Close

    • Porsche 991.2 ECU pulls timing/power if intake manifold temperature measurements get too hot?

      This is an interesting topic as it seems to be similar to issues that plagued Porsche 991.1 Turbo models. You likely recall that the 991.1 Turbo and Turbo S were woefully inconsistent. Why? The ECU would retard timing and cut boost in certain situations.


      Those situations were when temps got too hot. The cure? Upgraded intercoolers and an ECU tune.

      It seems we have a similar situation with the Porsche 991.2 Carrera 3.0 9A2B6 powered models.

      Just look at this dyno test of a 991.2 Carrera T equipped with Vektor headers. It shows gains everywhere until up top in the rev range:


      The explanation is the central fan on the front of the car during the dyno test with no air going to the rear toward the intercoolers:


      Compare and contrast these results to Vektor testing with fans directed at the intercoolers in a wind tunnel setup. The graph shows large gains up top with proper fan placement.

      Either way it seems to properly test aftermarket manifolds/headers will require some real world testing outside of a dyno cell. Additionally, upgraded intercoolers are a must on the platform especially when tuned.

      Our own 60-130 testing shows top end power quickly drops off after repeated pulls.

      This article was originally published in forum thread: Porsche 991.2 ECU pulls timing/power if intake manifold temperature measurements get too hot? started by Sticky View original post
      Comments 20 Comments
      1. F16HTON's Avatar
        F16HTON -
        Once the impllers hit maximum speed no more power to be had
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by F16HTON Click here to enlarge
        Once the impllers hit maximum speed no more power to be had
        I can't explain why the other graphs show gains up top but this one is all down low.
      1. F16HTON's Avatar
        F16HTON -
        I can, open up the exhaust allows you to spin the impeller more quickly down low, therefore more boost down low.

        Hopefully they data-logged the session and monitored the AIT's, waste gate duty cycle, ignition retard and other tuning protective measures.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by F16HTON Click here to enlarge
        I can, open up the exhaust allows you to spin the impeller more quickly down low, therefore more boost down low.

        Hopefully they data-logged the session and monitored the AIT's, waste gate duty cycle, ignition retard and other tuning protective measures.
        These headers are supposed to alleviate a restriction up top so it's weird to me every other header dyno shows exactly that. The impeller does not need to spin more to show a gain if you are getting the gain by reducing boost and improving breathing.

        I'm going to test this stuff myself. Any recommendations on testing procedures?
      1. VektorPerformance's Avatar
        VektorPerformance -
        We're going to be testing a direct forced air setup (in the right direction) on the intercoolers soon. Should provide some interesting data based on what we've seen with some improvised setups (or lack of) on the dyno.
      1. F16HTON's Avatar
        F16HTON -
        Click here to enlarge Originally Posted by VektorPerformance Click here to enlarge
        We're going to be testing a direct forced air setup (in the right direction) on the intercoolers soon. Should provide some interesting data based on what we've seen with some improvised setups (or lack of) on the dyno.
        It is not really rocket science when it comes to proving power...

        The dyno is probably the least accurate because it can never truly simulate the vehicle being driven on a road, under load.

        Time over distance...

        You find a stretch of road, measure how long it takes to go A - B at a 100% throttle. Establish a baseline and go from there.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by VektorPerformance Click here to enlarge
        We're going to be testing a direct forced air setup (in the right direction) on the intercoolers soon. Should provide some interesting data based on what we've seen with some improvised setups (or lack of) on the dyno.
        I'm open to testing your set in the real world against the competition on my car with upgraded turbos... doesn't get any better than that.
      1. spdracerut's Avatar
        spdracerut -
        Click here to enlarge Originally Posted by F16HTON Click here to enlarge
        It is not really rocket science when it comes to proving power...

        The dyno is probably the least accurate because it can never truly simulate the vehicle being driven on a road, under load.

        Time over distance...

        You find a stretch of road, measure how long it takes to go A - B at a 100% throttle. Establish a baseline and go from there.
        IF you can get consistent weather conditions. So both temperature and wind. There's a spot where I go cycling and some days there's a ~10-20mph wind depending on the day. Or no wind at all. Going into the wind, 13mph. Going with a tailwind, 26mph. I've played beach volleyball where the wind direction shifted almost 180 degrees over a 3 hour period.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        IF you can get consistent weather conditions. So both temperature and wind. There's a spot where I go cycling and some days there's a ~10-20mph wind depending on the day. Or no wind at all. Going into the wind, 13mph. Going with a tailwind, 26mph. I've played beach volleyball where the wind direction shifted almost 180 degrees over a 3 hour period.
        How would you go about testing?

        I will have multiple manifolds shortly.

        I intend to do 60-130 runs but not sure what I can use for data logging on this platform.
      1. F16HTON's Avatar
        F16HTON -
        Cobb AP has built in data logging.
      1. spdracerut's Avatar
        spdracerut -
        Click here to enlarge Originally Posted by Sticky Click here to enlarge
        How would you go about testing?

        I will have multiple manifolds shortly.

        I intend to do 60-130 runs but not sure what I can use for data logging on this platform.
        Well, first, what are you trying to capture? Just top-end power or full powerband? 60-130 will capture the upper-end of the rpm range. But headers also improve spool-up and low-end. To capture that, you have to start a pull from 1500rpm. Figure 3rd gear if you have the space because I think it tops out around 100mph. So maybe 2nd gear if you don't.

        The road needs to be as straight and level as possible of course. You need to do at least two pulls in each direction, preferably three, and take the average and check for data consistency. By going in both directions, it minimizes the impact of wind and elevation. Of course, you need something to datalog at a high sample rate. Always start with a full tank of gas so as to minimize weight variation. And try to only test when the temperatures are the same, as much as possible. If you test one when it's 80F and another when it's 90F, it's not a fair comparison.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        The road needs to be as straight and level as possible of course. You need to do at least two pulls in each direction, preferably three, and take the average and check for data consistency. By going in both directions, it minimizes the impact of wind and elevation. Of course, you need something to datalog at a high sample rate. Always start with a full tank of gas so as to minimize weight variation. And try to only test when the temperatures are the same, as much as possible. If you test one when it's 80F and another when it's 90F, it's not a fair comparison.
        This is all performance 101.

        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        Well, first, what are you trying to capture? Just top-end power or full powerband? 60-130 will capture the upper-end of the rpm range. But headers also improve spool-up and low-end. To capture that, you have to start a pull from 1500rpm. Figure 3rd gear if you have the space because I think it tops out around 100mph. So maybe 2nd gear if you don't.
        I'm interested primarily in top end gains.

        That said, the Vektor test at World showed no gains down low on the dyno but gains up top.

        Then they stated the time to boost was improved based on dyno logs:

        Click here to enlarge

        Why wasn't this reflected on the dyno graph though?
      1. spdracerut's Avatar
        spdracerut -
        Click here to enlarge Originally Posted by Sticky Click here to enlarge

        That said, the Vektor test at World showed no gains down low on the dyno but gains up top.

        Then they stated the time to boost was improved based on dyno logs:

        https://www.BoostAddict.com/images/i...4pxmPZQL-1.png

        Why wasn't this reflected on the dyno graph though?
        For some reason, World plotted the dyno data starting at 2500rpm. Notice from the datalogs the dyno pulls were actually started at around 1500rpm. Look closely at the datalog graph and the Vektor header pull was started later and at a lower rpm. Where MotoIQ has the vertical blue line plotted, the engine speeds are the same which looks to be 2100rpm, and the Vektor headers are at a higher boost level.

        You've seen Vektor's own dyno plot along with their customer plot which both showed massive spool-up and torque gain everywhere. So the MotoIQ dyno is a bit odd in that the torque was about the same until 4300rpm.

        So.... what is unknown is what mode the car was in when they dyno'd it. Sport? Sport Plus? etc. Maybe just in base mode, the car limits the mid-range torque/load? Who knows what funky things the ECU is doing. When you datalog the car, you'll see the actual throttle, not TPS, is moving around quite a bit which I think is to control load. So it's using the throttle plate to quickly adjust airflow/boost because it's faster reacting than regulating the turbo wastegates. Look at the datalog plot at around 4100-4700rpm and the car is running lower boost with the Vektor headers compared to stock. So maybe the car is trying to hit a preset 'load' target and therefore reducing boost. This car was completely stock tune. Not sure how the Porsche algorithms work, especially in different modes.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        For some reason, World plotted the dyno data starting at 2500rpm. Notice from the datalogs the dyno pulls were actually started at around 1500rpm. Look closely at the datalog graph and the Vektor header pull was started later and at a lower rpm. Where MotoIQ has the vertical blue line plotted, the engine speeds are the same which looks to be 2100rpm, and the Vektor headers are at a higher boost level.
        Ah, good catch!

        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        hen you datalog the car, you'll see the actual throttle, not TPS, is moving around quite a bit which I think is to control load. So it's using the throttle plate to quickly adjust airflow/boost because it's faster reacting than regulating the turbo wastegates. Look at the datalog plot at around 4100-4700rpm and the car is running lower boost with the Vektor headers compared to stock. So maybe the car is trying to hit a preset 'load' target and therefore reducing boost. This car was completely stock tune. Not sure how the Porsche algorithms work, especially in different modes.
        Ok well I will do dyno testing as well on a Mustang with load. I'm exploring data logging options. If you have any suggestions, feel free to add them.
      1. spdracerut's Avatar
        spdracerut -
        Well, MotoIQ used an AEM CD-5, so that's one option.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        Well, MotoIQ used an AEM CD-5, so that's one option.
        I looked it up but don't really understand it. Comes up as a digital display.
      1. spdracerut's Avatar
        spdracerut -
        Click here to enlarge Originally Posted by Sticky Click here to enlarge
        I looked it up but don't really understand it. Comes up as a digital display.
        Yup, it's a dash the plugs into the OBD-II port and can read most of the parameters transmitted over the CAN bus. It can be bought with optional built-in datalogging and also GPS which would come in handy for acceleration testing. In the photos in the MotoIQ article, you can see where they used it to keep tabs on all the parameters during the dyno testing. It's also what allowed them to know when the coolant, oil, and intake manifold air temps were as identical as possible to start each dyno pull to minimize variation. Of course, there's nothing you can do about ambient air temps, but they did testing on back-to-back days and the ambient air temp was within 4F IIRC from the datalogs.
      1. Sticky's Avatar
        Sticky -
        Click here to enlarge Originally Posted by spdracerut Click here to enlarge
        Yup, it's a dash the plugs into the OBD-II port and can read most of the parameters transmitted over the CAN bus. It can be bought with optional built-in datalogging and also GPS which would come in handy for acceleration testing. In the photos in the MotoIQ article, you can see where they used it to keep tabs on all the parameters during the dyno testing. It's also what allowed them to know when the coolant, oil, and intake manifold air temps were as identical as possible to start each dyno pull to minimize variation. Of course, there's nothing you can do about ambient air temps, but they did testing on back-to-back days and the ambient air temp was within 4F IIRC from the datalogs.
        Great info, thank you.
      1. VektorPerformance's Avatar
        VektorPerformance -
        Something else to consider here when looking at different plots is the dyno itself; Inertia, Eddy Current, mechanical brake... Every manufacturer uses different methods that will provide different results. results from a Dyno-Jet will not look the same as results from a Mustang, or SuperFlow or Land & Sea....

        One thing I have noticed when overlaying Plots and just looking at RPMs (as power has increased) is that the time it takes from start-to-finish of pull has reduced. We can do 1/4 mile runs on our dyno, but it's a bit too abusive for my liking to do over and over and over and over and.... Click here to enlarge
      1. F16HTON's Avatar
        F16HTON -
        Click here to enlarge Originally Posted by VektorPerformance Click here to enlarge
        We can do 1/4 mile runs on our dyno, but it's a bit too abusive for my liking to do over and over and over and over and.... Click here to enlarge
        Anyone who squawks about dyno plots is truly uninformed and it takes a great deal of education to overcome the fallacy.

        Cast, log style manifolds are used by manufacturers because they are affordable (in bulk quantity), they are predictable (dimensionally and reliability), and the accomplish the basic task of scavenging exhaust from a cylinder head.

        Cast, log style manifolds are rarely used in any type of situation where performance output is a key factor, ie. motorsports.