Page 1 of 5 12345 LastLast
Results 1 to 20 of 93

Thread: Transport Delay /Transport Time Constant

  1. #1
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458

    Transport Delay /Transport Time Constant

    According to the Cookbook, Transport Delay and Transport Time Constant should be significantly increased. As a general rule, I have been increasing most LT header equipped engines 10-20% over stock values (both transport delay and time constant). Don LaSota actually recommends more.

    These questions are for anyone who has tested the different values on a dyno or someone who is really up to speed on this subject. I am not.

    What is the correlation between the time constant and the transport delay?

    For those who have tested different values, what range of increase seems to be best with a Coyote with an NA engine? Or a PD blower?

    Are these parameters a linear increase with blower engines as LaSota has suggested with NA engines? Or do PD blower engines require less increase in time value over stock values beginning at the point of boost? ie - a non linear progression?

    Are the Cookbook's suggestion of 50% increase on both the delay and the time constant in the ball park?

    Can the increase in transport delay and time constant be asymmetrical? ie - multiply transport delay by say 1.5 and time constant by say 1.25?

    Thanks for providing some clarity.
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  2. #2
    Advanced Tuner
    Join Date
    Apr 2015
    Posts
    212
    I don't have any dyno time or anything like that. But on my na car with 17/8 jba long tubes, I went higher in the lower rpm than I did at the top. I think I'm around 25% increase down in the lower areas and only 15 up in the higher rpm.

    It seemed to be happier with those changes and not just a blanket multiple. I'm also using shitty ford 80s lol.

  3. #3
    Advanced Tuner
    Join Date
    Mar 2014
    Posts
    309
    50% is too much. I also vary depending on the header used. Kooks are mid length so I don't increase the % as much. Full LT's I shoot for 25% and that has worked well.

  4. #4
    Advanced Tuner
    Join Date
    Sep 2010
    Posts
    872
    I usually increase the delay 20% for long tubes...50% sounds like too much of an increase. While I don't know the math behind it, the stock value are for the OEM distance, even if the sensors moved 3ft further downstream it won't take 50% longer to reach them over its current position since the exhaust gasses will already be traveling at maximum velocity as they pass the OEM location, so it's a good assumption that it will only take a little bit more to travel the extra few feet
    Last edited by Jn2; 03-02-2017 at 03:55 PM.

  5. #5
    Senior Tuner
    Join Date
    May 2016
    Posts
    1,134
    Air moves slower through a larger volume pipe if pushed with the same pressure. This means the size of the primaries need to be taken into consideration as well as the distance that the sensors have moved. Velocity gets lower and exhaust gas gets cooler as you get downstream further as well.

    To do this mathematically and get it perfect would be pretty involved. All things to consider if you're looking for a mathematical approach. I tend to do that with tuning problems because my background is computer science and mathematics but sometimes close really has to be taken as good enough.

  6. #6
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458
    Thanks to all for your comments. I just found this comment from murfie from last year on this very subject: Here are his comments:

    The transport delay is basically the length of time the air fuel mixture leaves the combustion chamber and reaches the sensor. This is effected by changing different headers that place the sensor closer or further from the combustion chamber.

    There is also a response delay that oxygen sensors have due to the way they operate as a buffered system.

    Regulations require that these sensors have a monitoring system in place. This is usually what causes the check engine light to come on in higher mileage Honda's and Toyota's. The codes are "catalytic inefficacy" and usually turn out to be a bad sensor. The way they are usually monitored requires the engine to go into open loop mode(ecu ignores sensors and runs off tables) as the sensors are being checked. This is why you can clear the code and it doesn't return for a few thousand miles until the next sensor check. This also can cause the engine to run a nonstoich afr for a short period, usually rich.

    What Ford has done is take advantage of the inherent response delay in oxygen sensors in order to monitor them even while they are in closed loop mode(controling the cars fueling). This is where the time constant comes in. By measuring the time constant from the result of for different equations and comparing it to a maximum response rate they can monitor and determine a bad sensor with out the need for open loop o2 sensor checks. The sensor reports bad if the time constant is measured higher than the maximum set response rate.

    I see no need to change this value unless you are using after market sensors connected to the ecu(not recommended) or what to know if your sensor is going bad sooner than factory tolerances you could possibly lower the value.


    And here is the Ford Patent he is referencing: https://www.google.ch/patents/US5305727. Now I hope murfie chimes in here as I would ask him if he is only referring to changes to those parameters in cars with stock manifolds? Does his reasoning hold true for LTs over the stock manifolds?

    So, now we have yet another position. I have looked at a few modified tunes from the repository vs stock tunes of the same vehicle type and some tuners do change these parameters, but some do not. For instance, in the repository is a GT500 stock vs one labeled GT500 modified. The GT500 modified has both delay and constant increased by 25%. If you compare the stock GT500 to one labeled GT500 modded by Lund, Lund changed the values in the delay significantly with large changes at low RPM with increasingly smaller changes at higher RPM. And the Lund tune did not modify the Time Constant values.

    I assume the empirical evidence would be best provided by a significant amount of dyno testing (which I know both Jon Lund and Jon Jr have done, especially on the 5.8). And as well, when you change these values, do those changes impact the fuel trims? For my part, I will run my blower car at the strip this weekend. I will log two or three runs with my calibration that has the MAF transfer and STFTs dialed-in (but no changes to the delay or time constant values). I will then flash the same calibration with the transport delay and the time constant increased by 25%, log three runs and see if I see changes to the STFTs and as well, if there is a repeatable delta in the MPH of each set of runs. Comments on this?
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  7. #7
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    I was incorrect in how I described Transport delay. Spreading the misinformation. Here is the patent from ford describing it https://www.google.com/patents/US4397278

    Basically because there is a delay caused by air moving the distance from cylinder to senor a back and forth cycle(lean/rich) is used to control fuel. the transport delay time is how long is should sense a commanded condition before it switches to command the opposite condition. Ideally you want that to be 0, but you probably wont get that in the real world. Raising this value only allows more overshoot and less fluctuations. That in turn causes inaccuracies in your fuel trims in the lower %'s. Not a good idea for "dialing in" your MAF curve. To dial this value in, first raise it 20% and get your MAF curve close, then compare your commanded and actual and lower the value until you get minimal actual overshoot. Then recheck your MAF curve. As you lower it the fluctuations will increase so it will take more accurate polling to keep up with it. if your actual is not overshooting your commanded you will get inaccurate trims, you will need to increase this value to fix that. If your actual is overshooting your commanded you will get low resolution trims, you will need to lower it to get cleaner trims. The stock method is great for having trims apply faster when there's a large difference and slower when there's a small difference. It also reduces an over reaction to transient conditions so the lower you get your transport delay the more you need to go back over your transient tables.

    Time constant- Don't change its a senor validation check thing. If the sensor is failing the transport delay will be off and this value determines that. you can keep it in the same stock ratio above your new transport delay values if you want to.

    I added a picture of dialing in a larger MAF housing on a stock value low transport delay value with out raising it first. There's no overshoot of the command because its moving back and forth so fast. The lambda stays where it should but command is jumping around too much causing fuel trim error thats lost in the averaging out(I drew a red line to show this). Raising the value should give a smoother commanded and allow actual to follow commanded easier. This allows a more accurate MAF. With a more accurate MAF curve the commanded and actual will follow each other better and trims will be more accurate. The back and forth you can get the transport delay quite low.

    TransDelay.PNG

    Reduced scales for better resolution of view.
    TransDelay.PNG
    Last edited by murfie; 03-03-2017 at 02:22 AM.

  8. #8
    Tuner in Training
    Join Date
    Feb 2015
    Location
    Indianapolis
    Posts
    36
    Quote Originally Posted by Jn2 View Post
    I usually increase the delay 20% for long tubes...50% sounds like too much of an increase. While I don't know the math behind it, the stock value are for the OEM distance, even if the sensors moved 3ft further downstream it won't take 50% longer to reach them over its current position since the exhaust gasses will already be traveling at maximum velocity as they pass the OEM location, so it's a good assumption that it will only take a little bit more to travel the extra few feet
    I tried 50% after putting long tube headers because I had read in a few places that 50% was a good place to start after putting on long tubes, but it was too much. I ended up using 15% and it was fine. I actually drove the car with the factory delay setting for a bit and it seemed OK, but I knew there must be "some" difference between the long tubes vs. stock manifolds. In the end, it seemed that just a bit more than stock was better than some of the 50% stuff I had read.

  9. #9
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458
    murfie - Thank you for your explanation. That provides some clarity to my questions and assumptions. I suspected that the fuel trims were affected by the delay values but never bothered to empirically test my assumption. I have already created 2 tunes for the strip this coming weekend, one with stock values and one with 25% delay. I will datalog 3 runs with each calibration and should see some variation in the fuel trims.
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  10. #10
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    You won't see it in the fuel trims, they get blurred out in the average. You will see it in commanded vs actual lambda and how they follow each other.

  11. #11
    Senior Tuner
    Join Date
    May 2016
    Posts
    1,134
    So I guess we should be tuning MAF on these with EQ Error instead of fuel trims? I've gotten in the habit of using fuel trims.

  12. #12
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458
    murfie: I already log the commanded vs actual. On a log I took a few days ago, My WB EQ Bank 1 & 2 are reading 1.0 Lambda at idle, and the Equivalence Ratio Commanded holds relatively steady around .98 -.99. STFTs vascillate between +.8 - +1.5. That is with a 20% increase in transport delay. I will log it Sunday as is and also with transport delay back to stock.



    EDIT: Saturday AM: I will datalog the car next weekend. The races were canceled this weekend because of the cold temps.

    2nd EDIT: Next weekend, the forecast is for cold and some snow. So I guess we will shoot for the following weekend, assuming the forecast does not change.
    Last edited by 15PSI; 03-07-2017 at 10:14 AM.
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  13. #13
    Tuner
    Join Date
    Apr 2017
    Location
    Dearborn, MI
    Posts
    92
    15PSI do you have an update? I'm a little confused, are you and murfie saying that MAF must be adjusted when adjusting the delay for LT's?

  14. #14
    Senior Tuner
    Join Date
    May 2015
    Posts
    1,907
    I thought the Time Delay was learned over time? I agree increasing it from the get go helps the car to run better on initial fire up, but after a dozen full drive cycles I would think you could simply log the time delay and paste it into the tune so that after any further reflashes it is spot on.

  15. #15
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458
    I am going to let murfie chime in here. My results were inconclusive when I tested in April. My testing methodology was flawed, at least I came up with nothing provable, possibly because there was not enough adaptation time (10 sec 1/4 runs vs. a nice long drive cycle). I just had the O2 Transport Delay Total added to my OSID (in Beta), so when I have some more free time, I will do some additional logging.
    Last edited by 15PSI; 06-23-2017 at 02:09 PM.
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  16. #16
    Tuner
    Join Date
    Apr 2017
    Location
    Dearborn, MI
    Posts
    92
    Quote Originally Posted by murfie View Post
    That in turn causes inaccuracies in your fuel trims in the lower %'s. Not a good idea for "dialing in" your MAF curve. To dial this value in, first raise it 20% and get your MAF curve close, then compare your commanded and actual and lower the value until you get minimal actual overshoot. Then recheck your MAF curve.
    TransDelay.PNG
    So are you saying that changing the transport delay can effect the MAF curve or just the FT's which can hurt if you are trying to change those.

  17. #17
    Advanced Tuner
    Join Date
    Dec 2015
    Posts
    203
    Do the coyote's not have the capability to log transport delay yet? Or is this different because of widebands? On the older stuff I thought you just logged what was learned then put it into the table

  18. #18
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458
    If I remember correctly, Eric added the O2 transport delay in the 2015+ S550s some time ago. My 2012 did not have this parameter. Eric added it.
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  19. #19
    Senior Tuner
    Join Date
    May 2015
    Posts
    1,907
    its there to log on the 2015+, I have it in my scanner.

  20. #20
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Quote Originally Posted by HextallS550 View Post
    15PSI do you have an update? I'm a little confused, are you and murfie saying that MAF must be adjusted when adjusting the delay for LT's?
    No, The delay will be adjusted for exhaust and sensor position modifications OR rate of airflow through the engine is altered significant enough the learn is way off and compensating a lot.

    Quote Originally Posted by 15PSI View Post
    I am going to let murfie chime in here. My results were inconclusive when I tested in April. My testing methodology was flawed, at least I came up with nothing provable, possibly because there was not enough adaptation time (10 sec 1/4 runs vs. a nice long drive cycle). I just had the O2 Transport Delay Total added to my OSID (in Beta), so when I have some more free time, I will do some additional logging.
    Transport delay will not apply to OL fuel modes. I would say enrichment rate is used in OL modes. Just like transport delay, I also don't agree with raising this to prevent fueling delays. The ford GT has it at .025 and other eco boost have it pretty low as well. That patent explains the time delay and constant are used to determine a response rate which is similar to the enrichment rate. This is compared to a maximum response rate to make sure it doesnt get out of control. I would guess as the engine makes more power both enrichment rate and time delay will be lowered. The time constant is a value specific for the O2 sensor. Its also used to determine if the sensor is good or not.

    Quote Originally Posted by HextallS550 View Post
    So are you saying that changing the transport delay can effect the MAF curve or just the FT's which can hurt if you are trying to change those.
    If the MAF curve was dialed in on a flow bench the transport delay will not effect MAF values directly other than making the FTs react slower or fast to rich or lean conditions, depending on the change made.
    Just when dialing in the maf, I would disable the learn so that the back and forth rich/lean behavior isn't moving around the actual lambda/ commanded lambda on its own. Unless you are paying attention to which side of rich/lean, the average of the data you are applying, is spending more of its time on.
    With good MAF data use the Learn PID it will be more accurate than just math in the scanner.
    Last edited by murfie; 06-23-2017 at 11:17 PM.