Page 1 of 2 12 LastLast
Results 1 to 20 of 32

Thread: Late model Ford's Torque Control ETC System

  1. #1
    HP Tuners Support
    (foff667)
    Bill@HPTuners's Avatar
    Join Date
    Jun 2004
    Location
    Hailing from Parts Unknown
    Posts
    28,247

    Late model Ford's Torque Control ETC System

    I asked for a little information on how the torque system worked and got a wealth of information that I thought I should share, this was provided by Steven@hptuners:

    Some terminology here:
    - Net/Brake Torque - Actual output torque of the engine into the transmission.
    - Indicated Torque - The ideal torque the engine would output at stoich & MBT with no torque losses due to anything (i.e. friction losses, pumping losses, AC load, etc)
    I.e. Net Torque = Indicated Torque - Torque Losses

    Driver Demand Torque comes from your tables. Its the desired NET torque of the engine from your pedal input.
    The desired NET torque can be limited by many different torque requests, like traction control, engine net torque limits, as well as things like idle torque request. Your driver demand is only one of many possible requests.
    There are many limiters of max NET torque, and there are also clips for minimum NET torque (to keep the engine from stalling during idle and so on).

    The final "net" torque request is called something like "ETC Torque Request" in our software. This is the final desired NET torque that is actually converted into an airflow command. If you have no torque limits, this will be driver demand, but other systems can override it as mentioned.
    Now if we just used this torque to control the engine, we'd always end up with less torque out than what we wanted. This is because there are torque losses actively happening, caused by internal engine friction, pumping losses, accessories like the alternator and A/C, as well as spark and fuel effects.
    To compensate for this, the ECU calculates the current losses, and adds them to your net request. This brings your torque request into the "Indicated Torque" domain.

    Other compensations are applied at this part as part of the "CLIP/ADD" system.
    CLIP (Closed Loop Intervention Prevention) is responsible for making sure the output torque of the engine MATCHES what we request. Its designed to prevent us from having a conditions where more torque comes out than requested, i.e. unintended acceleration. As such its calculating a feedback torque to add to our final torque to keep requested EQUAL to actual. I'll talk about how actual is calculated later.
    ADD (Adaptive Driver Demand) is responsible for adapting your driver demand table so that the requested torques from driver demand match up with actual torque outputs. If your CLIP/ADD mode is "Intervention Prevention" you only have the realtime feedback, if its "Adaptive" you also get an adaptive table that slowly learns the feedback values, thus correcting your driver demand.
    I will be adding logging to all Ford vehicles soon that will show you what the "CLIP/ADD" torque feedback actually is for you.

    This final indicated torque request is called "Scheduled Torque", i.e. the torque we actually need to convert into an airflow request to get our desired net torque.
    To convert this into an airflow, we use the Torque Model > Inverses. This converts our scheduled torque into a desired air load.
    Now, at this point, EVEN more limits can be applied to cap our load. You can log these as "Torque Airlimit Source".
    They will restrict your load and thus your final airflow for various safety conditions.

    Load is then converted into an actual airflow via math, and this airflow can also be clipped. These will also show up in "Torque Airlimit Source" if active.
    These sources amongst others will also show up in your "Torque Driver Demand Limit" indicating they are ultimately limits on the driver demand request as well, though that pretty much just tells you what you already know from Torque Airlimit Source.

    This final desired airmass is used to ultimately control the throttle,and in factory turbo/super cars the turbocharger and superchargers as well.

    However, its not the final limiter of torque. Two more terms to express you will commonly see. They are "Fast Path" and "Slow Path" torque requests. This corresponds to how fast these torques can actually be controlled. Airflow is your SLOW path. Spark and Fuel are your FAST path. This is because it takes time for air to travel to the cylinder from the throttle, but you can cut spark or fuel nearly instantly for a given cylinder. You may also see these referred to as "Base" (slow) and "Instantaneous" (fast) torque requests.

    Some torque sources control the slow path torque only, some control fast path only, and some (like traction control, engine speed limiters) can control both.
    As such, Fords "SAI Torque Reduction" system can come into play and further reduce torque. I have a post on the forum about this.

    That hopefully answers any questions (or makes new ones) about the desired torque system.
    Now as I mentioned earlier, there's also an ACTUAL torque system, used to create feedback for CLIP/ADD torques.

    This system uses the CURRENT airflow, CURRENT spark, and CURRENT fuel to calculate an estimated torque.
    Your CURRENT load goes through the Torque Model Engine Torque tables, to give an INDICATED estimated torque.
    Current losses from accessories, as well as losses from actual spark and fuel are applied. This gets you your current NET estimated torque.

    This is compared to your desired NET torque inside of the CLIP module to calculate what feedback is needed to line the two up. It errors on the side of preventing overtorque.
    As well, the INDICATED estimated torque is used by IPC (Independent Plausibility Checker) to decide if the engine torque is plausible for current conditions. This uses the torque min/max tables inside the IPC.
    If you're outside the range, it increments the wheel torque error with how much you're off, and if you eventually go beyond it, you get the wrench light limp mode.


    Thats a brief overview of Ford's Torque Control ETC system.

    I will sticky this thread shortly as well.
    It doesn't have to be perfect, it just needs to be done in two weeks...

    A wise man once said "google it"

  2. #2
    Advanced Tuner
    Join Date
    Jul 2017
    Posts
    529
    This is info I have been searching for, and partly figured out by myself, through trial & error.
    Thanks for writing it all down, it for sure answers a lot of Q's for me, and verifies ideas I have had, about how it works.
    The new torque inverse calc in the newer software versions should make this a lot easier to dial in. I will be trying this my next tuning session for sure!

  3. #3
    Advanced Tuner 4wheelinls1's Avatar
    Join Date
    Jun 2009
    Location
    Melbourne Australia
    Posts
    283
    Thank you.

  4. #4
    Advanced Tuner
    Join Date
    Sep 2010
    Posts
    872
    Quote Originally Posted by MRRPMBRP View Post
    This is info I have been searching for, and partly figured out by myself, through trial & error.
    Thanks for writing it all down, it for sure answers a lot of Q's for me, and verifies ideas I have had, about how it works.
    The new torque inverse calc in the newer software versions should make this a lot easier to dial in. I will be trying this my next tuning session for sure!
    I used it on my personal car recently, it really smoothed out the throttle on my Paxton car. Pedal feels linear like it did when it was NA, only issue is hovering around -10in.hg(which is where the BOV closes). If I hover in that range of the throttle it will buck a bit if my foot moves a hair off the gas from the BOV opening and closing. IF I am in the -5in.hg then it will keep the BOV shut the whole time and even if my foot lets off a hair the BOV stays shut so no issues.

  5. #5
    Tuner
    Join Date
    Jun 2017
    Location
    South East Georgia
    Posts
    127
    Another thread to add to the bookmarks. It’s always interesting how things that you’ve inferred a function for from testing turn out right or wrong when information like this is posted. Thanks for clarifying this Bill, it’s definitely more accurate than the previous post I’d made.

  6. #6
    Advanced Tuner
    Join Date
    Jul 2017
    Posts
    529
    I'm still trying to figure out how to use the torque inverse calculator. Anyone know how?
    I click on the calculator in the software, the calc opens up, won't calculate anything until you change a number, in the torque table.
    Then, after you hit calculate, it seems to do nothing at all. If it's calculating anything, where's the calculated output going? I have looked at the inverse table to see if it's upadating, but it isn't.

  7. #7
    HP Tuners Support
    (foff667)
    Bill@HPTuners's Avatar
    Join Date
    Jun 2004
    Location
    Hailing from Parts Unknown
    Posts
    28,247
    Quote Originally Posted by MRRPMBRP View Post
    I'm still trying to figure out how to use the torque inverse calculator. Anyone know how?
    I click on the calculator in the software, the calc opens up, won't calculate anything until you change a number, in the torque table.
    Then, after you hit calculate, it seems to do nothing at all. If it's calculating anything, where's the calculated output going? I have looked at the inverse table to see if it's upadating, but it isn't.
    works fine for me in the newest 3.7 beta, you can always open the file you are modifying against itself to see the changes you are making. Maybe you are using old software.
    It doesn't have to be perfect, it just needs to be done in two weeks...

    A wise man once said "google it"

  8. #8
    Advanced Tuner
    Join Date
    Jul 2017
    Posts
    529
    Quote Originally Posted by Bill@HPTuners View Post
    works fine for me in the newest 3.7 beta, you can always open the file you are modifying against itself to see the changes you are making. Maybe you are using old software.
    How is it supposed to work? I don't know if I'm doing it correctly, or if something's wrong with the software.

  9. #9
    HP Tuners Support
    (foff667)
    Bill@HPTuners's Avatar
    Join Date
    Jun 2004
    Location
    Hailing from Parts Unknown
    Posts
    28,247
    you enter in the values, then click calculate inverse, then go to the next one, rinse & repeat.
    It doesn't have to be perfect, it just needs to be done in two weeks...

    A wise man once said "google it"

  10. #10
    Advanced Tuner
    Join Date
    Jul 2017
    Posts
    529
    Quote Originally Posted by Bill@HPTuners View Post
    you enter in the values, then click calculate inverse, then go to the next one, rinse & repeat.
    Are you meaning calculate inverse one cell at a time?
    When I open the inverse calculator, it's already populated with the torque values, no need to enter anything that I can tell, although you must change some value in the torque table to get the "calculate inverse" to not be greyed out.
    Where do the new inverse table values go? The inverse torque table doesn't update, does the inverse calculator output elsewhere, then you must get the values & put them into the inverse table?

  11. #11
    Advanced Tuner
    Join Date
    Jul 2017
    Posts
    529
    Figgerd it out, first, updated HPTuners, then, figgerd out that you can't look at the inverse table while calculating.
    I had the inverse table open so I could see if it actually updated, it didn't. I shut the inverse table down, read it again, it was updated. Seemed to work good like that. I haven't flashed the calculated figures in the pcm yet, but will later on, see how they do.

  12. #12
    Advanced Tuner GapRider's Avatar
    Join Date
    Nov 2015
    Location
    South Carolina
    Posts
    246
    Bill, thanks so much for that overall explanation - it confirms and fills gaps.
    2019 C7 Stingray M7 - long tube headers, 6.30/6.22 226/238 cam, supporting stuff, DOD and VVT delete.
    Stock everything else

  13. #13
    Advanced Tuner 4wheelinls1's Avatar
    Join Date
    Jun 2009
    Location
    Melbourne Australia
    Posts
    283
    With this information has anyone set a PID to really tidy up IPC error? I find I'm still juggling IPC MAX and Torque inverse tables, I seem to struggle to eliminate IPC under high load 40-60% throttle 2000-3600 rpm. I have been using the Rouch OS with stock throttle settings as the throttle is stock and I seem to get variation in IPC error.

    I can identify when IPC faults and normally it is when the Brake TQ exceeds the scheduled TQ, that makes sense to me so I go and raise the schedule and do the inverse for the Mapped point in use, say 14 and retry but the error is still there.

    I also noticed the IPC error never really matched the discrepancy, is this a counter hence why it faults after 40 000 IPC errors in a stock file, does that reset on key off? Is there a friction TQ that impacts that we are not able to log that explains the discrepancy?

    In regards to IPC min and IPC min tables I notice Whipple don't even scale into loads over 1 and the values in the max are quite low, I find I have issues if I try the same.


    Oddly I ended up modifying the ETC predicted angle despite having a stock throttle and stock data. Perhaps the air flow characteristics change with differing inlet tract and design.
    Last edited by 4wheelinls1; 03-21-2018 at 12:31 AM. Reason: More information available

  14. #14
    Senior Tuner CCS86's Avatar
    Join Date
    Nov 2017
    Location
    Austin
    Posts
    1,089
    After reading this description, I have a couple questions:


    In the screenshot below, the car is cruising at a steady 39 mph.

    Based on the Driver Demand table and Pedal Position, we get the ETC Torque Request of 44 ft lbs, yes?

    The ECU is expecting ~54 ft lb of loss, and schedules 98 ft lb, to achieve a net torque of 44.

    In my case, has it estimated wrong? The Engine Brake Torque reported is only 13 lb ft. Shouldn't this be equal to ETC TQ Request if tables are correct?

    TQ-question.png




    In this case, the car is idling stably, yet it is reporting an Engine Brake Torque of -10 lb ft. How can this be?

    TQ-idle-question.png
    Last edited by CCS86; 06-22-2018 at 05:21 PM.

  15. #15
    Advanced Tuner
    Join Date
    Apr 2018
    Posts
    605
    Quote Originally Posted by CCS86 View Post
    After reading this description, I have a couple questions:


    In the screenshot below, the car is cruising at a steady 39 mph.

    Based on the Driver Demand table and Pedal Position, we get the ETC Torque Request of 44 ft lbs, yes?

    The ECU is expecting ~54 ft lb of loss, and schedules 98 ft lb, to achieve a net torque of 44.

    In my case, has it estimated wrong? The Engine Brake Torque reported is only 13 lb ft. Shouldn't this be equal to ETC TQ Request if tables are correct?

    TQ-question.png




    In this case, the car is idling stably, yet it is reporting an Engine Brake Torque of -10 lb ft. How can this be?

    TQ-idle-question.png
    At 35MPH , 1100rpm on my 14GT:
    ETC Tq request=60 lb ft.
    Engine Brake Tq=65 lb ft.
    Indicated Tq=88 lb ft.

    Usually my ETC Tq request and engine brake tq are always very near each other in value when cruising at a steady speed or a slight coast/decel. If they were way off in steady conditions I'd be worried and think the tables were wrong.

    As for the -10 lb ft. this is just deceleration of the engine in some way... very slight I might add.

  16. #16
    Senior Tuner CCS86's Avatar
    Join Date
    Nov 2017
    Location
    Austin
    Posts
    1,089
    But if it is idling stably, nothing is decelerating. Something is off a bit.

  17. #17
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    The torque values are more about increase/decrease airflow into load than RPM accelerating and decelerating. Generally more airflow usually means accelerating(but sometimes its just making up for losses), less usually means decelerating(again losses or gains in this case), but if the ECU is seeing more airflow than expected or less than expected you get error. Hence the compare of the estimated net torque and the desired net torque and lining them up.

    In your case your estimated NET torque is being reduced form your desired TA being 2* and actual being 1.5*. DD torque is too high. throttles trying to open too much, getting too much airflow.
    Last edited by murfie; 06-22-2018 at 11:55 PM.

  18. #18
    Advanced Tuner
    Join Date
    Dec 2017
    Posts
    767
    What a great write up, makes it easier to grasp

  19. #19
    Senior Tuner CCS86's Avatar
    Join Date
    Nov 2017
    Location
    Austin
    Posts
    1,089
    Quote Originally Posted by murfie View Post
    The torque values are more about increase/decrease airflow into load than RPM accelerating and decelerating. Generally more airflow usually means accelerating(but sometimes its just making up for losses), less usually means decelerating(again losses or gains in this case), but if the ECU is seeing more airflow than expected or less than expected you get error. Hence the compare of the estimated net torque and the desired net torque and lining them up.

    In your case your estimated NET torque is being reduced form your desired TA being 2* and actual being 1.5*. DD torque is too high. throttles trying to open too much, getting too much airflow.


    Murfie, that was my bad. In the screenshot I posted, the desired angle is rounded to one decimal (it keeps going back to that). It is rounding 1.55* up to 2*. When expended to the same precision, they essentially match.

    Should "ETC TQ Request" match "Engine Brake TQ"? If so, would you correct errors like this in the Indicated TQ / Inverse tables, or Driver Demand?

  20. #20
    Senior Tuner CCS86's Avatar
    Join Date
    Nov 2017
    Location
    Austin
    Posts
    1,089
    On a side note, how can we control the amount of expected loss? For example, with a positive displacement blower, there is now more frictional torque loss. Especially at low rpm/load, it is probably significant enough to make a difference. I haven't found any tables that show these kinds of torque losses. Are we left to just "subtract" it from the indicated/inverse TQ tables?