding these together you get roughly 4 dimensions of calibration which is hell of a lot of detail.
    Over testing i have come to my own rules and its goes as follows: 1) make the base calibration of any sensor or any system as smooth as utterly possible with the minimum allowed error to retain a smoothed form. 2) from here the multipliers and corrections added to this can be slightly less refined as long as the results remain positive. (this does not mean i will let it be peaky and illshaped) there is almost always more than 1 correction factor so how well you blend all these correction factors into a smoothed final result will define how good you really are.
    2000 Ford Mustang - Top Sportsman

  7. #27
    Tuner in Training
    Join Date
    Nov 2016
    Posts
    13
    Thanks.
    Again, I'm completely aware of standard testing methods, versus application. When to use something as a basis and when to deviate based on knowns.
    So, I am using the stock MAF housing with the stock MAF.

    I had gone out to refine my intake once again.

    The MAF housing is in a factory location. The factory post MAF tubing drops down to like 2.25 inch going into the factory turbo, which is obviously not going to work since the turbo is a 4" inlet
    So I used all 3" tubing which is pretty equivalent to the ID of the factory MAF housing, and slightly smaller than the factory intake bits pre-maf, but they don't make 3.1x "ID tubing
    But it all goes through the stock MAF housing and location. It's in the only straight section of the tubing as is stock. It might work better closer to the air filter actually, farther away from the recirc valve.

    So we'll see how it goes on my way home this evening, I've reset all of the trims etc.
    I'd feel much better with trims under 8% I don't mind making that kind of an adjustment to the MAF.

    Thanks again for the clarification about the Hz vs voltage. That is also what I had figured but, there are several "stickies" which have this "Hz ONLY" mindset... So.

  8. #28
    Senior Tuner cobaltssoverbooster's Avatar
    Join Date
    Dec 2008
    Location
    Nevada
    Posts
    4,452
    i checked all physical items possible on my yukon and had to apply a 23% global correction in order to get it to adjust. sometimes you cant be afraid to make the change.
    good on you for testing different locations. as long as you adhere to the rules of basic calibration then you shouldnt be worried like you are.
    2000 Ford Mustang - Top Sportsman

  9. #29
    Tuner in Training
    Join Date
    Nov 2016
    Posts
    13
    So, I found a very minor leak, and made some more changes to the intake.
    Brought me into more normal ranges. Not great, but, pretty respectable.

    CSSOB, I assume my coils are dying, but once I hit boost over the MAP value limit car breaks up. Doesn't feel like a hard cut and doesn't go crazy lean or anything but it does break up.
    Plugs are relatively new and gapped properly.
    So man, I really wish I could just properly scale a sensor with this software

  10. #30
    Senior Tuner cobaltssoverbooster's Avatar
    Join Date
    Dec 2008
    Location
    Nevada
    Posts
    4,452
    Whats Gap? at 25 Psi i ran my cobalt ss at .024-.026" gaps
    ive ran them down as low as .019" by 35 psi
    2000 Ford Mustang - Top Sportsman

  11. #31
    Tuner in Training
    Join Date
    Nov 2016
    Posts
    13
    Turns out I think the "blow-out" or whatever might be a cut when exceeding the MAP sensor. So back to, needing the MAP sensor calibration ability to tune, again.
    I was running .711mm Gap

  12. #32
    Senior Tuner cobaltssoverbooster's Avatar
    Join Date
    Dec 2008
    Location
    Nevada
    Posts
    4,452
    it very well could be its unhappy
    2000 Ford Mustang - Top Sportsman

  13. #33
    Potential Tuner
    Join Date
    Jul 2022
    Posts
    1
    I have a 2009 9-3 Aero 2.8 XWD does anyone have a stage 1 or so tune file they have tested on a stock car like mine? I am open to pay someone for it.. I have HP tuners... for another car figured I would see what is out there. Or if someone can send me a link to any tune file libraries.
    Thx

  14. #34
    bump a dead thread but i recently got a 06 9-3 2.8 with the A6 and would just like to know any tips on just increasing the horsepower some on a mainly factory setup. any suggestions on the approx spark adv, afr pe to set, as well as safe boost increase on either 93oct or even e47? thanks in advance!

  15. #35
    so i licensed my 06 aero A6 and noticed the lamda table is mainly all 1.0 values meaning the AFR the ecu is trying to command 14.7 even at WOT? am i missing something that would then retard timing when it sees 14.7 at full throttle? lastly i noticed the trans tables are all blank?