Ecoboost Speed Density tuning

[Blown383]

Alright, I've got a pretty firm grasp on how to manage boost on an Ecoboost. The part that is a bit over my head is the fueling side. Most tunes I've seen from most major shops have not touched the speed density tables at all and rely solely on the closed loop operation to manage fueling. I've been tweaking on my personal Focus ST and I've hit a point where I need more fuel but I look at the speed density tab and cannot put my head around where I need to add my correction. I understand that each point on each table for each mapped point is a quadratic formula that plots a curve to predict fueling. If I need 20% more fuel somewhere, where do I add it? I'm running into a problem where the ECU cannot add the fuel fast enough and it runs lean for a bit before it catches up. I either need to make the PID controller more sensitive and respond faster for making corrections or do it the right way and put the fuel in. If anyone has any insight on how to adjust fueling, please post. I know there have been several posts about this but so far they have all been heavy on the math and no answers on exactly what to do. I'm still waiting on Eric to wrap up the logger for 2015+ Focus ST so I can see which mapped points to add fuel. Please share if you've had success adding fuel via speed density on an Ecoboost. Thanks!

[metroplex]

We had several threads about this, but bottom line is that the best/most you will find here are theoretical discussions on how the speed density system works (or links to the patents) but I have yet to see any practical application on how to actually correct the fueling.

[Blown383]

Yeah that's my problem. I have a basic understanding on how the system works but how to add fuel is beyond me. It seems like most companies just add additional fueling with aux systems or meth and leave the tables stock. Seems kinda hackish to me. I understand using such systems for increasing beyond the factory HPFP limits but I've only added like 3psi. There's gotta be a way.

[Blown383]

Alright, so after quite a bit of experimenting, it looks like we can apply corrections to the slope. You have to apply the additive inverse to the slope, i.e. fuel trims +25%, take -25% to the slope. I don't have mapped points available to log on my Focus ST yet but I did manage to get the fuel trims quite a bit closer applying the correction to all the tables. Now all we need is HPTuners to add weighted histograms so we can have a histogram for each mapped point and dialing in the fuel should be easier, or at least get it close enough because the Tri-Core ECUs do an excellent job at applying closed loop corrections. If anyone has a better way, I'd love to hear it.

[metroplex]

Are we monitoring STFT by itself, or taking STFT+LTFT?

[Blown383]

STFT+LTFT is best. I really need to have mapped points available to log to see if this method is best. So far it seems most promising compared to other parameters I tried.

[metroplex]

Looking at slope, there are only values based on mapped point and RPM. If we make changes to the values, wouldn't it impact the fuel trims for all of the loads at that RPM?

[Higgs Boson]

Alright, so after quite a bit of experimenting, it looks like we can apply corrections to the slope. You have to apply the additive inverse to the slope, i.e. fuel trims +25%, take -25% to the slope. I don't have mapped points available to log on my Focus ST yet but I did manage to get the fuel trims quite a bit closer applying the correction to all the tables. Now all we need is HPTuners to add weighted histograms so we can have a histogram for each mapped point and dialing in the fuel should be easier, or at least get it close enough because the Tri-Core ECUs do an excellent job at applying closed loop corrections. If anyone has a better way, I'd love to hear it.

you make your own histograms/charts.

[metroplex]

For tuning MAF systems I normally log LTFT+STFT against RPM and load. I see that on my SHO when it was factory stock, the fuel trim is lean at certain regions by up to 13%. This didn't change with the added boost. If I go into the slope tables and make the changes, it would seem to impact that entire RPM row regardless of load.

So if I am 13% lean at 4k RPM at load of 1.5, but I am say 0% at the same RPM but at load of 0.9, the changs I make at 4k would mess it up at 0.9 load...

Someone else suggested using the fuel imbalance tables that do break it out by cylinder, load, and RPM. I have not experimented with either route, just trying to learn.

[Blown383]

you make your own histograms/charts.

Weighted histograms so there is a histogram for each mapped point and the logger automatically does the weighting/blending between the points for each one. Rather than 14-26 histograms all showing similar data.

[Higgs Boson]

Weighted histograms so there is a histogram for each mapped point and the logger automatically does the weighting/blending between the points for each one. Rather than 14-26 histograms all showing similar data.

What we need is a Virtual VE table like we have for GM speed density applications.

[Blown383]

Well Eric got the logger out for me and I dialed in the fueling applying corrections to the slope. It looks like it works perfectly! I managed to increase boost and pretty much max out the turbo. I'm keeping timing relatively conservative as the car will be receiving methanol and an intercooler soon. I did get the fueling down to +/-5% using this method. Metroplex, since the mapped point weighing changes with varying load, it doesn't ever get excessively rich or anything using this method. I'm sure you could find a certain situation where it may move into one of the modified mapped points but the fuel trims should take care of the rest. After I dialed in the WOT in different gears and varying loads, fueling above 2500 at low load stayed relatively consistent and there were no extreme trims. I believe weighted histograms that correspond to the mapped points weights would be extremely beneficial to tuning an Ecoboost. It would simplify tuning timing and fueling, imo. Pretty excited about this! Thanks for your hard work Eric!

[metroplex]

Awesome! Is HPT planning to release this within VCM Scanner or is it a separate program/spreadsheet?

[Blown383]

It's just an idea to help tune an Ecoboost. Hopefully someone from HPT will see this and add it!

[murfie]

These tables all show a part of manifold absolute pressure in "pressure of air mass" in slightly different ways and where the pressure is going to go. As your MAP(therefore load) varies the different air mass values (derived from PV/RT=n) of the pressure measured by MAP sensor. The calculated air mass is then multiplied by the numbers in these tables as the "x" value in (Quad)x^2+(slope)x+(btslope)x+(offset)=MAP. The ECU compares the calculated MAP and actual MAP and adjust fuel appropriately after the complex cylinder air calculation.

So yes while modifying these values will affect all loads at given RPM, Varying loads at given RPM will also affect the final calculated MAP. The offset is the air that was left in the manifold, the blow through slope is the air that went straight to the exhaust, The slope is the main amount of air that got into the cylinder and therefore will have the most affect on the fuel added, and quadratic is also air that got into the cylinder to a finer degree so it will have an affect on fuel just not as great.

You can try set it up in the scanner to compare your calculated manifold pressure and absolute manifold pressure. Start by correcting the slope values to get fuel roughly close with commanded and actual WB EQ then use quadratic tables to fine tune it. The remaining difference will be either added in the Blow through slope when calculated is higher then measured too bring it down. Adding offset when measured is higher than calculated.

Weighted histograms/charts are already a thing you use the filtering input and add a function to the histogram/chart like PID MP 2 needs to be greater than 30%.

[gudmundur]

Thank you all for this good information.
I am helping a friend with a Focus RS 2016.

After increasing boost to 1.5 bar we need to add some fuel and as I understand the first step would be to add fuel to Slope tables.
But there are 16 of them 0-14 and optimal power, any idea on how we can find out when each table is used ?
Ford RS speed density.JPG

[metroplex]

How do you know you need more fuel? Maybe the SHO is different, but when factory stock with 11 psi max boost on 87 octane (about 9-10 with 93 octane), the LTFT+STFT was as high as +13% in some RPM/load regions but this pretty much stayed the same when I increased boost to 18 psi. I never had to touch any of the speed density settings. I even kept the fuel pump pressure settings stock.

[murfie]

Log your MP weights or get/wait for support to add them to your car.
MPweights.PNG

use the predefined math "EQ ratio error" and set up 16 graphs with a filter for each mapped point against the RPM range of your SD tables.
EQ error v MP.PNG

you can modify the percentage or cell hits required for as much or as little accuracy as you want.

[murfie]

The SD is there for sensor check and sensor failure. In GMs its popular to use it as sensors like the MAF limit the size of the intake. It also can be much more accurate than information sensors give you. It is not necessary on most ford tuning.

[metroplex]

I only have 6 mapped points on the SHO (0 thru 5). What is the value you are using for the filtering? A mapped point?

We then apply these corrections to the slope or quadratic for the corresponding MP? Are the adjustments rhe inverse of the fuel trims?