Hello
Can someone explain, how the quatratic term and the slope tables need to tune?
I read a lot about it, mainly with forced induction, but it looks for the most tuners is this part like a black hole and they don't touch this tables. But it seems to be very important to adjust them.
Or its possible to get a guide for that anywhere?
thanks Henry!
You can gain a bit of understanding by starting with these posts:
https://www.hptuners.com/forum/showt...uadratic+slope
2012 Mustang GT with S/C
4Runner with S/C
Turbo/NOS Hayabusa - 320RWHP
https://www.easycalculation.com/stat...regression.php
Heres a website you can use to find to regression line based on your collected data.
https://www.varsitytutors.com/hotmat...tic-regression
This explains what that website is doing.
Y variables would be MAP values in inHg
X variables would be Lbm values of air
Map you can get with an external gauge. Make sure it's inHg not PSI like mst gauges are for boost.
LBm you can get from (LB/min)/(REV/MIN) to get LB/rev. There's 2 revolutions per intake stroke so divide that by 2. You will find LBm will follow engine load.
These values would be from different RPMs trying to keep the RPM as steady as possible while varying MAP pressure.
You can try out some made up values in the website to see how the curves in the lines you can make are effected.
.0001- .006 is a typical range of LBm and 8 - 30inHg for MAP in a NA application. The lines best fit when you have a full range of values.
Attached is some made up numbers I put in the website to show you the line and values of A,B,C you get.
Quadratic regression.PNG
Last edited by murfie; 07-17-2017 at 04:29 PM.
thanky guys,
thats a lot of input and a lot to read. But its a begin to understad it better.
Since the FRPP/Roush tunes are cracked and the speed density values we can use in our own boosted tunes, the cars will going better possible to drive. Thats my reason to understand it how it works.
@murfie, MAP is a PID in the Scanner, why should I use a external gauge?
you can use the inferred MAP pid, but a gauge is best to make sure you are correcting and not making things worst.
Okay, my Dyno has a external pressure gauge. I can test and compare with the scanner paid.
Make sure you bump up max MAP in the tune in the Speed Density section or it will never infer higher than what's in that field. I'm guessing you already know this but it can help others reading this too.
@murfie
if I have the variables for MAP and LBm, where I get the values for the quatratic term tables? your recommed website show only lines?
Is your data filtered per mapped point and per RPM?
It gives you the best fit equation Ax^2+Bx+C and a nice graph to see what the lines like.
A- Quadratic
B- Slope
C- Intercept
Quadratic regression.PNG
...not 100%sure it I understand your right. maybe you can explain it on a actual speed density calculation from a stockfile? step by step?
Sorry for the late response. This data is collected from my 2015 GT with Steeda intake and MBRP exhaust.
SD MP1 500-1325RPMs.Layout.xml
Here's an example of a Graph layout used to collect the data needed. I had to used calculated MAP as I did not have a external sensor hooked up. Its better to use a external sensor for more accurate results.
You will need to create a user math. Heres a snip of my LBm.
usermath LBm.PNG
Example of outcome data:
1325RPM MP1 Graph.PNG
You have to manually enter data in website as "," are needed and copy pasting doesnt work. luckily You only will need to change the "x" data.
1325RPM MP1 values into quad reg.PNG
This is what would be applied to the tune, but for all RPM's filtered by MP.
Before:
1325RPM MP1 SD before.PNG
After:
1325RPM MP1 SD after.PNG
Its very time consuming. For the best results use a MAP range the MP you are working on may actually apply to.
Last edited by murfie; 07-22-2017 at 09:44 PM.
I had the filters wrong in that layout. Heres an updated one.
SD MP1 500-1325RPMs.Layout.xml
Another example for multiple MPs
SD MP1 500-1325RPMs and MP2 1025-1325RPMs.Layout.xml
From these you can see where its hitting the MP more and increase the resolution to have more data to put into the website.
Last edited by murfie; 07-23-2017 at 01:18 AM.
@murfie
thanks for your explaination. start to understand, how it works.
But its difficult.
I have some questions:
1. Why you not use the Manifold absolute pressure PID, rather the calculated?
2. How I understand it right, I need to enable only one MP and collect the LBm values for every RPM and evry MAP!?
then I enable the next MP and do the same again?
Did you get any significant differences after calculate quatratic term, slope and offset?
Thanks Henry
1. My car does not have a MAP sensor, so any manifold absolute pressure is calculated from the MAF and manifold volume. If your MAF curve is wrong so will your manifold pressure PID. Its best to just use an external sensor and get actual values.
2. Enabling only one MP and collecting LBm is one way of doing it. I think cobb suggest tuning like that. I just filter the data I collect by MP and RPM. either way the more data you get for a range of manifold pressure values the better the regression numbers will be.
The only time I get a significant difference is when I don't have a wide enough range of MAP or have too wide of a range of MAP. The more resolution in MAP values and data points the better the regression will fit.
Wide range:
1,2,3,4,5,6,7,8,9,10,15,20,25,30,35,40
More resolution:
8,8.1,8.2,8.3,8.4,8.5,8.6,8.7,8.8,8.9,9,9.2,9.4,9. 6,9.8,10,10.3,10.6,11,11.5,12,12.5,13,13.5,14,14.5 ,15
I think, the best way is to collect the data by enable only one MP. Step by step.
Today I played a litle bit with and I think more and more I understand it. Maybe on a NA engine you need not change it the factory values are good. but if installing a supercharger, then we need to calculate it new.
I saw, Roush enabled only MP 14-26 in theyr's tune. but with the roush calculated speed density values the supercharger is much more drivable than any VMP or other tuners version.
I work often with KenneBell SC. They enable all MP's like the factory settings and don't touch the speed density, only the slope values. But not a new calcuation rather a multipy by...% calculation. I think there is room for more power!
how much MAP should I use for SC engine?
Do you think, the website equation what calculate the quatratic term is simply put in a excel sheet? For easy using copy and paste?
Henry
Putting it into a spreadsheet can be done. I might play with the idea later if I find time.
Formula:
Quadratic Regression Equation(y) = a x^2 + b x + c
a = { [ Σ x2 y * Σ xx ] - [Σ xy * Σ xx2 ] } / { [ Σ xx * Σ x2x 2] - [Σ xx2 ]2 }
b = { [ Σ xy * Σ x2x2 ] - [Σ x2y * Σ xx2 ] } / { [ Σ xx * Σ x2x 2] - [Σ xx2 ]2 }
c = [ Σ y / n ] - { b * [ Σ x / n ] } - { a * [ Σ x 2 / n ] }
Where ,
Σ x x = [ Σ x 2 ] - [ ( Σ x )2 / n ]
Σ x y = [ Σ x y ] - [ ( Σ x * Σ y ) / n ]
Σ x x2 = [ Σ x 3 ] - [ ( Σ x 2 * Σ x ) / n ]
Σ x2 y = [ Σ x 2 y] - [ ( Σ x 2 * Σ y ) / n ]
Σ x2 x2 = [ Σ x 4 ] - [ ( Σ x 2 )2 / n ]
x and y are the Variables.
a, b, and c are the Coefficients of the Quadratic Equation
n = Number of Values or Elements
Σ x= Sum of First Scores
Σ y = Sum of Second Scores
Σ x2 = Sum of square of First Scores
Σ x 3 = Sum of Cube of First Scores
Σ x 4 = Sum of Power Four of First Scores
Σ xy= Sum of the Product of First and Second Scores
Σ x2y = Sum of Square of First Scores and Second Scores
SLOPE = COVAR(MAP Ref, Aircharge) / VARP(Aircharge)
OFFSET=AVERAGE(MAP Ref) - Calculated Slope * AVERAGE(Aircharge)
That's what I use for the formulas in an Excel solution.
Last edited by tbrtuning; 07-23-2017 at 02:18 PM.
Do you have a graph or a layout for that? What pid numbers you use? How do you calculate the quatratic term?
My spreadsheet is confusing because I have a bunch of stuff in it from testing different things and have never made a better version. I honestly don't even use it for the Coyote either. I just turn off Cylair Anticipation and never mess with it. I'll see if I can put something together in a few minutes with only the basic things you would need.
Reply With Quote