Torque and Inverse Torque tuning discussion

[Matt@MMP]

Discussion of the day.. Let's talk Torque and Inverse Torque tuning for the dreaded throttle surge on late model Ford's. For discussion sake I'll keep it basic and use a GT500 for example because we are only dealing with 1 torque and inverse torque table. I've figured out the method for using actual torque from dyno, converting it to engine torque and making sure it's in the right measurement (nm, ftlbs, etc) then plotting this data, interpolating the unknown's and then building the inverse table and repeat.


Let's say you didn't have a dyno and were remote tuning this vehicle. Would you use wheel torque error and just fudge your way through it? I know you can go in and cut throttle area based on error as well (have done this also) but this will usually result in a lazy pedal feel.


I'm just curious on some other methods being used out there and had some good discussion going behind the scenes with a few others regarding this topic. I figure this could make for an interesting thread.

[Higgs Boson]

If the car is stock (or a known good combination) and the throttle body is changed, use throttle angle error to correct pedal pulse/jerk/surge.

If the car is modified from a known good TB combo then you might use torque tables.

I added a GT350 intake and TB and screwed around with all kinds of tables with no good result (literally hundreds of flashes in experimentation) and finally just used TA Error to correct the Predicted table and it's smooth as butter now.


If you really want to play with the TQ tables then you can log Engine TQ Reference to see where it is.....I don't think anyone (around here at least) has a foolproof method for targeting each table with any accuracy although I know some of us are working at it.

[AKDMB]

Let's say you didn't have a dyno and were remote tuning this vehicle. Would you use wheel torque error and just fudge your way through it? I know you can go in and cut throttle area based on error as well (have done this also) but this will usually result in a lazy pedal feel.


I'm just curious on some other methods being used out there and had some good discussion going behind the scenes with a few others regarding this topic. I figure this could make for an interesting thread.

I always wondered how this would work as well. So you need to tune the torque tables in the instance of

1. Different throttlebody (apparently you can use throttle angle error to tune the predicted angle table)

2. Significantly changing how much power the car makes (supercharger, turbo, etc.)

Any other reason you have to tune the torque tables?

[Higgs Boson]

The load row goes to 1.1 and 515 or whatever ft lbs. so it has high enough limits to accommodate anything NA. I guess it would be possible to increase power at part throttle mid RPM and screw it up but I doubt it.

Also, the ECM uses more than just those tables, it also uses Driver Demand, which is where the process typically starts so adjustments for part throttle can also be made there.

Of course, the end result of the tables is at the throttle body, so theoretically, leaving the torque model stock and applying angle error to the TB table is an easy way to get the same result. I did not have good luck having an effect with the TQ/Inv tables.

[murfie]

Yup we need to figure out snap to point and snap to line and their relationship to distance arrays. Then we can start targeting specific tables. I've been going over many different tunes and have not come across a set pattern. All I can see is that snap to line describes some kind of throttle change path and how the different arrays interact with each other. The 2015 f150 eco boost shows the stability array and emission array right in the middle of the fuel economy array. The 2011-14 are pretty straight forward and use a linear fuel economy array and the other arrays are just built into it. The 2015 gt is really strange but you see the beginning of the fuel economy array and the end of it. Where the middle is coming from I haven't figured out. My best guess is the OP target at the bottom as 14-22 and the emissions and stability which are both 0-0 in the upper part of snap to line. Simplified it looks like it goes fuel economy array> stability/emissions> fuel economy>OP> fuel economy as the path through the arrays.


This is just an educated guess as the descriptions are just full of ????


Scratch the middle fuel economy array and substitute IMRC arrays and 6-7-7-8-8-9 makes sense. As 6/7 are closed 8/9 are open.

[Eric@HPTuners]

This is just an educated guess as the descriptions are just full of ????

Sorry for that. I'll see if I can get that updated.

[murfie]

Sorry for that. I'll see if I can get that updated.

Nice! Thank you so much. It was going to be the first thing I was going to bug support about after I got my own cable.

[jchambers]

Matt,

Post up a screenshot and log of some surge. It'll give us something to talk around with a good visual. I think I've got a few from an EB recently that I can try to post up as well.

[Higgs Boson]

Jeff, what did you log to "see" the surging? I didn't see it in torque values or spark values, the only place I could see it was throttle angle error over about 1.5 degrees. Under 1 degree of error was not noticeable through the car, but maybe around 75% throttle and middle RPMs was the worst, around 3 or 4 degrees of error.....

[murfie]

Reading about torque based engine management I've come across a ford patent that says manual transmission and automatic transmission ecu use a different torque reference value. Manuals use the torque the engine produces directly. Auto transmissions use torque based on output shaft torque. If this is true the gear ratio of the selected gear in an automatic would need to be factored into the torque request and give much higher values than the manual engine torque. Thoughts on this?

From memory I think the driver demand table in my car was output shaft vs rpm.

One advantage to a torque based control is pedal feel in all conditions remains constant.

Patent I'm referring to:

http://www.google.ch/patents/US6953023

[Higgs Boson]

Driver demand is pedal vs rpm, now rpm might be crank or output shaft but I think it's crank. Will always be pedal since it is driver demand....

The manuals have dd tables per gear, the auto has one.

[murfie]

Driver demand is pedal vs rpm, now rpm might be crank or output shaft but I think it's crank. Will always be pedal since it is driver demand....

The manuals have dd tables per gear, the auto has one.

It was the pedal map ratio table I was thinking of. OSS rpm vs pedal%.

[Higgs Boson]

It was the pedal map ratio table I was thinking of. OSS rpm vs pedal%.

Yes, that is auto only, not on the manuals.

[murfie]

Yes, that is auto only, not on the manuals.

So Im guessing low range table is torque when the TCC is unlocked that's why the values are much lower than engine torque. The engine RPMs follow the TCC map slip values pretty close if not exactly. Then the ratio map gets applied to both tables.

[Higgs Boson]

Maybe. I figured it was a latent table for trucks, sort of like all the TCC tables for different truck categories like Mud and 4x4.

[etm351]

I have played with torque table and torque inverse tables on the dyno with a stock can and cars with headers and air kit and i haven't found any difference in power. Now it seems power on the street from 0-60 or so seems better but hard to say for sure, anyone else seen power gains?

[kaineoutlaw]

Is this in the 11-14s autos?

[TunedByNishan]

The only thing I have never seemed to understand is why one would need to change the torque values in the table for a given load look up?

Air Load is proportional to Torque Since it represents some factor of cylinder filling (cylinder pressure) and the OEM has already modeled this up to 1.20 load.

I can't see why adding a power adder, let's say supercharger for example, would alter how much output torque the same amount of air load provides (up to 1.20).

Yes, the engine will flow more air now, but all that will do is make more air load and shouldn't change how much torque the engine produces at lower airflow rates. The engine will likely just produce a higher air load at the same pedal position due to the additional of a supercharger.

My approach has been to take the 1.20 Load Row, copy it, and paste it one row up and then use the last row for extending up to 1.5 or 1.6 load (depending on what I need/how much power the car is making) and in some of the roush files and other known good tunes seem to do exactly the same.

I've also heard that some strategies don't produce any surging whatsoever with the same calibration file.

[Higgs Boson]

if there was only one Mapped Point (and therefore no Mapped Points) then that would possibly be a solution however, with the spark and cam events bouncing around 27 different MP's the torque numbers are not always spot on at part throttle.

The main issue with the surging is when the IMRC's open and the load is split between 14 and 21. If you correct MP 14 and 21 85% of the surging will be gone. The rest is fine tuning.

[thegraystang]

if there was only one Mapped Point (and therefore no Mapped Points) then that would possibly be a solution however, with the spark and cam events bouncing around 27 different MP's the torque numbers are not always spot on at part throttle.

The main issue with the surging is when the IMRC's open and the load is split between 14 and 21. If you correct MP 14 and 21 85% of the surging will be gone. The rest is fine tuning.

this might work for the 15+ but the 11-14 dont have IMRCS or 28 points.