Results 1 to 7 of 7

Thread: TPS at 0.4% at idle

  1. #1
    Tuner in Training fjb2069's Avatar
    Join Date
    Mar 2008
    Location
    Chicago
    Posts
    31

    TPS at 0.4% at idle

    LS6 in 69 camaro. cam, heads, manual t56.

    Having issues with engine stalling/rpm diving periodically when car is coming to a stop or coasting. The TPS seems to hang at 0.4% at times instead of 0.0%. Seems to only does this when wheels are turning.

    I put in a new tps and retightened the grounds. I have a separate ground cable from the engine block to the battery. I also changed the tps voltage to readout in mV.

    KOEO 686.3mv 0% tps

    Car running 666.7mv 0% tps.

    I just ran car on my lift in shop. If let the rear tires spin and go up to 30 mph, then let off, the tps sometimes shows 0.4% until I hit the brakes and stop the tires. (Tires not spinning 0.0% TPS) But when the tps is reading 0.4% or 0.0% the voltage is still reading 666.7mv!!!

    NOTE (0% TPS 0.6863 volts 100% tps 4.683volts, therefore I should take about 16mV to cause a 0.4% change in TPS) It also appears that the TPS has a resolution of 0.1%, so I would expect to see tps readings of 0.1, 0.2, 0.3% periodically, but I only see 0.4%


    How can the tps be two different values at the same voltage? If there is a grounding issue you should see a change in voltage?.

    From this, I assume the TPS % reading is calculated from more than just the TPS sensor output voltage.

    Can anyone shed some light on this.

  2. #2
    Senior Tuner kingtal0n's Avatar
    Join Date
    Feb 2017
    Location
    miami
    Posts
    1,799
    For the sake of tuning the car in terms of what the software is doing and understands, 0.4% is the same as 0.0%.

    Microcontroller analog inputs work in terms of ADC counts, they convert the analog voltage to a step-wise function which digital electronics can use.


    When measuring analog voltage using a meter or when the ECU is measuring it can only compare voltage against some ground, the minor differences you see are due to small differences in measured analog voltage against some ground pathways. When current flows through wires the ground is raised, and modern ECU and electronics are very sensitive to small changes, even just by using the meter to measure some voltage you are changing it slightly.

  3. #3
    Tuner in Training fjb2069's Avatar
    Join Date
    Mar 2008
    Location
    Chicago
    Posts
    31
    Quote Originally Posted by kingtal0n View Post
    For the sake of tuning the car in terms of what the software is doing and understands, 0.4% is the same as 0.0%.

    Microcontroller analog inputs work in terms of ADC counts, they convert the analog voltage to a step-wise function which digital electronics can use.


    When measuring analog voltage using a meter or when the ECU is measuring it can only compare voltage against some ground, the minor differences you see are due to small differences in measured analog voltage against some ground pathways. When current flows through wires the ground is raised, and modern ECU and electronics are very sensitive to small changes, even just by using the meter to measure some voltage you are changing it slightly.
    Thanks for your response.

    I believe 0.4% does cause a difference in the operation of the system because the Throttle Follower Airflow Table typically has lower cells that only activate on 0.0% not 0.4% TPS.

    I am also basing this on an expert tuner telling me that this will cause an issue with idle and changing Airflow table to 0.4% to compensate is a bandaid.


    I am not an expert tuner, but I do have 30 years experience in embedded systems and I do agree with your statement about A/D converter operation, but assuming the GM ECU has at least a 10 or 12 bit resolution, +/- a couple bits due to ground noise should still vary the TPS reading 0.1, 0.2, 0.3, and not go from 0.0 to 0.4% repeatedly. And the system should show SOME change in voltage in order to change the TPS value regardless if the voltage change is cause by noise or ground path variance.

    Was hoping someone from HP would chime in to explain what is going on.

  4. #4
    Senior Tuner kingtal0n's Avatar
    Join Date
    Feb 2017
    Location
    miami
    Posts
    1,799
    Quote Originally Posted by fjb2069 View Post
    Thanks for your response.

    I believe 0.4% does cause a difference in the operation of the system because the Throttle Follower Airflow Table typically has lower cells that only activate on 0.0% not 0.4% TPS.

    I am also basing this on an expert tuner telling me that this will cause an issue with idle and changing Airflow table to 0.4% to compensate is a bandaid.


    I am not an expert tuner, but I do have 30 years experience in embedded systems and I do agree with your statement about A/D converter operation, but assuming the GM ECU has at least a 10 or 12 bit resolution, +/- a couple bits due to ground noise should still vary the TPS reading 0.1, 0.2, 0.3, and not go from 0.0 to 0.4% repeatedly. And the system should show SOME change in voltage in order to change the TPS value regardless if the voltage change is cause by noise or ground path variance.

    Was hoping someone from HP would chime in to explain what is going on.
    Programmers and users (depending on the software) can setup a software to differentiate between 0% and 0.4% of course. However it is common practice to tune the 0% and 0.4% regions for any/all maps similarly as you have noticed the ECU may have situations where 0% = 0.4% and you need the ECU to treat them similarly.

    For example Megasquirt ECU allows you to define the closed throttle portions for it's various map features and you would generally choose some higher value, say 1% or 0.8% or something between 0 and 1% as the definition for "closed" because as you have noticed there are times when current in the wires, ground schemes, induced current in wires, etc... may raise the ADC (analog voltage) by some small amounts.

  5. #5
    Tuner in Training fjb2069's Avatar
    Join Date
    Mar 2008
    Location
    Chicago
    Posts
    31
    Quote Originally Posted by kingtal0n View Post
    Programmers and users (depending on the software) can setup a software to differentiate between 0% and 0.4% of course. However it is common practice to tune the 0% and 0.4% regions for any/all maps similarly as you have noticed the ECU may have situations where 0% = 0.4% and you need the ECU to treat them similarly.

    For example Megasquirt ECU allows you to define the closed throttle portions for it's various map features and you would generally choose some higher value, say 1% or 0.8% or something between 0 and 1% as the definition for "closed" because as you have noticed there are times when current in the wires, ground schemes, induced current in wires, etc... may raise the ADC (analog voltage) by some small amounts.
    Yes, I agree. I have tuned several of our race cars with Holley EFI and Bigstuff. These engines have significantly larger cams and I have never had any issue with idling whatsoever.

  6. #6
    Senior Tuner kingtal0n's Avatar
    Join Date
    Feb 2017
    Location
    miami
    Posts
    1,799
    Pls allow me try again to assist

    Quote Originally Posted by fjb2069 View Post
    Having issues with engine stalling/rpm diving periodically when car is coming to a stop or coasting.
    This is a tuning issue, somebody can look at your tune file and help you fix this without doing anything to the TPS sensor or worrying about the mV induction from sensors.
    You should post file and data logs to illustrate your points so people can help, nobody can tell what is going on without those.
    I am merely here to provide forward theory and ideas. For all we know there is some big issue in your file or log.

    The TPS seems to hang at 0.4% at times instead of 0.0%. Seems to only does this when wheels are turning.
    This is a very specific and widely interpreted issue, which depends heavily on the setup. For example if the throttle body is very large, 0.4% may actually be a lot more airflow than on a OEM throttle body. Also, some throttle body do not fully close to the same position exactly. And there is vibration and clearance to consider, both in the valve itself and the union between Sensor and throttle body. All which may affect sensor behavior and the outcome from tuning strategies which may be used to alter/calm behavior of the engine.
    Then, there is the shielding. Some TPS wires do not shield well enough for a custom setup if you've changed anything from OEM, like aftermarket wires or moved wires, bundles of electronics, new placement for electrical items, new current routes, they all may affect current and voltage in the wires around them. A fan comes on through wires that were never in the original application and suddenly you've got more voltage in a wire somewhere else, etc...

    Thus, your specific application and position for electronics and wires, and the union of sensor and throttle body, size of throttle body, etc... everything need be considered. If you are concerned about analog voltages then I recommend re-shielding (or extra shielding) the existing wiring in question, add new ECU grounding and wire sheaths and move them away from other wires. I always use new shielded wire for my TPS and similar sensors for custom applications because you really can't be too careful.




    I put in a new tps and retightened the grounds. I have a separate ground cable from the engine block to the battery. I also changed the tps voltage to readout in mV.

    KOEO 686.3mv 0% tps

    Car running 666.7mv 0% tps.
    With these readings we should consider a couple things. First keep in mind the ECU has it's own 5v output to power sensors, and 5v is far less than battery (Even dead) voltage, so the differences you are seeing engine on / off should not reflect any difference in ECU output or battery condition. It is more likely related to ground offsets and magnetic/electric field affects of nearby wires which flow more current when the engine is running. The ECU shares many common ground features so for example injector drivers and coil drivers, even though they are very small currents, collectively may raise the ECU common ground enough for it to cause detection of a higher analog voltage that you would measure externally via using a multimeter on the sensor wire compare with directly to the ground of the battery terminal.

    Next, we should remember the scanner probably isn't reporting exact numbers or very detailed information. The ecu may very well know the difference between 0.0% and 0.1% internally, but for the sake of cleanliness and simplicity, both visually and programming features wise, it may have been decided to simply show 0% and then 0.4% as the two options instead of every increment. The programmers probably considered the 0.4% a reasonable compromise of resolution to accuracy given the programming features provided. In other words, there aren't any setups which should or could use 0.1% throttle position as a reliable, repeatable trigger for any features, it would be irresponsible and unlikely to maintain, as even the sensor mechanism experiences some wear, as does the throttle body/blade position cannot remain perfect all the time for 20 years, and so forth.


    How can the tps be two different values at the same voltage? If there is a grounding issue you should see a change in voltage?.

    From this, I assume the TPS % reading is calculated from more than just the TPS sensor output voltage.

    With most microcontroller there is a wide range of input analog voltages where the ADC can read a similar sensor values (TPS% and others) because it is comparing internally generated voltage with an ECU provided ground, sort of a 'mini island' in the middle of an ocean of crazy currents going on outside, by design. Our perception of those voltages is varied, depending on whether we are looking from within the ECU (Using a scanner or by taking off the cover the ECU) or outside (by choosing other grounds and external wires). Furthermore there is a difference between what the ECU will 'see' and what we will measure using a device such as multi meter- the two analog voltages rarely line up and just by connecting the measurement device you can create a small offset.

    with respect to the HPtuners scanner and it's reporting of voltages- the scanner is what I would consider a 'blunt' tool and not to be taken too exactly. I find many of it's reported details to vary, sometimes significantly, with what I measure or notice outside the ECU. In other words, I would not freak out over 0% vs 0.4% in the scanner details. If you are truly concerned about the mitigation of these offsets, from the perspective of there being only so many things you can do, here is a list of what you can actually apply:
    1. Superior grounding for ECU, alternative ground pathways
    2. Extra shielding and relocation of sensor wires
    3. inspect and clean throttle body mechanism including where it attached the sensor portions
    4. Connect external multimeter device and compare analog voltage readings

    You may find the external input analog voltages do not change much or at all while the internal measurements ('scanner reported values') tend to change more often. Or something like that. Some TPS sensors will jump around, you should definitely check using an external measurement device and not assume the scanner is correct or the TPS is working perfectly.

  7. #7
    Tuner in Training fjb2069's Avatar
    Join Date
    Mar 2008
    Location
    Chicago
    Posts
    31
    I did not post my tune or log because my question was more based on system operation vs tuning. But at your request, I have attached it now. See near end of log file when RPM drops to about 200 rpm and TPS is 0.4%

    This is what I was assuming, thank you for confirming: "Next, we should remember the scanner probably isn't reporting exact numbers or very detailed information. The ecu may very well know the difference between 0.0% and 0.1% internally, but for the sake of cleanliness and simplicity, both visually and programming features wise, it may have been decided to simply show 0% and then 0.4% as the two options instead of every increment. The programmers probably considered the 0.4% a reasonable compromise of resolution to accuracy given the programming features provided. In other words, there aren't any setups which should or could use 0.1% throttle position as a reliable"
    Attached Files Attached Files