Results 1 to 13 of 13

Thread: Advanced Filtering

  1. #1

    Advanced Filtering

    Does anybody know if there is a way setup a filter on chart ve afr error%. I am trying to filter out transition data so say if throttle position changes more than x% in 500ms filter it out.

  2. #2
    Senior Tuner mowton's Avatar
    Join Date
    Mar 2007
    Location
    Florida
    Posts
    3,773
    [50090.156]>50 AND [50090.156.avg(500)]>50 AND (50090.156.avg(-500)]>50

    So TPS has to be >50% for 500ms and remain 50% for at least 500ms more for filter to result in True (show the data). Filters on the positive going pulse and negative edges. The time isn't exact so you may have to play with the values if you are looking or exactly 500ms but the TPS percents will be. I also use this approach for Commanded Lambda to filter that lean glitch we always see when the PE engages. Throw in a Fuel trim cell at 0-15 and the Commanded Lambda of 1.0 and .85 and you have all filtering you should need for consistent street tuning. I also substitute the Acc Pedal en lieu of the TPS% as the PCM can sometimes take control of it :-)

    Ed M
    Last edited by mowton; 04-13-2016 at 10:11 PM.
    2004 Vette Coupe, LS2, MN6, Vararam, ARH/CATs, Ti's, 4:10, Trickflow 215, 30# SVO, Vette Doctors Cam, Fast 90/90, DD McLeod, DTE Brace, Hurst shifter, Bilsteins etc. 480/430

    ERM Performance Tuning -- Interactive Learning ..from tuning software training to custom tunes
    HP Tuners Dealer- VCM Suite (free 2hr training session with purchase), credits and new Version 2.0 turtorial available
    http://www.ermperformancetuning.com
    http://www.facebook.com/ERMPerformanceTuning

    [email protected]

  3. #3
    ok getting somewhere now just trying to wrap my head around it that seems like it would work good with anything over 50% throttle. What I am trying to do is basically emulate the data you would get from a load bearing dyno so i dont have to worry about trying to hold in a cell while also trying to concentrate on driving. Maybe hptuners can't do this.
    Last edited by richsoucie; 04-13-2016 at 10:55 PM.

  4. #4
    so say there is more than a 5% variance over 500ms filter that data out

  5. #5
    [17.156]>2 AND [17.156.avg(500)]=[17.156] where 17.156 is tps. So that would require the tps to be steady for 500ms to show the data. Now how would i go about adding a variance so that [17.156]+-5% is also shown

  6. #6
    well that escalated quickly

    [17.156]>2 AND [17.156.avg(1000)]=[17.156] OR [17.156.avg(1000)]<[17.156] + 1 OR [17.156.avg(1000)]>[17.156] - 1 AND [17.156.avg(-1000)]=[17.156] OR [17.156.avg(-1000)]<[17.156] + 1 OR [17.156.avg(-1000)]>[17.156] - 1

  7. #7
    [5.242]>170 and [17.156.slope(500)]<1 and [17.156.slope(500)]>-1 AND [17.156.slope(-500)]<1 and [17.156.slope(-500)]>-1

    Where 5.242 is coolant temp and 17.156 is tps
    If I am understanding what I did correctly than anything under coolant temp of 170 will filter out and any change in the tps +-1% over 1 second will be filtered out
    Last edited by richsoucie; 04-14-2016 at 12:30 PM.

  8. #8
    Moderator
    Join Date
    Mar 2014
    Location
    Raleigh, NC
    Posts
    6,347
    Do you really get much data when filtering it down to +/- 1% per second? I haven't checked over the formula, just curious about the concept.

  9. #9
    [5.242]>170 & [2114.156.slope(500)]<2 & [2114.156.slope(500)]>-2 & [2114.156.slope(-500)]<2 & [2114.156.slope(-500)]>-2

    5.242 is coolant temp and 2114.156 is accelerator pedal position

  10. #10
    3.1.png

    Left is with no filter Right is with [5.242]>170 & [2114.156.slope(500)]<2 & [2114.156.slope(500)]>-2 & [2114.156.slope(-500)]<2 & [2114.156.slope(-500)]>-2

    not so much getting data as it is getting good data. You can get crap data all day long and have 1000's of cell hits but be chasing your tail trying to tune ve on the street. You can actually take a high positive or high negative number down to 1 or 2 just blipping the throttle thats what we are trying to get rid of
    Last edited by richsoucie; 04-14-2016 at 04:22 PM.

  11. #11
    Moderator
    Join Date
    Mar 2014
    Location
    Raleigh, NC
    Posts
    6,347
    I'm fully aware of the difference. I was just asking if it was working well for you or not. It's something I've been meaning to spend some time on but haven't had a chance yet.

  12. #12
    Advanced Tuner
    Join Date
    Aug 2014
    Location
    Central, LA
    Posts
    737
    Quote Originally Posted by mowton View Post
    [50090.156]>50 AND [50090.156.avg(500)]>50 AND (50090.156.avg(-500)]>50

    So TPS has to be >50% for 500ms and remain 50% for at least 500ms more for filter to result in True (show the data). Filters on the positive going pulse and negative edges. The time isn't exact so you may have to play with the values if you are looking or exactly 500ms but the TPS percents will be. I also use this approach for Commanded Lambda to filter that lean glitch we always see when the PE engages. Throw in a Fuel trim cell at 0-15 and the Commanded Lambda of 1.0 and .85 and you have all filtering you should need for consistent street tuning. I also substitute the Acc Pedal en lieu of the TPS% as the PCM can sometimes take control of it :-)

    Ed M
    Could you clarify what you meant by the portion that I've placed in bold? Do you mean you add a filter for fuel trims and commanded lambda?

    Sorry my brain isn't functioning well enough right now I guess!

  13. #13
    Tuner in Training
    Join Date
    Apr 2016
    Posts
    31
    Quote Originally Posted by mowton View Post
    [50090.156]>50 AND [50090.156.avg(500)]>50 AND (50090.156.avg(-500)]>50

    So TPS has to be >50% for 500ms and remain 50% for at least 500ms more for filter to result in True (show the data). Filters on the positive going pulse and negative edges. The time isn't exact so you may have to play with the values if you are looking or exactly 500ms but the TPS percents will be. I also use this approach for Commanded Lambda to filter that lean glitch we always see when the PE engages. Throw in a Fuel trim cell at 0-15 and the Commanded Lambda of 1.0 and .85 and you have all filtering you should need for consistent street tuning. I also substitute the Acc Pedal en lieu of the TPS% as the PCM can sometimes take control of it :-)

    Ed M
    Will this math work for any filtering needs? I'm thinking this is a good application to cut down on the anomalies for many things!