Page 1 of 4 1234 LastLast
Results 1 to 20 of 71

Thread: Ford Speed Density Calculator

  1. #1

    Ford Speed Density Calculator

    Available in the latest VCM Suite beta is a new calculator for the Ford Speed Density model. This calculator is quite sophisticated, so an understanding of the process is vital to help get the best results from the calculator. The calculator is capable of giving you the quadratic coefficients including blowthrough, with the goal being to hasten/simplify the process of calibrating this model.

    Our generated table is shown as load. As will be explained, a classical sense of VE % doesn't necessarily fit the calculator.

    Speed Density Refresher
    A quick refresh on the goal of Speed Density: convert various sensor measurements of air pressure, temperature, and other conditions to get an estimation of the amount of air in a cylinder when combustion occurs.
    Typically, you're looking at Manifold Absolute Pressure (MAP), Manifold Charge Temperature (MCT), Exhaust Manifold Absolute Pressure (ExMAP), etc. to give you a good estimation based on the ideal gas law.

    Starting with the basic equation, we can refine it to suit our needs:
    MzxhZ3b.png

    To convert from "moles" of stuff, to an actual mass, we use a specific gas constant for our gas (air at our reference conditions)
    ptJW1Xe.png

    And then we arrive at our final equation for estimating cylinder aircharge using the pressure, volume, and temperature for the gas in our cylinder! Easy!
    ewNaDQ2.png

    This would be the amount of air in the cylinder if we had those ideal conditions. Unfortunately, measuring the pressure and temperature of the air once the intake valve is closed is non-trivial, so instead we estimate it based on manifold conditions.
    vQA5F9z.png

    But, that's in the ideal world. In reality, the cylinder may not have enough time to fill, some of the air may be pushed back out into the intake manifold due to valve timing, our temperature may not be an accurate representation, etc, so a "fudge factor" known as "volumetric efficiency" is introduced. This makes our final equation roughly:
    uOX3q7H.png

    The basics of calibrating speed density revolve around getting volumetric efficiency right in as many conditions as you can.

    How is Ford different?
    It isn't really, it still bases its model off of measured/inferred manifold pressure, but it of course has a wrench thrown into it to improve accuracy in many varying conditions.

    You may have heard the model is quadratic in nature. This is only partly true. The real model is actually multi-faceted, and has several parts.

    There are several terms we need to define:
    Maximum Trapped Aircharge - This is the maximum amount of air that can be in a cylinder for a given cylinder condition (pressure and temperature). By definition, you cannot get more air in there, you would need a higher pressure, lower temperature, etc. This is defined by physics.
    Standard Aircharge - This is the maximum amount of air that can be in the cylinder at STANDARD conditions. (For most Ford vehicles, this is 29.95inHg of pressure, 60*F air temperature). This is the value labelled "Engine Displacement" in HPTuners.
    Blowthrough - This is a condition where we are flowing more air through the cylinder than the maximum trapped aircharge. This happens when the intake and exhaust valve are both open at the same time, and MAP > ExMAP. Also known as scavenging.
    Load - This is a ratio of the current aircharge in the cylinder to the standard aircharge.

    As such maximum trapped aircharge is like our 100% VE aircharge amount.
    So, how does the ECU calculate maximum trapped aircharge? First, takes the standard aircharge calibrated under known standard conditions, and corrects this for current conditions. Some of these can be the effect that engine coolant temperature and air temperature might have on charge density (see table below), as well as others. The idea is to get close to the best guess at in-cylinder conditions and use the ideal gas law to get this aircharge.
    5t9XfJY.png

    So, rolling this all into one value, which Ford simply calls "C", probably for "Constant at all the conditions we just took out", we can now calculate the maximum trapped aircharge simply as:
    DZXWidD.png

    What about greater than 100% VE?
    This can happen, when conditions in the cylinder are better than the raw equation might suggest. Typically, the weak link is our use of manifold pressure to estimate cylinder pressure. Certain conditions can actually cause our cylinder pressure to be higher than measured manifold pressure, for example, well tuned "resonant" intake runners can make use of the energy of the air moving to help pack more air into the cylinder.

    To correct for this, Ford has the "Aircharge Multiplier" tables. These directly multiply the calculated C value from above, to allow you to compensate for these effects (whether helpful >1, or harmful <1). Remember, this compensates the actual in-cylinder aircharge. If that air is going into the cylinder, but not taking part in combustion due to leaving through the exhaust valve for example (aforementioned blowthrough), that does NOT become a multiplier to the table.

    Can I shove more air in than this 100%? What is blowthrough?

    You may be thinking, "with boost I have VE greater than 100%!", this would be true if your equation DIDN'T already account for this with its dependence on MAP. So what happens if we DID somehow setup a way to get more air in, say through some nice overlap with the exhaust cam, with our intake MAP > exhaust MAP?
    Well, you have to evaluate: Is this going to increase the air in the combustion chamber ABOVE my theoretical physics calculation? Will this air stay in the cylinder, or will it blowthrough it because of overlap?
    If it stays in, Aircharge Multiplier adjustments are needed.
    If it blows through and into the exhaust, blowthrough tuning is needed.

    Why does Ford track blowthrough?

    Well, fresh air in the exhaust messes up our oxygen sensor readings. We now have excess oxygen in the exhaust, so our measured exhaust lambda is now leaner than what actually happened in the cylinder. However, if we know how much fresh air made it into the exhaust, we can compensate our measured exhaust lambda to get an idea of the in-cylinder lambda. This means we can preserve closed loop fueling in these conditions.

    It should be noted, blowthrough is really only calibrated by Ford on the Ecoboosts, a lot of the other vehicles have it disabled as it is unneeded. These applications didn't have the ability to dynamically vary overlap with older VCT, or they were NA applications that couldn't necessarily hit the conditions to have enough blowthrough to worry about. Roush and Whipple calibrations dont enable blowthrough logic even though they could see it, as they often aren't run in conditions where it would make enough of an effect that would warrant it. GT500 didn't have VCT and had a set overlap without blowthrough even under boost, so its uncalibrated as well.
    Look below for a section on whether or not you need to consider calibrating for it in your application.

    Okay, so enough about max trapped aircharge, what about how it actually calculates the air in the cylinder?
    We needed to get to the blowthrough part of the talk so that we could finally answer how the quadratic terms and blowthrough slope play together.

    Ford came up with this model after extensive measurement. Using the figure from their patent on this model with blowthrough (US 20130111900 A1):
    US20130111900A1-20130509-D00002.png

    The line representing our maximum trapped aircharge is line 202. 220 represents actual measured points of MAP and total aircharge. They noticed once blowthrough begins, the points start to follow a linear line, and so blowthrough has a single "slope" value. The points below that are quadratic in nature, and so have a quadratic fit (204).

    So, once you've reached conditions where blowthrough can happen (you'll notice it usually conveniently starts around when MAP > ExMAP), our in-cylinder aircharge becomes our max aircharge, and our extra airflow that fools the oxygen sensor is tracked via blowthrough slope.

    So, we kind of have three calculations going on:
    - Maximum trapped aircharge
    - If we are flowing more than maximum trapped aircharge, incylinder airflow is capped to the maximum aircharge. Blowthrough air is tracked via the slope from the point of where it started
    - If we are flowing less than the maximum trapped aircharge, incylinder airflow follows the quadratic curve. Blowthrough airflow is zero.

    As well, when solving the quadratic by inputting MAP and getting out aircharge, there are several cases:
    - If both roots are real and negative, or our quadratic term is zero, fall back to the linear model. (Solve the linear equation of MAP = Slope*Aircharge + Offset)
    - If both roots are real and positive, we use the smaller root.
    - If both roots are real and only one is positive, we use the positive root.
    - If both roots are imaginary, airflow is set to the maximum trapped airflow.

    As well, some of you may be aware that the quadratic term is "orthogonal". This means that if you zero it, you should still have a decent enough fit for the data. The quadratic term actually affects the slope and offset of the function while finding the quadratic solution. As such you cant just take a "normal" quadratic solution to the numbers you see, they need to be adjusted with respect to the quadratic as well.

    Once we do all this, we have a nice calculated aircharge, but its uncorrected. So, we correct for the same effects we did the maximum trapped aircharge. Air density, ExMAP (which is another model entirely that may need calibration), coolant temp, etc.
    This final corrected value is then our aircharge.

    This is really complicated
    We agree. That's why we made the calculator to handle all the background math the same way the ECU does, and calculate the correct "orthogonal" quadratic curve for you.

    Calibration
    Why do I need to calibrate this?
    The following systems use this to help predict aircharge/infer MAP, and as such dictate the importance of this calibration:
    • Manifold Filling/Transient - Inferred MAP/Aircharges use these values extensively
    • Throttle Body Control - Inferred MAP and aircharges are used to predict throttle airflow, as such are direct players in the throttle body control
    • Injector Flow Rate - Inferred MAP on Port Injected motors directly goes into the flow rate calculation
    • Sensor Failure - Even if your MAF fails, this acts like the modern Load With Failed MAF, amongst other things
    • Actual Airflow - On Speed Density applications like Ecoboost, this is the air model. You don't have a MAF to feed input into all the models, you have this.


    How-To
    You'll notice the table is RPM vs. MAP vs. Load. The idea is, you hit a target RPM and MAP in steady state, and adjust the load to either the actual load if you're a MAF car, or adjust it based on fuel trims assuming your fuel model is correct (You are using good injector data aren't you?) if you're an SD car.
    Lock your car in a given Mapped Point, or use the logged parameters to make sure you update the correct mapped point! It's really that straight forward.

    For example, if we're on a MAF Equipped Coyote, and we hit steady state, we may notice the load in our table is off from what we actually have measured via our installed MAP sensor and actual load. If we dial this in, we improve airflow anticipation and throttle control as the ECU can more accurately calculate airflows downstream of the MAF. By remaining in steady state, we aim to take out the effect any transient corrections based on the current model may have. Without a physical MAP sensor, this calibration is also just guess and check at that point. You cannot use the inferred MAP to try to correct it, as that's what you're trying to correct!

    For SD cars, this is all you have to estimate cylinder airflow! You need this dialed in as it is the vehicles air model, which everything else based on load is based off of. Its the only way the engine knows how much air is actually entering the cylinders.

    Blowthrough
    Calibration of blowthrough is more complicated. First, you need to decide if blowthrough is something that can happen on the vehicle, and if its effect is great enough that it warrants need. If it can, you need to enable it.
    How do you know? Well you have a good MAF curve / SD calibration, but you notice during conditions that could cause blowthrough, that your fuel trims go lean. This cant be MAF error (you dialed that in!)
    If you start to see this happen and you're solid on your MAF curve and its only happening for these airflows when your cam positions could be causing it, that's how you know its time to enable it.
    On SD, you calibrate blowthrough just like normal, its just more important your maximum aircharge line is dead on so its not actually injecting more fuel, but simply correcting the readings for blowthrough.

    Then, just adjust your load values in those cells with the percent you're off like normal, and you should notice it goes away. To be sure it's calculating it as "blowthrough" you can log InjPW or Fuel Mass injected to verify that's constant. You're seeing the effect of blowthrough "fixing" your readings.
    To properly calibrate this, we should mention, the load you see from the scanner WILL NOT match the load in the calculator during blowthrough. The Load in the scanner is incylinder aircharge, where as the calculator is showing a load based on total aircharge (including blowthrough)! The reason to do this is it allows you to still make the same percentage changes to get the desired effect and allows us to keep it all in one table.

    Caveats
    Why does my load suddenly spike or go negative at low MAP values?
    What you see is what the ECU would output in that given condition. Typically, what is happening is you are below the offset value. I.e. the MAP it shows would never be reached by the vehicle running at that RPM and cam position. Depending on whether it devolves to the slope based model (which will give you negative loads!) or caps it to max trapped airflow (suddenly larger loads), the calculation is still correct as to what the ECU would infer the load to be for those conditions (which wouldn't realistically be reached anyways!). Don't worry about calibrating these points unless you think you will have a valid value for them. The calculator attempts to intelligently work around them!

  2. #2
    Potential Tuner
    Join Date
    Nov 2014
    Posts
    9
    Playing with this on a Focus RS 2.3L Ecoboost GRCK25T.

    If I add 20% to Mapped point 0 within the speed density calculator, then calculate coefficients it results in 0's from 5250-6000 in MAP Slope VCT Mapped Point 0 Slope vs Airmass vs RPM.

    Is this correct?

    Stock values:
    inHg/lbm 0
    700 26527.69921875
    1000 25300.30078125
    1300 24802.76953125
    1400 24938.48046875
    1500 24459.80078125
    1700 23976.740234375
    2050 23481.5390625
    2200 23439.640625
    2500 23585.869140625
    2700 23204.7109375
    3100 22507.609375
    3300 22712.689453125
    3500 22126.509765625
    3600 21894.1796875
    3900 21402.5390625
    4100 20876.66015625
    4500 20589.779296875
    5250 20618.3203125
    5500 20578.5703125
    6000 21659.48046875
    rpm

    After 20% multiplier values:
    inHg/lbm 0
    700 22731.470703125
    1000 21415.84765625
    1300 20811.740234375
    1400 20804.77734375
    1500 20670.802734375
    1700 20289.056640625
    2050 19981.150390625
    2200 19845.162109375
    2500 19853.115234375
    2700 19665.103515625
    3100 19130.5703125
    3300 19255.228515625
    3500 18786.2265625
    3600 18435.8984375
    3900 18074.517578125
    4100 17781.966796875
    4500 17359.298828125
    5250 0
    5500 0
    6000 0
    rpm

  3. #3
    Quote Originally Posted by Nader View Post
    Playing with this on a Focus RS 2.3L Ecoboost GRCK25T.

    If I add 20% to Mapped point 0 within the speed density calculator, then calculate coefficients it results in 0's from 5250-6000 in MAP Slope VCT Mapped Point 0 Slope vs Airmass vs RPM.

    Is this correct?
    Yes. In this case the data you get back out (I tested myself) is correct. What youve essentially done by multiplying it all by 20% (20% more air at the SAME manifold pressure) is push everything past the maximum aircharge line at those RPMs. Notice now that you generate blowthrough slopes on every RPM now. You're hitting blowthrough everywhere much quicker, and the fit you get for your non-blowthrough region is degenerate. It only has a few points to match with incredibly small errors, and that solution gave the closest load numbers to what you entered into the calculator.

    All of the points you created are higher than the maximum trapped aircharge almost immediately.
    As such your fit ends up looking like this:

    tzEFfqV.png

    Notice blowthrough is happening at a load of 0.3ish or so. Are you really flowing so much air that its coming out the exhaust with so low load? Didn't think so, but that's what your changes are claiming.

    As explained, this model is not that simple. The numbers you get are correct for what info the calculator has available which put the max trapped aircharge BELOW your data points. You can't just hack up a curve and hope it works.

  4. #4
    Tuner in Training
    Join Date
    Jun 2016
    Posts
    42
    Quote Originally Posted by Steven@HPTuners View Post

    Caveats
    Why does my load suddenly spike or go negative at low MAP values?
    What you see is what the ECU would output in that given condition. Typically, what is happening is you are below the offset value. I.e. the MAP it shows would never be reached by the vehicle running at that RPM and cam position. Depending on whether it devolves to the slope based model (which will give you negative loads!) or caps it to max trapped airflow (suddenly larger loads), the calculation is still correct as to what the ECU would infer the load to be for those conditions (which wouldn't realistically be reached anyways!). Don't worry about calibrating these points unless you think you will have a valid value for them. The calculator attempts to intelligently work around them!
    This is the exact issue I've been running into. I've been working on a Whipple 2016 F150, and I have this problem. Load will suddenly drop, which causes the throttle body to open to increase load, then load spikes up, and begins a throttle oscillation until the throttle is tipped in more aggressively. This oscillation makes it pretty much impossible to log any sort of steady-state error. Is there any method that I could use to close the gap a bit so I can log some steady state error? For the record, torque tables are very close, and CLIP ADD Switch is currently disabled. So I'm confident the oscillation is not related to the torque model.

  5. #5
    Tuner in Training
    Join Date
    Aug 2016
    Location
    Blackpool, UK
    Posts
    46
    This is VERY impressive Work. I am going to have to sit and read this a few times to understand even 60% of it. Thanks for pushing the envelope with tuning tools guys. This must have been extremely complex to achieve.

  6. #6
    Advanced Tuner 15PSI's Avatar
    Join Date
    Feb 2016
    Location
    East Coast Somewhere
    Posts
    458
    Steven - Excellent write-up. This will be so helpful to all of us.
    2012 Mustang GT with S/C
    4Runner with S/C
    Turbo/NOS Hayabusa - 320RWHP

  7. #7
    Advanced Tuner
    Join Date
    Jun 2008
    Location
    Tyler/Longview, TX area
    Posts
    746
    Thank you! Thank you! Thank you! Awesome information!
    Last edited by txcharlie; 11-03-2017 at 12:26 PM.

  8. #8
    Advanced Tuner
    Join Date
    Jun 2008
    Location
    Tyler/Longview, TX area
    Posts
    746
    ....

  9. #9
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Quote Originally Posted by Steven@HPTuners View Post
    Yes. In this case the data you get back out (I tested myself) is correct. What youve essentially done by multiplying it all by 20% (20% more air at the SAME manifold pressure) is push everything past the maximum aircharge line at those RPMs. Notice now that you generate blowthrough slopes on every RPM now. You're hitting blowthrough everywhere much quicker, and the fit you get for your non-blowthrough region is degenerate. It only has a few points to match with incredibly small errors, and that solution gave the closest load numbers to what you entered into the calculator.

    All of the points you created are higher than the maximum trapped aircharge almost immediately.
    As such your fit ends up looking like this:

    tzEFfqV.png

    Notice blowthrough is happening at a load of 0.3ish or so. Are you really flowing so much air that its coming out the exhaust with so low load? Didn't think so, but that's what your changes are claiming.

    As explained, this model is not that simple. The numbers you get are correct for what info the calculator has available which put the max trapped aircharge BELOW your data points. You can't just hack up a curve and hope it works.
    Wouldn't the quadratic term zero out, not the slope term zeroed? Mx+b not Ax^2+C which is still a curved line.

  10. #10
    Tuning Addict
    Join Date
    Jan 2015
    Location
    Franklin, NC / Gainesville, Ga
    Posts
    6,696
    Is there a write up anywhere on how to tune in the SD? Such as using fuel trims, afr errors, load errors or whatever you need to do this? I also understand you need to use a separate external map to get correct data? Any recommendations on any of this? Knew I should have jumped deeper into Fords long ago...
    2010 Vette Stock Bottom LS3 - LS2 APS Twin Turbo Kit, Trick Flow Heads and Custom Cam - 12psi - 714rwhp and 820rwtq / 100hp Nitrous Shot starting at 3000 rpms - 948rwhp and 1044rwtq still on 93
    2011 Vette Cam Only Internal Mod in stock LS3 -- YSI @ 18psi - 811rwhp on 93 / 926rwhp on E60 & 1008rwhp with a 50 shot of nitrous all through a 6L80

    ~Greg Huggins~
    Remote Tuning Available at gh[email protected]
    Mobile Tuning Available for North Georgia and WNC

  11. #11
    Quote Originally Posted by OzTuning View Post
    This is the exact issue I've been running into. I've been working on a Whipple 2016 F150, and I have this problem. Load will suddenly drop, which causes the throttle body to open to increase load, then load spikes up, and begins a throttle oscillation until the throttle is tipped in more aggressively. This oscillation makes it pretty much impossible to log any sort of steady-state error. Is there any method that I could use to close the gap a bit so I can log some steady state error? For the record, torque tables are very close, and CLIP ADD Switch is currently disabled. So I'm confident the oscillation is not related to the torque model.
    You can try to infer what the load should be for those areas and generate coefficients for it and see if it gets you close instead of shooting the load up on you.

    Quote Originally Posted by murfie View Post
    Wouldn't the quadratic term zero out, not the slope term zeroed? Mx+b not Ax^2+C which is still a curved line.
    In this case as you can see, there's still a curved line fit on the lower values. The linear solver we use in this case will keep a curved fit because you have to remember, the way the Ford logic works, it does not become Ax^2+C. The Quad when run through the actual algorithm creates offsets to the slope and offset, so in this case its actually Ax^2 + (0 + B_offset)x + C + C_offset.

    As such there's still a "hidden" linear term there.

    Quote Originally Posted by GHuggins View Post
    Is there a write up anywhere on how to tune in the SD? Such as using fuel trims, afr errors, load errors or whatever you need to do this? I also understand you need to use a separate external map to get correct data? Any recommendations on any of this? Knew I should have jumped deeper into Fords long ago...
    The very small how-to section. Its pretty straightforward. Error % being multiplied onto the values on SD applications, or using actual load at steady state when you measure MAP on MAF applications.

  12. #12
    Tuning Addict
    Join Date
    Jan 2015
    Location
    Franklin, NC / Gainesville, Ga
    Posts
    6,696
    I'm sorry, but I must be going blind. Do you have a link. Didn't see anything mentioning ford SD in the how to section.
    Thank you again.
    2010 Vette Stock Bottom LS3 - LS2 APS Twin Turbo Kit, Trick Flow Heads and Custom Cam - 12psi - 714rwhp and 820rwtq / 100hp Nitrous Shot starting at 3000 rpms - 948rwhp and 1044rwtq still on 93
    2011 Vette Cam Only Internal Mod in stock LS3 -- YSI @ 18psi - 811rwhp on 93 / 926rwhp on E60 & 1008rwhp with a 50 shot of nitrous all through a 6L80

    ~Greg Huggins~
    Remote Tuning Available at gh[email protected]
    Mobile Tuning Available for North Georgia and WNC

  13. #13
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Its in the original post of this thread. Third paragraph/section from from the bottom.

  14. #14
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Quote Originally Posted by Steven@HPTuners View Post
    In this case as you can see, there's still a curved line fit on the lower values. The linear solver we use in this case will keep a curved fit because you have to remember, the way the Ford logic works, it does not become Ax^2+C. The Quad when run through the actual algorithm creates offsets to the slope and offset, so in this case its actually Ax^2 + (0 + B_offset)x + C + C_offset.

    As such there's still a "hidden" linear term there.
    So line 204 is calculated using the conventional quad, slope, and offset. thats what I see in the lower values. Above point 208, when calculated line 204 and the maximum line intersect, curve 210(blow through line) is calculated. Line 210 which starts from point 208 is calculated. According to the patent that line does not have a quad coefficient and uses point 208 as an offset, bold in the equation below. Is ford using the quad term when calculating the blow through not reverting to the non quadratic equation, like in this patent?

    P inf  ( k ) =  P exh  ( k ) 30  [ in - Hg ]  max  { 1 c norm * m c  ( k ) , [ 1 c norm * m c  ( k ) +slp * ( m air x  ( k ) - m c  ( k ) ] }


    I just downloaded a focus tune and tried it. I think your calculator is putting the slope term in the quadratic and zeroing out the slope.

    SD Focus.PNG
    Last edited by murfie; 11-03-2017 at 09:03 PM.

  15. #15
    Tuning Addict
    Join Date
    Jan 2015
    Location
    Franklin, NC / Gainesville, Ga
    Posts
    6,696
    Sorry Murfie - thought he was talking about a different How to - didn't read it close enough... Thanks...
    2010 Vette Stock Bottom LS3 - LS2 APS Twin Turbo Kit, Trick Flow Heads and Custom Cam - 12psi - 714rwhp and 820rwtq / 100hp Nitrous Shot starting at 3000 rpms - 948rwhp and 1044rwtq still on 93
    2011 Vette Cam Only Internal Mod in stock LS3 -- YSI @ 18psi - 811rwhp on 93 / 926rwhp on E60 & 1008rwhp with a 50 shot of nitrous all through a 6L80

    ~Greg Huggins~
    Remote Tuning Available at gh[email protected]
    Mobile Tuning Available for North Georgia and WNC

  16. #16
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    Quote Originally Posted by GHuggins View Post
    Sorry Murfie - thought he was talking about a different How to - didn't read it close enough... Thanks...
    No worries.
    Its still in beta so there will be "bugs" to work out and to watch for. They give us the tools, its up to one of us to write a good "How to" on using it. Seeing as its just been release to the public I would give it some time for people to come up with more in depth guides.

  17. #17
    Quote Originally Posted by murfie View Post
    So line 204 is calculated using the conventional quad, slope, and offset. thats what I see in the lower values. Above point 208, when calculated line 204 and the maximum line intersect, curve 210(blow through line) is calculated. Line 210 which starts from point 208 is calculated. According to the patent that line does not have a quad coefficient and uses point 208 as an offset, bold in the equation below. Is ford using the quad term when calculating the blow through not reverting to the non quadratic equation, like in this patent?

    P inf  ( k ) =  P exh  ( k ) 30  [ in - Hg ]  max  { 1 c norm * m c  ( k ) , [ 1 c norm * m c  ( k ) +slp * ( m air x  ( k ) - m c  ( k ) ] }


    I just downloaded a focus tune and tried it. I think your calculator is putting the slope term in the quadratic and zeroing out the slope.

    SD Focus.PNG
    Blowthrough slope is separate and still there. When you're in blowthrough, the incylinder airflow is the maximum trapped, the total airflow is that plus the blowthrough slope from that point.

    The calculator is just that, a calculator. Its following the exact logic the ECU is to give you the values you see. The problem with the zero slope popping up is the input data is way far advanced of maximum trapped airflow, and to fit that data the linear equation solver is degenerating. It's not realistic in the slightest to raise the whole table 20%, it's not a bug, its just what the solver does to get the OUTPUT of running the coefficients through Ford's algorithm to match as close as it can to what you give it.

    Garbage in, garbage out. The calculator will just try to match the data you give it to Ford's model, its up to you to actually measure it and be smart enough to know when maximum trapped airflow should be bumped or if what you're seeing is even realistic. Raising the whole table 20% is not realistic, even individual points that much, you're not going to suddenly do that. That's a huge VE change, indicating that the cylinders maximum trapped airflow is now almost certainly higher. This may be because of raised standard aircharge (more displacement or much more efficient current displacement), or need to be corrected in the Maximum Aircharge Multiplier because you performed some magic.

  18. #18
    Advanced Tuner Auto-Hentzschel's Avatar
    Join Date
    Sep 2015
    Location
    Schleid, Germany
    Posts
    341
    I played also with the calculator. But what ever I do, add/remove 1% 10% or something else. Everytime the calculator zero out the Mapping point's area where I do the changes. Where is my mistake?
    SD_calc.jpgSD_calc_error.jpg

  19. #19
    Tuning Addict
    Join Date
    Jan 2015
    Location
    Franklin, NC / Gainesville, Ga
    Posts
    6,696
    Auto, you plugging in a decimal or a comma 01? Mine doesn't do any of that unless I use a comma and even then it defaults back to original numbers? Which update are you using?
    2010 Vette Stock Bottom LS3 - LS2 APS Twin Turbo Kit, Trick Flow Heads and Custom Cam - 12psi - 714rwhp and 820rwtq / 100hp Nitrous Shot starting at 3000 rpms - 948rwhp and 1044rwtq still on 93
    2011 Vette Cam Only Internal Mod in stock LS3 -- YSI @ 18psi - 811rwhp on 93 / 926rwhp on E60 & 1008rwhp with a 50 shot of nitrous all through a 6L80

    ~Greg Huggins~
    Remote Tuning Available at gh[email protected]
    Mobile Tuning Available for North Georgia and WNC

  20. #20
    Senior Tuner
    Join Date
    Jan 2013
    Location
    Hawaii
    Posts
    2,101
    I'm finding if I don't smooth out the new load values, It will either revert back to what it was or replace the whole column with 0's depending how far off or close they were.
    Other times if I don't smooth it before calculating new coefficients and it does give me new values that are very choppy or leave random large load holes or spikes if I didn't change enough cells.

    3D New coe MP 25 VE.PNG

    Kind of likes a balance of smoothing enough cells and calculating coefficients until all MAP values are close to the loads seen. It certainly gets things closer than calc. MAP max and cylair multipliers alone for FI.

    Stock VE bumps removed.PNG

    MP 25, 4350-5250 RPM, 22-23PSI MAP, Trying to get smooth load at 1.4 load and not have it indicate the two spikes as the stock values do.