Results 1 to 19 of 19

Thread: Decel issues at low rpm. Car wants to keep going for a moment...

  1. #1
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605

    Decel issues at low rpm. Car wants to keep going for a moment...

    My 14 GT has a weird decel issue. If I'm at under 2k rpm and goose the throttle a little, it gives me an 'echo' of rpm jump after it starts decelerating. See below photo please.

    Where this occurs is exactly when TQ Based Decel switches to Target N (vertical line in photo) and ETC request goes from 5lb ft. to 48 lb ft. Throttle and timing jump to the task as well. Is there a way to keep it in TQ based decel till it hits sub 1000rpm? The vertical line is the exact moment it switches to Target N Torque source. I don't believe tuning driver demand will affect Target N. Also torque based decel changes have no affect either. It's target N where the change needs to occur I believe.

    Anyone have any ideas why the car would be requesting torque when I'm not on the pedal? It only happens in drive, not neutral and from 1000-2000 rpm range. Specifically when the torque source changes in the log I guess when nearing idle, and I've inputted some throttle recently. I have some spark and fuel changes in the tune to allow crackling but I believe this has existed even without those changes. It will be what I try next

    Decel Issue.JPG
    Last edited by blackbolt22; 04-22-2019 at 10:22 PM.
    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

  2. #2
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Have you done anything to make your decel torque decay rate more aggressive? Or have you increased the decel RPM max rate tables? They don't seem to be agreeing with how quickly your RPM should return to Idle. Then again they could be saving you from stalling.

    I think the RPM decel max tables are pretty straight forward, you may need to increase them, you may not.

    I modified the axis and table values of the Torque decay rate table to get a better 3D look at the area you say you are having trouble with. I would say increasing the values for the 5.5 and -5.5 values would make it more aggressive, and agree more with what your airflow is calling for, this would have a cascading effect and you would eventually need to adjust the other rows and columns as well to avoid inconsistent decel rates.

    Torque decay rate.PNG

    I'm just going to attach this diagram from Fords patent again as I think it applies a lot, in giving a general idea of what is going on in the ECU and controlling the engine from idle to WOT.

    US6279531-drawings-page-5.png

  3. #3
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    I can't remember exactly when this started but I think it coincided with my 4C converter install. Now that you mention Decel Max it might be that the converter is looser allowing faster decel of the engine in drive than allowed so target N is trying to 'catch' it from falling and adding in torque to slow the decent.

    Initially when I had the issue these tables were stock settings. The only thing I have adjusted thus far is Decay Rate to try and fix this issue. Decay Max for Neutral and Drive is stock. I didn't think much of it because the RPM axis stops at 880 and the setting seems to allow for more decel than neutral. 880=2000 in D, 900=700 in N.


    Decel Issue2.JPG

    This decel max seems like it 'could' be the cause but 880 is set to 2000rpm/sec. No way is it decelerating that fast in the log when it occurs. The photo in my original post has RPM max scaled at 2000 just to demonstrate the echo but in reality according to the log it's:

    Decel start 8.143 @ 1750rpm
    Decel end 8.522 @ 1284 rpm

    .379 seconds @ 466rpm changed

    1/.379 = 2.6385 * 466 = 1103 rpm per sec

    Still I should go double it later when I get off of work and see!! lol.
    Last edited by blackbolt22; 04-23-2019 at 07:05 AM.
    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

  4. #4
    Advanced Tuner bbrooks98's Avatar
    Join Date
    Oct 2011
    Location
    Gainesville, VA
    Posts
    301
    I had a similar issue due to target N. When I went through the calibration I noticed the tq source was zeroed out for target N. Once I set that to stock it got much better, then I adjusted the number a little bit and found some improvement. Just something to try..

  5. #5
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Quote Originally Posted by bbrooks98 View Post
    I had a similar issue due to target N. When I went through the calibration I noticed the tq source was zeroed out for target N. Once I set that to stock it got much better, then I adjusted the number a little bit and found some improvement. Just something to try..
    I'm assuming you mean torque ratio? This was one of the things I checked. All 3 Target N values are set to stock.

    If Decel MAX doesn't fix this later for me though, I will try and work with the bordeline and torque ratio tables since the transition to MP0 (stock as well) when I let off the throttle does cause a peak in timing during tq. decel and a subsequent dip when it changes to Target N - though that might be a bandaid 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

  6. #6
    Advanced Tuner bbrooks98's Avatar
    Join Date
    Oct 2011
    Location
    Gainesville, VA
    Posts
    301
    Quote Originally Posted by blackbolt22 View Post
    I'm assuming you mean torque ratio? This was one of the things I checked. All 3 Target N values are set to stock.

    If Decel MAX doesn't fix this later for me though, I will try and work with the bordeline and torque ratio tables since the transition to MP0 (stock as well) when I let off the throttle does cause a peak in timing during tq. decel and a subsequent dip when it changes to Target N - though that might be a bandaid lol.
    Yeah, torque ratio I meant.
    2011 Mustang GT TT A6
    1998 Eclipse GSX Awd

  7. #7
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    OK I doubled Decel Max in Drive, no change. Set a few other items back to stock, no change. A side question: If I increase torque decay so the engine decels faster, I am increasing the value, right?

    Murfie I'm not quite sure what you meant by increasing 5.5 and -5.5 - you mean in this axis?

    Here is the stock torque decay table:

    Stock Decel Rate.JPG

    and mine:

    My Decel Table.JPG <- this is just me trying to speed up the decel and I think I've had this stock as well but it still 'echo's will try stock on these tomorrow. Might even roll with a 100% stock tune just to see!

    Each deceleration happens like this, regardless of RPM:
    Torque Source = Tq Based Decel
    Spark Source = Closed throttle decel

    this transitions to..
    Torque Source = Target N
    Spark Source = Torque control

    The difference is, when the problem does NOT occur, it only blips the top set of sources and then sits in Target N the entire way down. When it does occur is stays in the first set of sources then when it transitions to target N, this is when the echo occurs as shown at the line (begin second set of sources). As a result I'm not sure Target N stepping in is abnormal, it's more that the transition to it takes longer at lower rpms where the rpm decel rate is actually slower.

    Decel Issue 3.JPG

    Finally in the log above I am bumping the throttle but I can also be steady throttle/rpm/mph and just let off and it 'echos' an RPM 'bump' just like you see here. So it doesn't matter what precedes the decel really. It's as if there is a HIGH SPOT in the accel pedal right before 0% i.e. as if 0% is 0% throttle, 5% is 20% throttle, and 10% is 10% throttle. That transition point between on and off pedal is where the tq bump happens. Keep in mind the tune is mostly stock but tomorrow I will go full stock and test lol.
    Last edited by blackbolt22; 04-23-2019 at 11:03 PM.
    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

  8. #8
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    If that is your torque decay table, I would say thats too aggressive for positive torque and not aggressive enough for negative torque values.
    I was looking at a boss calibration, that is where that 3d plot is from. they don't have a 0 in the axis it goes from 5 to -5. Idle RPM's target is 0 torque for ETC and therefore engine brake torque.

    To understand schedule torques target:
    From the patent image I posted, and just that patent in general, you can take scheduled torque, and subtract ETC torque and you will see your "idle torque". For the most part this is usually a pretty constant value, something like ~35-50 ft lbs at any given time. If your ETC request drops faster than your scheduled torque does, scheduled torque will bring the request back up to that 35-50Ftlb difference as it decays. Scheduled torque is the controller of your airflow.

    Keep in mind during this transition from above idle to idle, "idle torque" is determined by your torque model by giving a MBT spark and torque value, but idle does not use the MBT spark value directly. Idle uses a torque reserve logic or idle spark logic, so that it knows the max spark to use at idle and maximum torque it can make at idle this is usually some thing like 70-80ftlbs. Having this extra torque in reserve allows it to determine a less than MBT spark to use and it can quickly increase or decrease torque, as needed, just by spark control. Its goal is a target RPM AKA target N. For this RPM, it has an airflow it should target as well. these two things are determined in the RPM and airflow tabs. It can not go into this idle reserve mode until it knows target N will not drop too fast for recovery. Hopefully this clears up your questions about the sources you are seeing. Basically one set means return to idle, and the other means idle control. Airflow is relatively slow, and at low RPMs is slower and could be too late to prevent a stall, the only thing that can happen quick at low RPMS with any accuracy is spark and fuel control, but if the air isn't there nothing is going to help. So it is cautious and makes sure the air thats needed is there when it needs it. This is why dash-pots for carburetors exist. This same logic is why traction control system are not that good, as they are working off a guess at needed airflow in the future, and most of the time that guess is wrong or overly cautious due to the time delay.

  9. #9
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Quote Originally Posted by murfie View Post
    If that is your torque decay table, I would say thats too aggressive for positive torque and not aggressive enough for negative torque values.
    I was looking at a boss calibration, that is where that 3d plot is from. they don't have a 0 in the axis it goes from 5 to -5. Idle RPM's target is 0 torque for ETC and therefore engine brake torque.

    To understand schedule torques target:
    From the patent image I posted, and just that patent in general, you can take scheduled torque, and subtract ETC torque and you will see your "idle torque". For the most part this is usually a pretty constant value, something like ~35-50 ft lbs at any given time. If your ETC request drops faster than your scheduled torque does, scheduled torque will bring the request back up to that 35-50Ftlb difference as it decays. Scheduled torque is the controller of your airflow.

    Keep in mind during this transition from above idle to idle, "idle torque" is determined by your torque model by giving a MBT spark and torque value, but idle does not use the MBT spark value directly. Idle uses a torque reserve logic or idle spark logic, so that it knows the max spark to use at idle and maximum torque it can make at idle this is usually some thing like 70-80ftlbs. Having this extra torque in reserve allows it to determine a less than MBT spark to use and it can quickly increase or decrease torque, as needed, just by spark control. Its goal is a target RPM AKA target N. For this RPM, it has an airflow it should target as well. these two things are determined in the RPM and airflow tabs. It can not go into this idle reserve mode until it knows target N will not drop too fast for recovery. Hopefully this clears up your questions about the sources you are seeing. Basically one set means return to idle, and the other means idle control. Airflow is relatively slow, and at low RPMs is slower and could be too late to prevent a stall, the only thing that can happen quick at low RPMS with any accuracy is spark and fuel control, but if the air isn't there nothing is going to help. So it is cautious and makes sure the air thats needed is there when it needs it. This is why dash-pots for carburetors exist. This same logic is why traction control system are not that good, as they are working off a guess at needed airflow in the future, and most of the time that guess is wrong or overly cautious due to the time delay.
    Thanks for your words! I went back to the stock tune as a test and it was still there, just as strong. I tried the boss decel table + axis and it was still there, just as strong! The only thing that has helped -somewhat- was delaying the upshift schedule so that the car can stay at a higher rpm more often. There are still times when I'm sitting in 2nd on a back street though @ 1500 rpm/steady mph, let off the throttle and it still occurs!!

    I had an idea to try and match the transition between tq. based decel and target N. If you look at the image below (I made your Idle Torque line) it appears etc/scheduled, throttle angle, and spark all increase in a quick ramp then Target N Kicks in. Also TA error increases. What do I need to change to get rid of this ramp of etc/scheduled torque? Driver demand, first row?

    Decel Issue 6.JPG
    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

  10. #10
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Idle> airflow, usually the interpolation of the engine coolant temp axis in the multiplier tables are over looked.

    Make the drive table more like the neutral table in the lower RPMs.
    Idle airflow to Idle load.xlsx

    The other thing you may want to look into is the idle torque IPC tables. They line up with what you are logging fairly well. IPC is another thing thats over looked when it comes to torque request. The ECU needs to think the torque is an achievable goal, not too high not too low.

    Just realized the 11-14's don't have these idle torque IPC tables, might be the way torque and spark are used differently for idle between the older and newer.
    Last edited by murfie; 04-25-2019 at 01:54 AM.

  11. #11
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Quote Originally Posted by murfie View Post
    Idle> airflow, usually the interpolation of the engine coolant temp axis in the multiplier tables are over looked.

    Make the drive table more like the neutral table in the lower RPMs.
    Idle airflow to Idle load.xlsx

    The other thing you may want to look into is the idle torque IPC tables. They line up with what you are logging fairly well. IPC is another thing thats over looked when it comes to torque request. The ECU needs to think the torque is an achievable goal, not too high not too low.

    Just realized the 11-14's don't have these idle torque IPC tables, might be the way torque and spark are used differently for idle between the older and newer.
    I really appreciate the spreadsheet. I learned a bit and also modified it so I could get my airflow values below. This is what I will try next. I'm assuming it's ok to scale up to 2000? The end result of this should be less airflow as I decel...hopefully

    Decel Issue 7.JPG
    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

  12. #12
    Advanced Tuner
    Join Date
    Feb 2017
    Location
    TX
    Posts
    761
    I don't think this was mentioned.
    The 11 - 14s have spark decel. Stock, these values are set to 60* giving torque control, complete control. Changing that table. Reduces high spark during decel which does not help slow down the rotation. You could set 1000 to 2500 to 8 - 12*. Set 850 to 15*. It puts you in better control of decel. Maybe not better, but different control.

  13. #13
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Tried everything suggested. Still unsolved, no change. Tried a bunch of my own ideas including very high decel max and rates, no change at all. I cannot figure out why it's requesting torque on a decel from sub 2k rpm but it definitely occurs the moment it switches to Target N. Timing is naturally higher here as well but throttle opens up too. ETC request goes UP for a moment in the transition from tq based decel to target N. I did some testing and it actually also does it in Neutral but not as noticeable! If throttle and ETC request didn't also increase I would say it's just because of the timing transition... but they do.

    If there is any situation where the trans slows the engine down like in 1st gear, it doesn't happen at all.. there is no request jump. It's only when the rpms decrease quickly... Really seems like a decel limit but I've adjusted all of those - they don't seem to be in effect when in Target N.
    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

  14. #14
    Advanced Tuner
    Join Date
    Feb 2017
    Location
    TX
    Posts
    761
    Did you post your file?

    Have you changed any of these parameters. What does your driver demand look like?

    enable rpm error.PNG
    Last edited by Thatwhite5.0; 04-28-2019 at 09:25 PM.

  15. #15
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Quote Originally Posted by Thatwhite5.0 View Post
    Did you post your file?

    Have you changed any of these parameters. What does your driver demand look like?

    enable rpm error.PNG
    All those are stock. Driver demand is stock. Driver demand is only active when driver demand is the source but I did try it just for funzies. I lowered torque in the first few rows from 3k to 1150 or w/e and it made no difference.

    This issue exists on 100% stock tune (which I tried just to rule out my custom settings) . This issue coincided with my 4C converter swap. Something about the looseness of the converter in 2nd/3rd/etc. gear at low rpm (1000-2000) makes the torque system act weird. Not sure if there is even a fix. The next thing I will try is lowering timing in the region where the decel occurs. Target N woes... lol.
    Last edited by blackbolt22; 04-28-2019 at 10:04 PM.
    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

  16. #16
    Advanced Tuner Witt's Avatar
    Join Date
    Aug 2006
    Location
    Pittsburgh, PA
    Posts
    380
    I have a blown Copperhead/GT with a 3C and a turbo swapped TC1797/Ecoboost with a 5C and will say everything is there to make it decel and idle correctly with those converters. Without a tune and log posted I couldn't even begin to guess where the problem is. Good luck.

  17. #17
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Just a follow up on this issue...

    The summary of the problem is as follows:

    -At light throttle, upon a decel under 2000 rpm Torque based decel is used initially and then Target N around 1300-1500 rpm. During the transition there is a large differential of power creating the issue of the car trying to take off like if I blipped throttle again.

    -At light throttle, upon a decel above 2000 rpm Target N was only used and there was no issue at all.

    The solution:

    I turned off torque based decel and tuned dashpot parameters instead. The changes I made had an impact as opposed to torque based. As a result only Target N is used during decel and I gotta say I wish I had done it alot sooner. Problem resolved. The car decels and drives perfectly.
    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

  18. #18
    Tuner in Training
    Join Date
    Mar 2018
    Posts
    19
    Quote Originally Posted by blackbolt22 View Post
    Just a follow up on this issue...

    The summary of the problem is as follows:

    -At light throttle, upon a decel under 2000 rpm Torque based decel is used initially and then Target N around 1300-1500 rpm. During the transition there is a large differential of power creating the issue of the car trying to take off like if I blipped throttle again.

    -At light throttle, upon a decel above 2000 rpm Target N was only used and there was no issue at all.

    The solution:

    I turned off torque based decel and tuned dashpot parameters instead. The changes I made had an impact as opposed to torque based. As a result only Target N is used during decel and I gotta say I wish I had done it alot sooner. Problem resolved. The car decels and drives perfectly.
    Hi blackbolt22,

    Could you tell me which dashpot parameters have you tuned? because stock file comes with all zeroed. Only tables with values are idle dashpot decay.

    I have a 14' with a VMP supercharger. The issue was when I installed VMP Twin Throttle 69mm. I set up the gt500 throttle parameters as they told me but there is nothing in this world that give me the control of the idling. I'm getting crazy. I have like 5 days moving every table that controls idle, throttle, airflow, everything.

    Thanks so much.

  19. #19
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Here is a good thread which contains settings and tunes. GapRider's post has a tune similar to mine you can use for initial dashpot values https://forum.hptuners.com/showthrea...tuning+dashpot . After that you have to tune it for your preference/car, like I did. Read the descriptions of each parameter and think about what it's doing.

    If it's out of control with the new throttle body , did you check for physical leaks in tubing, connections, gasket etc.? If the throttle body model is off compared to the actual, it will be out of control. What's the ETC Angle errors looking like? Any IPC errors? If you aren't logging those you should. If TB model or TQ model are off it will cause strange issues. Torque based decel can work better if the data is accurate to the actual.
    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