Sent
802SHO 2010 Build
- Thread starter 802SHO
- Start date
Disclaimer: Links on this page pointing to Amazon, eBay and other sites may include affiliate code. If you click them and make a purchase, we may earn a small commission.
Exactly. The goal is to scale the model to allow the new safe power and stop intervening but to not disable it. We want it intelligently protecting. Right now it’s over protecting and intervening to the lowest possible power rating the new hardware makes.
SM105K
Streetlight Grand Prix Champ/ IG @fafomotorsports
Who are you and what are you driving now? I see AZ and terminator and interests peak.
One thing I’ve been digging into recently is the relationship between airflow scaling, torque scaling, and inverse torque modeling on these early EcoBoost torque-based strategies.
The simple answer is supposed to be easy. Increase airflow capacity. Increase torque limits. Let the software calculate the inverse side. Done.
The problem is these ECUs aren’t calculators. They’re predictive torque management systems. The PCM is constantly trying to determine how much torque it thinks the engine is making, how much torque the airflow should represent, whether the airflow agrees with the model, and whether intervention is necessary.
What my buddy Brendan explained to me a while back was interesting. He said on these early EcoBoost strategies some tuners would scale the main torque and load side upward while allowing the software to automatically populate the inverse side. The math looked correct, but the behavior didn’t. In practice the car would sometimes target less airflow, intervene sooner, or generally become less cooperative even though everything appeared correct on paper.
His experience was that the inverse side often needed to be scaled upward along with the torque side so the model remained coherent. In other words, both sides of the equation had to move together.
The more I look at recent logs, the more I think about that conversation.
Airflow is improving. Actual load is improving. MAP and TIP are getting closer together. The hardware is clearly moving in the right direction. Yet the system is still showing torque intervention behavior.
ETC Torque Request sits around 369. Brake Torque Desired sits around 369. Engine Brake Torque approaches that area. Spark Source starts reporting Torque Reduction. Throttle gets chopped. Scheduled Torque continues higher.
I’m not claiming that 369 is definitely a hard limit. I’m not claiming this is the answer. What I am saying is that the airflow side of the model appears to be improving faster than the torque approval side.
What I found encouraging is when I brought this up to Ryan, he immediately knew what I was talking about and explained that he manually handles the scaling rather than relying on the software to automatically populate values. He literally said, “I have a separate excel table I do everything in and then I import it. I don’t let the software auto scale anything.” That was actually reassuring to hear because it tells me he’s looking at behavior, not just math.
At the end of the day, the goal isn’t to make a spreadsheet happy. The goal is to make airflow, load, torque, throttle, spark, and shift behavior all tell the same story.
Because on a strategy developed in 2009, running hardware that would have looked completely ridiculous in 2009, that’s probably the entire battle.
The simple answer is supposed to be easy. Increase airflow capacity. Increase torque limits. Let the software calculate the inverse side. Done.
The problem is these ECUs aren’t calculators. They’re predictive torque management systems. The PCM is constantly trying to determine how much torque it thinks the engine is making, how much torque the airflow should represent, whether the airflow agrees with the model, and whether intervention is necessary.
What my buddy Brendan explained to me a while back was interesting. He said on these early EcoBoost strategies some tuners would scale the main torque and load side upward while allowing the software to automatically populate the inverse side. The math looked correct, but the behavior didn’t. In practice the car would sometimes target less airflow, intervene sooner, or generally become less cooperative even though everything appeared correct on paper.
His experience was that the inverse side often needed to be scaled upward along with the torque side so the model remained coherent. In other words, both sides of the equation had to move together.
The more I look at recent logs, the more I think about that conversation.
Airflow is improving. Actual load is improving. MAP and TIP are getting closer together. The hardware is clearly moving in the right direction. Yet the system is still showing torque intervention behavior.
ETC Torque Request sits around 369. Brake Torque Desired sits around 369. Engine Brake Torque approaches that area. Spark Source starts reporting Torque Reduction. Throttle gets chopped. Scheduled Torque continues higher.
I’m not claiming that 369 is definitely a hard limit. I’m not claiming this is the answer. What I am saying is that the airflow side of the model appears to be improving faster than the torque approval side.
What I found encouraging is when I brought this up to Ryan, he immediately knew what I was talking about and explained that he manually handles the scaling rather than relying on the software to automatically populate values. He literally said, “I have a separate excel table I do everything in and then I import it. I don’t let the software auto scale anything.” That was actually reassuring to hear because it tells me he’s looking at behavior, not just math.
At the end of the day, the goal isn’t to make a spreadsheet happy. The goal is to make airflow, load, torque, throttle, spark, and shift behavior all tell the same story.
Because on a strategy developed in 2009, running hardware that would have looked completely ridiculous in 2009, that’s probably the entire battle.
Zinprogress
Member
I’m a 39yr old feral car guy tragically turned to a domesticated man bitch. I went from whipples, turbos, nitrous and high 7’s to a Buick enclave over night lol. I had an 03 cobra about 14 years ago, it’s one of the 2 out of many cars I regret ever selling. I live in Peoria, currently just have the 13 SHO. That’ll change very soon as kids are getting older.
Zinprogress
Member
When I say that’ll change I don’t mean the Taurus, I love the car and I honestly can’t see myself parting with it. I’d like to do a C7 with fender mounted turbos. Try to keep it as useable as possible. We will see.
Like Feras Qartoumy’s Time Attack Corvette C6 turbo placement?
Zinprogress
Member
Yeah, something similar. I’ve had a build in my head for several years now that changes as the technology does. Originally it was a C6, budget and cost has upgraded that to a C7. I have at least one last grand build left in me, just need to tie off a few more obligations first. New edges still catch my eye so we will see where that goes.
Bronco2fan
SHO Member
Just need to wrap them now more or less. 
DadMobile
Worlds 3rd least slowest Ecoboost SHO
Dang, I figured it would be running down the track with all the amazing progress y’all have made. Super duper cool steering wheel!! I love all the lights, what was your inspiration for the interior design, stunning. I’ll check back in another week. I’m super hopeful, you can do it!
I panel 90% complete. Just need to use solid rivets on the WG dump metal protection plate. 
Gloss black PPF. Matte black PPF Ghostface graphic. That’s with flash on so it’s actually more stealth.
Gloss black PPF. Matte black PPF Ghostface graphic. That’s with flash on so it’s actually more stealth.
DadMobile
Worlds 3rd least slowest Ecoboost SHO
Ok, I have to admit, this is so gnarly I had to log back in just say comment on how rad your latest modification is. You’re building a unique race car with UNLIMITED potential and MAXIUM style that nobody ever in this world will be able to replicate.
DadMobile
Worlds 3rd least slowest Ecoboost SHO
Thanks for sharing more pictures, will you be selling this super dope design??? Where can I get one?!
Ryan sent a new map. I have more to share about the calibration, my theory and conversation with Ryan. Stay tuned
One thing about me is once something doesn’t make sense, I can’t leave it alone. If something isn’t right I’m stuck in diagnostic mode, I can’t turn it off. My brain is stuck in HUNT.
That’s how I found the grounds. That’s how I found the PCM power issues. That’s how I found the fuel system issues. That’s how I found the BOV problem. It’s how I found half the stuff on this car.
So when I started looking at these latest logs and saw the PCM absolutely strangling the car at only 15 psi on wastegate spring, my brain immediately went into diagnostic mode. The calibration window looked tiny. Almost absurdly tiny. To the point where I started joking that it looked like the car was being tuned like a single turbo G25-550 4-cylinder Focus instead of a twin G25 3.5L V6.
Half joking, Then I started looking at the data. The more I looked, the less funny it became.
The logs were showing the car getting forced into the same airflow numbers over and over again. Turbo Airflow and Average Air Mass were flatlining at 43.23 lb/min while the PCM was simultaneously closing the throttle, pulling spark, triggering Engine Speed Limit intervention, and throwing overspeed codes.
The hardware kept saying one thing. The PCM kept saying another. That’s when I brought it to Ryan. My theory was simple: the calibration window looked way too small for what the hardware was physically capable of doing.
Ryan pushed back initially saying this is why he prefers live tuning and showed me the airflow tables. Fair enough. I wasn’t looking at tables though. I was looking at what the car was actually doing. I said I’m only looking at the log. I said, “I see the screenshot man, and yeah I see the load isn't capped there. I was only looking at the raw data from the log. The datalog is literally showing the turbo airflow flatlining hard at exactly 43.23 lb/min for two whole seconds right before it shuts the throttle. The data is definitely showing a hard clip somewhere, so it has to be buried in one of those turbo or speed density limits.”
After some back and forth, his response got my attention. He doesn’t think it’s a speed density limit. He thinks it may be a turbo FMEM related limiter and specifically mentioned running into situations on 4-cylinder applications where not all the tables were fully defined.
That was the first thing I’d heard that actually lined up with what the logs were showing. The interesting part isn’t whether my theory was exactly right, the interesting part is that the data pointed us toward something real. That’s the whole reason I keep digging.
I stare at logs for hours because every major problem I’ve solved on this car started exactly the same way. Something didn’t fit reality. The difference this time is the problem isn’t mechanical anymore. We’re teaching the PCM what this hardware actually is. And honestly, I think that’s the step that was skipped years ago. The original approach was basically “add boost.” What we’re doing now is making the model agree with reality first.
After watching the PCM fight 15 psi this hard, I understand now why that step matters so much. New map is loaded. The hunt continues. Ryan promised faster response times too.
That’s how I found the grounds. That’s how I found the PCM power issues. That’s how I found the fuel system issues. That’s how I found the BOV problem. It’s how I found half the stuff on this car.
So when I started looking at these latest logs and saw the PCM absolutely strangling the car at only 15 psi on wastegate spring, my brain immediately went into diagnostic mode. The calibration window looked tiny. Almost absurdly tiny. To the point where I started joking that it looked like the car was being tuned like a single turbo G25-550 4-cylinder Focus instead of a twin G25 3.5L V6.
Half joking, Then I started looking at the data. The more I looked, the less funny it became.
The logs were showing the car getting forced into the same airflow numbers over and over again. Turbo Airflow and Average Air Mass were flatlining at 43.23 lb/min while the PCM was simultaneously closing the throttle, pulling spark, triggering Engine Speed Limit intervention, and throwing overspeed codes.
The hardware kept saying one thing. The PCM kept saying another. That’s when I brought it to Ryan. My theory was simple: the calibration window looked way too small for what the hardware was physically capable of doing.
Ryan pushed back initially saying this is why he prefers live tuning and showed me the airflow tables. Fair enough. I wasn’t looking at tables though. I was looking at what the car was actually doing. I said I’m only looking at the log. I said, “I see the screenshot man, and yeah I see the load isn't capped there. I was only looking at the raw data from the log. The datalog is literally showing the turbo airflow flatlining hard at exactly 43.23 lb/min for two whole seconds right before it shuts the throttle. The data is definitely showing a hard clip somewhere, so it has to be buried in one of those turbo or speed density limits.”
After some back and forth, his response got my attention. He doesn’t think it’s a speed density limit. He thinks it may be a turbo FMEM related limiter and specifically mentioned running into situations on 4-cylinder applications where not all the tables were fully defined.
That was the first thing I’d heard that actually lined up with what the logs were showing. The interesting part isn’t whether my theory was exactly right, the interesting part is that the data pointed us toward something real. That’s the whole reason I keep digging.
I stare at logs for hours because every major problem I’ve solved on this car started exactly the same way. Something didn’t fit reality. The difference this time is the problem isn’t mechanical anymore. We’re teaching the PCM what this hardware actually is. And honestly, I think that’s the step that was skipped years ago. The original approach was basically “add boost.” What we’re doing now is making the model agree with reality first.
After watching the PCM fight 15 psi this hard, I understand now why that step matters so much. New map is loaded. The hunt continues. Ryan promised faster response times too.
This is interesting to me so here it is:
The Tale of Two Logs: Breaking Down the Torque Strategy
If you look at the data visual I posted, we are comparing the longest Wide Open Throttle (WOT) pull from Revision 5 against the newest one, Revision 6.
To understand Ford's torque-based tuning, you just need to look at two relationships:
1 Top Graphs (Airflow vs. Throttle): The blue line is the physical air the turbos are shoving into the motor (lb/min). The red dashed line is the throttle blade—basically, how much the ECU is actually allowing the engine to breathe.
2 Bottom Graphs (The Torque Brain): The green dashed line is Desired Torque (what the ECU wants). The purple line is Actual Engine Torque (what the ECU calculates the engine is actually making).
Here is exactly what the logs are telling us.
Revision 5 (The Left Side) - The Chokehold
In R5, the car was on a massive leash. The moment I went WOT, the ECU commanded the throttle blade (red line) to instantly drop from 100% down to around 40%, and it held it there the entire pull. Because the throttle was choking the engine, the airflow (blue line) slowly crept up and maxed out at only 36 lb/min. The engine wanted to eat, but the ECU refused to open its mouth.
Revision 6 (The Right Side) - Hitting the Wall
In R6, things changed. Look at the bottom right graph—the purple line (Actual Torque) climbs up and beautifully matches the green line (Desired Torque). Because the torque model is finally happy, the ECU allows the throttle blade to stay open much wider.
With the throttle actually open, the twin G25s finally get to do their job. The airflow (blue line) shoots straight up... and then violently smashes into a mathematical brick wall at exactly 43.23 lb/min. It flatlines completely. Shortly after hitting that wall, the ECU panics, hits the FMEM safety protocol, and violently slams the throttle shut to 17%.
The Logical Conclusion
Until proven otherwise, it looks like every time we remove one software restriction, the hardware immediately runs into the next one. In R5, we were throttle-limited before we could even find the ceiling. In R6, the torque strategy improved, the throttle opened up, the turbos breathed deeper, and we instantly found the next hidden limiter at 43.23 lb/min.
Whether that 43 lb/min wall is a turbo FMEM strategy, an undefined scaling issue, or a hard-coded limiter remains to be determined. But the data isn't lying: the hardware is fully capable, and we are systematically hunting down the software limiters holding it back.
My last log was actually Rev6 I mislabeled it R5 again.
The Tale of Two Logs: Breaking Down the Torque Strategy
If you look at the data visual I posted, we are comparing the longest Wide Open Throttle (WOT) pull from Revision 5 against the newest one, Revision 6.
To understand Ford's torque-based tuning, you just need to look at two relationships:
1 Top Graphs (Airflow vs. Throttle): The blue line is the physical air the turbos are shoving into the motor (lb/min). The red dashed line is the throttle blade—basically, how much the ECU is actually allowing the engine to breathe.
2 Bottom Graphs (The Torque Brain): The green dashed line is Desired Torque (what the ECU wants). The purple line is Actual Engine Torque (what the ECU calculates the engine is actually making).
Here is exactly what the logs are telling us.
Revision 5 (The Left Side) - The Chokehold
In R5, the car was on a massive leash. The moment I went WOT, the ECU commanded the throttle blade (red line) to instantly drop from 100% down to around 40%, and it held it there the entire pull. Because the throttle was choking the engine, the airflow (blue line) slowly crept up and maxed out at only 36 lb/min. The engine wanted to eat, but the ECU refused to open its mouth.
Revision 6 (The Right Side) - Hitting the Wall
In R6, things changed. Look at the bottom right graph—the purple line (Actual Torque) climbs up and beautifully matches the green line (Desired Torque). Because the torque model is finally happy, the ECU allows the throttle blade to stay open much wider.
With the throttle actually open, the twin G25s finally get to do their job. The airflow (blue line) shoots straight up... and then violently smashes into a mathematical brick wall at exactly 43.23 lb/min. It flatlines completely. Shortly after hitting that wall, the ECU panics, hits the FMEM safety protocol, and violently slams the throttle shut to 17%.
The Logical Conclusion
Until proven otherwise, it looks like every time we remove one software restriction, the hardware immediately runs into the next one. In R5, we were throttle-limited before we could even find the ceiling. In R6, the torque strategy improved, the throttle opened up, the turbos breathed deeper, and we instantly found the next hidden limiter at 43.23 lb/min.
Whether that 43 lb/min wall is a turbo FMEM strategy, an undefined scaling issue, or a hard-coded limiter remains to be determined. But the data isn't lying: the hardware is fully capable, and we are systematically hunting down the software limiters holding it back.
My last log was actually Rev6 I mislabeled it R5 again.
SM105K
Streetlight Grand Prix Champ/ IG @fafomotorsports
Haha sounds a lot like me 41 and kinda domesticated. Never had a 7 second ride, I have had some cool stuff. A nice Termi is on my bucket list of cars. I am in North Phoenix. I used to have a 13 SHO, but still retardedly opine for a 3rd Gen SHO. I blame retardation and my hate of money.
SM105K
Streetlight Grand Prix Champ/ IG @fafomotorsports
Oh look, the groupie is back. Just like h e r p e s.
Similar threads
