Page 1 of 3 123 LastLast
Results 1 to 20 of 53

Thread: Struggling to find my error with gt350 throttle body

  1. #1

    Struggling to find my error with gt350 throttle body

    Im not sure what im missing, i have transferred the gt350 TB data over to a stock 5.0 tune and it isnt working properly. very mild startup like the TB isnt opening and part throttle medium load IPC errors start to crop up with surging being felt. also finding with key on, engine off, 100% pedal= 85* tb with stock tune. flash the frpp2 tune and im back at 90-92*.

    this is my stock tune and then the frpp2 tune that i bought from ford performance directly. frpp tune works great, stock tune with same changes and adjustments as frpp not so much. did compare the extra maps that we cant see in the frpp2 tune with the 2017 gt350 file in the repository and transferred those values over for start up TB angle and etc angle filtering.

    wanted to have this go flexfuel but im struggling to find what im missing about getting the gt350 tb to work on a stock OS.

    any suggestions or is it needed to tune the tq tables at part throttle to get this TB to work with out ipc error/surge.



    STOCKOS-GT350TB-FRPPTIMING.hpt
    Current FRPP2 with changes.hpt

  2. #2
    Advanced Tuner
    Join Date
    Feb 2017
    Location
    TX
    Posts
    761
    Need a log.
    Include Throttle Angle, Throttle Angle Desired, Throttle Area a long with all the "Torque" channels and IPC wheel error. And your regular channels.

    I could look at that Throttle Body Model and make corrections all day, I'd rather see the area your having trouble with and work on that.

  3. #3
    logs attached, i went back to stock tune and only changed the TB and Maf to rule out any mishaps on my adjusting. I dont have a normal list of things i log, i typically try and focus on one area and start scaling the polling speed to reduce least important. if im missing anything let me know and ill add and re-do.
    STOCKOS-GT350TB.hpt

    i have noticed FRPP and roush like to rescale the inverse TQ axis. on the NA tunes they dont actually rescale the inverse table itself, its left stock. so from stock the bottom row at 516 tq, to frpp at 700 tq.
    This does not work to simply rescale the stock tune to match the frpp one, nasty rev hangs etc... there are tables we cant see that eric has advised might be added in the future. fingers crossed he gets time. Im guessing there has to be some other area they are accounting for the change in the axis on the inverse tables.


    also did a cold start with each tune as well if it helps.

    frpp startup.png stock OS startup.png
    frpp2 startup only.hpl
    stock OS GT350 TB Startup only.hpl




    main log, largest errors at 1:45 / 2:04 / 2:26 / 3:14 / 3:38

    ipc error.png
    stock OS gt350 TB drive.hpl
    Last edited by Grim5.0; 03-31-2020 at 11:03 PM.

  4. #4
    Advanced Tuner
    Join Date
    Feb 2017
    Location
    TX
    Posts
    761
    There are different ways to tune out torque errors. I know every one thinks using a known tb cal will be problem free but thats not always true. You might have missing tables but your file has the necessary parameters.
    Torque errors are caused by improper calculations with airflow, the ecm is processing data while the cars running and the live data doesn't match the tune.
    Airflow issues can be fixed through the tb model, torque/inverse tables and even driver demand. We know your tb model is a true model, making proper adjustments can still solve this.

    Second and easiest is adjusting driver demand. Find a torque error in your log and compare etc torque request and engine brake torque during the torque error. If etc torque request is less than brake torque, increase driver demand in the correct cells to request more torque than engine brake torque. You can only do so much for torque errors using driver demand and you could so no improvement.

    Third option and what I'd call the "correct" solution is your torque/inverse tables. The stock tables aren't agreeing or calculating correctly with the parameters being calculated using the new tb model. Torque/inverse need adjusted to run with the known tb model.

    I still need to check your files and new log, minor changes could clean up the errors. I'll upload some pics.

  5. #5
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    MAF/MAP sensor
    airflow>load>torque

    driver demand/ TB model
    desired torque> desired load> desired airflow

    They work inverse of each other. You would have to skew the torque model to correct the TTL and LTT.

    You changed your TB model which calculates your desired airflow from a predicted angle, and adjust blade angle when its out of whack with what the MAF is reporting. Its why you only see like half a second of IPC, its a very fast adjustment of the blade angle.

    The driver demand is the appropriate place to correct this. Its the IPC the max/min limits(25-30ftlb window), that gets in your way. Desired torque changes scheduled/indicated torque directly proportional with the "idle torque". This is what the IPC watches. One of the first guys to install a GT350 mani and TB(higgs) pointed this out and said to tune it he just set them way out of the way to get it dialed in. The clip/add can get in the way too.

    The window, tightest right where everyone experiences the "surge" that they can't fix.

    IPC window.PNG
    Last edited by murfie; 04-01-2020 at 02:47 AM.

  6. #6
    Advanced Tuner
    Join Date
    Feb 2017
    Location
    TX
    Posts
    761
    The throttle body model looks nasty, I know it's a factory calibration but damn. Driver demand has a smooth increase as throttle/vacuum increases. The throttle body model is not.
    This is the main area of issue 28 to 35*
    Badtbmodel.PNG

    Increasing Driver Demand in this area and smoothing the 3D model could help. Works on first gens only so by so much. s550's are more torque based and some of the errors, engine brake is calculating more than schedule.
    This was for DD4, could do the same for 3,4 and 5.
    dd4.PNG

  7. #7
    So far im striking out, great opportunity to learn though. If im understanding the TQ pids in the logs, Desired Brake Torque is my rt foot from the DD tables and then ETC Torque request is the final output of the DD/throttle calculation. at idle this seems to include the idle reserve TQ, if i increase to idle reserve TQ im finding this increases too beyond the desired tq. the gt350 uses 7 ftlbs vs 4ftlbs for the 5.0. if thats incorrect please advise

    tried several variations of DD adjustments at the specific pedal/rpm points and the surrounding cells. Increased the cell 10% and then smoothed around. Ended up doing this 3 times and still getting large ipc errors. did try the DD suggestion pictured above as well. unfortunately similar results
    i just realized i did not follow exactly as instructed on these tables though by increasing the DD above EBT, ill give that a go tonight and get a fresh log.

    would mapping EBT against the DD map axis work to repopulate the DD?

    Higgs post was my first realization there was work ahead! Most of the threads on here for this TB end with someone saying to copy the gt350 TB data. that did leave me feeling optimistic it would be that simple. i now know the reason those threads end a short time later!
    Im not getting into limp mode yet, just the bucking sensation and spikes in IPC, do i need to move the max/min tables as higgs did? or is that when you start getting put into limp mode?


    Im also finding the cranking time has increased quite a bit on a hot restart too. sounds like its not going to catch. I increased the etc throttle angle to match the colder ect, doesnt seem to do anything during cranking though!
    TB crank angle map.pnglong crank time.png

    Thank you guys for what you have shared and the suggestions!
    Last edited by Grim5.0; 04-02-2020 at 07:56 PM.

  8. #8
    Heres log with EBT entered as DD values during IPC error events. Then roughly smoothed, rather large jumps from stock and not the greatest smoothing between pedal positions. I went through a log, ploted each large ipc error and the EBT indicated into the DD and then went back to smooth the area around each point.

    Im running into tq control/oscillation a lot now. Im finding that if the pedal position and rpm are constant and without ipc error, timing is in the floor and it feels like im pulling a parachute. IPC errors are showing EBT eclipsing the tq tables sched/indicated. Is the DD still the direction to go?

    IPC errors get worse the longer its ran which seems odd. cold start near perfect, minimal ipc errors until nearing op temp.

    Drive.png

    DD tables.png

    Drive.hpl

  9. #9
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    WHAT YOU NEED TO KNOW:
    Desired torque and ETC torque are the same, as they come from the same table values.

    Scheduled torque and indicated torque are the same, as they come from the same table values.

    Engine brake torque is derived from schedule/indicated torque. spark/fuel/ambient conditions/losses/ idle torque and probably other things i'm not mentioning are factors in arriving at this torque value. Its accuracy is dependant on how accurate your calibration is.

    When I say "Idle torque" Im refering to the difference between your requested torque and your scheduled torque. This is usually 50-70 ft lbs depending on torque model/ idle airflow. when at idle torque request is 0, engine brake torque matches so its 0, and schedule/indicated is this idle torque value. That difference needs to be constant at all torque request, with only small differences in large throttle changes.


    THE PROBLEM THE NEW TB HAS CAUSED:
    Desired/etc torque and engine brake torque need to be the same. feedback loops on both the dd table and TB model work to keep them the same. Above "WOT start" it only uses the TB model to keep this relationship the same, on ecoboost this is raised way up and not true in stock form. This can be lowered, eliminating one of the feedback checks, to make it easier to dial in. Throttle angle error is the error from this feedback loop. If it was left open loop, your pedal would not feel like its attached to your TB. You need to use your throttle angle error to correct your DD table to match what the TB model thinks it should be. Do not skew your known good models to accomplish this. Yes you can apply the error in those and fix it. The dd table is the torque at the rear wheels, and the TB is saying that is wrong, so fix it in the right spot. You get wrench light in situations when the TB model says its flowing enough air to make 500ftlbs of torque, but the pedal and RPM say it should be at 200Ftlbs of torque, throttle must be out of control shut it down.



    THE SECOND PROBLEM:
    IPC error is seperate from all this. IPC watches scheduled/ indicated torque as they are usually very constant(just look at the torque model). When this gets out of line from the max/min tables, IPC steps in and corrects. The problem comes from the loop correcting to what the TB is reporting causing the torque request to move the scheduled/indicated torque value out of where IPC thinks it should be. This needs to be moved out of the way while you correct the throttle angle error. Then it can be moved back to where the scheduled/indicated torque moved to.
    Last edited by murfie; 04-03-2020 at 01:03 AM.

  10. #10
    Senior Tuner veeefour's Avatar
    Join Date
    Oct 2016
    Location
    Poland
    Posts
    1,712
    GT350 TB needs some Speed Density changes to work properly. TB model(effective and predicted) is perfect from factory.

  11. #11
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    If you don't have a MAP sensor hooked up and logging in the scanner, Ignore the SD, you are still NA and nothing in there will make a big difference. The math(Regression analysis) of how it works and correcting it, is far over the average persons head, ability, and time constraints. With out the understanding of the math, any process or calculator made so far, will just make you quit trying and purchase a tune from one of the big names quicker.

  12. #12
    Senior Tuner veeefour's Avatar
    Join Date
    Oct 2016
    Location
    Poland
    Posts
    1,712
    I do not want to share too much Murf but I fixed a zillion GT350 TB tunes just by adjusting SD model without even hooking any MAP sensor.
    Trust me it's that easy, not sure why and what is special about this TB but it is special and super funky to tune.

    PM or email me should you want to, I will tell what I know.

  13. #13
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Share too much... or share anything for that matter. Seriously you think you will lose that much business?... acting like its covid19 you are containing.

    Then really how useful is this information, if you don't know why its fixing the issue, and what else your changes are effecting. If the SD alone could fix it, the MAF transfer alone could fix it, think about that one.
    Last edited by murfie; 04-03-2020 at 04:16 AM.

  14. #14
    Advanced Tuner
    Join Date
    Feb 2017
    Location
    TX
    Posts
    761
    Disable the oscillation switch or shuffle switch if you haven't.

    These are all different methods or techniques to accomplish the same goal. It all comes down to tuning a good calculation.
    Your TB model, torque request, engine brake torque and scheduled torque are not calculating properly.

    Those are some good explanations Murphie.
    In Murphies paragraph beginning "engine brake torque," explains the cause of your torque errors. In your datalogs found with your torque errors, you will see engine brake torque calculating close to the same value or more than scheduled torque.

    You have several solutions like i said, adjusting driver demand will only do so much.
    You have two ways to approach. You can adjust the torque/inverse table to calculate correctly with your throttle body model.
    Or adjust the throttle body model to calculate correctly with your torque tables. They advise not to and I'm not advising you to but I've found as long as the both throttle angle predicted and throttle area tables calculate correctly together. Adjusting the models where they need to be according to your datalogs will get rid of your torque errors and run perfect. You may only have to change cells of error by a few tenths to achieve a better relationship with your torque/inverse tables.

    Murphie also defines throttle angle error. Which is A+.
    My definition and understanding, throttle angle error is torque request in ft lbs calcuated into throttle angle via TB model and subtracted from throttle angle actual.
    You can adjust driver demand or the throttle body model based off throttle angle error.
    Driver demand is ETC Torque Request.
    Throttle angle predicted and ETC area can adjust Engine Brake Torque to a lesser value than scheduled.
    Torque/inverse can be adjusted to be greater than engine brake torque.

    Ideally you'd want adjust torque/inverse. Your torque errors happen when your car runs best drivability, blending from mp7 into mapped points 20 and 21 i believe. Torque table 7 has more torque scheduled in the same load rows. When your car blends into 20/21 with no change in throttle, your scheduled torque decreases below engine brake torque causing ipc errors and throttle oscillation. It will take a couple of tunes. Just make sure you're seeing progress, you might have to head the opposite direction or use a different mp.

    Or maybe veefour will share that easy fix.

  15. #15
    Senior Tuner veeefour's Avatar
    Join Date
    Oct 2016
    Location
    Poland
    Posts
    1,712
    Quote Originally Posted by murfie View Post
    Share too much... or share anything for that matter. Seriously you think you will lose that much business?... acting like its covid19 you are containing.

    Then really how useful is this information, if you don't know why its fixing the issue, and what else your changes are effecting. If the SD alone could fix it, the MAF transfer alone could fix it, think about that one.
    OK boss...

  16. #16
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    If any one cares about converting the angle error to a mass flow rate error, to load error, to a torque value to fix driver demand. Rather than guessing.

    https://patents.google.com/patent/US5526787A/en

    Fords patent on determining desired throttle position. It describes the flow and the math of the throttle body model.

    http://www.eng.auburn.edu/~pjones/ME....B.Heywood.pdf

    This is the paper the patent refers to for the throttle body equations, they are on pages 906-910.


    For you all bad at math, ignore the ideal gas law, critical pressure ratio, and discharge coefficient.

    mass flow = density of air x area x velocity

    dry air density is 4.33373843 ? 10^-5 pound / (inch^3)

    Area is effective throttle area (inch^2).

    Velcoity is 2362.2 x Sqrt((2xETC vacuum [Pa])/ 1.19957467 [kg / (m^3)]) [inch/min]

    lb/min = lb/inch^3 x inch^2 x inch/min
    -------------------------------------------------------------------------------------------------------------------------------------
    (((lb/min divided by RPM) divided by 4 intakes/rev) divided by engine displacement from the general tab) will get you load.
    -------------------------------------------------------------------------------------------------------------------------------------
    Load converts to a torque value from a direct 1.0 load = x amount of torque defined by your torque model.

    If you have 5* of throttle angle error(.035 in^2) (~1.8 lb/min)(.064 load@4250RPM)(23.6ftlb@12-25% throttle), look at your model to see approximately what effective area that is, and from that you can calculate a correction torque value.

    This is where and how much I see your DD table should be corrected, a little bit lower accelerator % and higher RPM than what has been suggested and where you showed your changes. Drivability issues are always going to be the slope from the 0 torque demand to the torque at the accelerator position the issue occurs. This is what defines accelerate, decelerate, or maintain speed. The torque request just needs to be a bit flatter in the higher RPMs, making this slope steeping to accommodate the airflow through the larger TB at lower angles.

    DD adjustments.PNG

    TA error in DD.PNG

  17. #17
    Advanced Tuner
    Join Date
    Mar 2019
    Posts
    350
    Quote Originally Posted by veeefour View Post
    OK boss...
    If you are on here for the sole purpose of keeping secrets in the effort to increase your business you are of no use to anyone. You are doing nothing but to slow down progress. Every time I see you "contribute" to someone else's thread it is a direct effort to confuse forum members, keep people from learning, and prevent people from figuring out how to resolve their problems. Because of this I would never, ever give you any of my business. I would much rather spend my money with someone who genuinely cares and wants to improve the community. This is a forum, it's about sharing information and working as a collective.

  18. #18
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Quote Originally Posted by murfie View Post
    Share too much... or share anything for that matter. Seriously you think you will lose that much business?... acting like its covid19 you are containing.

    Then really how useful is this information, if you don't know why its fixing the issue, and what else your changes are effecting. If the SD alone could fix it, the MAF transfer alone could fix it, think about that one.
    I think you mean "acting like you have the cure for covid-19" since it's quite un-contained at the moment .

    Anyways, thanks Murfie for being the person you are! I tried to give you more rep but I guess it's maxed out coming from me, haha . If we didn't have people like you here, we'd have nothing at all!
    Knock Retard is the reduction or prevention of knock by lowering ignition timing:

    (+) Adding Knock Retard = Reducing Timing. PCM is seeing knock.
    (--) Lowering Knock Retard = Increasing Timing. PCM isn't seeing knock.
    __________________________________________________ ________

    2014 Mustang GT Premium. VMP Gen2R Supercharged with an FTI 3000rpm Converter. JLT, BMR, Steeda, Viking, etc.
    Don't fix it if it ain't broken | Maximum effort gets maximum results

  19. #19
    Senior Tuner veeefour's Avatar
    Join Date
    Oct 2016
    Location
    Poland
    Posts
    1,712
    Quote Originally Posted by B E N View Post
    If you are on here for the sole purpose of keeping secrets in the effort to increase your business you are of no use to anyone. You are doing nothing but to slow down progress. Every time I see you "contribute" to someone else's thread it is a direct effort to confuse forum members, keep people from learning, and prevent people from figuring out how to resolve their problems. Because of this I would never, ever give you any of my business. I would much rather spend my money with someone who genuinely cares and wants to improve the community. This is a forum, it's about sharing information and working as a collective.
    I helped couple of thousands on this forum, free of charge so go the hell.

    On the other hand it drives me crazy that I spent weeks, driving the car around paying 8USD per gallon...and than go to forum share everything free of charge
    so other shops that are hiding under "private" nicks here can grab it for free and make money on my work? You gotta be sucka!!!

    Murf shares a lot and he's a smart guy although that is rather his hobby, but he is often wrong, I'm often wrong as well.
    I stopped giving simple advice's long time ago, after I realized how dark the FORD community is. 70% threads here are BUSINESS related, help me so I can make money.

    And than when I NEED HELP THERE'S NOBODY TO HELP EVEN SUPPORT.

    I don't think so!

    I have a lot of work, my business is doing fine trust me. I don't even need to be here but when I do I make sure to confuse every filthy shop owner asking for a free help LMFAO!

  20. #20
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    I see both of your points and the good have to pay for the bad I guess. If a shop is charging people for tunes then their posts should be about singular problems (not basics) and definitely not asking for full tunes from scratch!

    Personally I'm here to learn and help with what I do know but it's purely a hobby to me as well. I have a single Ford that I tune and it's my own. I've tuned a close friend's Silverado for fun (and at no charge to him) but he's stock and can't even afford 93 octane fuel, though he is ambitious about modding so maybe one day lol. Friends of friends have asked me but it's always been a NO. I don't want the liability and doing that would make it more of a chore instead of a hobby. A lot of us are probably like me so we're not all bad veeefour lol.
    Knock Retard is the reduction or prevention of knock by lowering ignition timing:

    (+) Adding Knock Retard = Reducing Timing. PCM is seeing knock.
    (--) Lowering Knock Retard = Increasing Timing. PCM isn't seeing knock.
    __________________________________________________ ________

    2014 Mustang GT Premium. VMP Gen2R Supercharged with an FTI 3000rpm Converter. JLT, BMR, Steeda, Viking, etc.
    Don't fix it if it ain't broken | Maximum effort gets maximum results