Ford's Engine Brake Torque Calculation

[CCS86]

Does anyone know how Ford is calculating this value? Apparently GM uses airflow for the calculation, but it doesn't appear that Ford does it this way.

I created a math parameter to take MAF rate (I am using a stock GT500 curve), in lb/min, and multiplies by 10 (super rough HP approximation). Then multiplies by 5252 and divides by RPM to get "torque". The absolute of this isn't accurate, but since it is airflow based, it should help to check for correlation. I plotted this against the Engine Brake Torque channel. In some areas, they have similar values, but they don't really trend together.

It is possible that it is simply looked up from the torque tables. You can grab a snapshot of RPM, absolute load, mapped point, and look up a value in the correct torque table. The reported Engine Brake Torque is significantly lower than the table value. Maybe it is factoring in losses? During a WOT run, the losses between the TQ table and reported Brake Torque were 15% @ 4000 rpm, 20% @ 5000 rpm, and 28% @ 6000 rpm. This seems to disprove it simply being looked up.

I took a WOT run, used the absolute load for each RPM to interpolate the TQ table values, and plotted them against reported Engine Brake Torque. I used the maximum reported values for each RPM in a longer log, because I did find that it seems to use spark timing in its calculation. On a WOT run that I started from low RPM, and picked up some knock retardation that carried through the run, the reported TQ values were significantly lower then in other matching conditions where timing was being added.

TQ-Channel-vs-table.png


This is a Coyote with an M122 blower at 10-11psi. It should be in the neighborhood of 460 wheel TQ, so all these values seem low if they are meant to be engine TQ.

[txtailtorcher]

Theres a very good thread on it. https://forum.hptuners.com/showthrea...rol-ETC-System

Your answer is in here.

[CCS86]

Theres a very good thread on it. https://forum.hptuners.com/showthrea...rol-ETC-System

Your answer is in here.

I don't see an answer in there. I posted a number of questions in that thread which were never answered.

[txtailtorcher]

I don't see an answer in there. I posted a number of questions in that thread which were never answered.

engine brk tq is calculated via lambda, spark and airflow; which is derived from the maf and statistical SD model as well as map and all the other inputs. Command is received from the throttle and filtered through DD and the other models. Then frictional loss multipliers are applied. That's the basic overview. It's not based off of one single calculation. That's how ford also accounts for altitude changes in factory ecoboost and other N/A cars.

Edit: Tq tables are indicated tq (look up the definition of this to understand what im saying) Tq tables are used as part of the ETC system similar to DD, its just another of the many filtered request's that can be used when accelerating. In the post I referenced bill said that if we used the first way to command tq then we would always end up with less than commanded. That's why some numbers are lower than others and thus why you need to add some to whatever the tq table is. You can play with this by adding and subtracting from the TQ tables. When you get to an unrealistic high demand the car will oscillate, opening the TB too much and itll buck. too low and itll just die. experiment with this and make your tq max and min unrealistically high and low so you can watch IPC error and how it reacts vs the changes you made in the tables. Also Disable CLIP/ADD or just switch to intervention protection and log it so you can see what its adding or subtracting.

[CCS86]

Thanks, that sounds close to what I was imagining.

I really wish we had access to frictional loss tables/multipliers.

[veeefour]

engine brk tq is calculated via lambda, spark and airflow; which is derived from the maf and statistical SD model as well as map and all the other inputs. Command is received from the throttle and filtered through DD and the other models. Then frictional loss multipliers are applied. That's the basic overview. It's not based off of one single calculation. That's how ford also accounts for altitude changes in factory ecoboost and other N/A cars.

Edit: Tq tables are indicated tq (look up the definition of this to understand what im saying) Tq tables are used as part of the ETC system similar to DD, its just another of the many filtered request's that can be used when accelerating. In the post I referenced bill said that if we used the first way to command tq then we would always end up with less than commanded. That's why some numbers are lower than others and thus why you need to add some to whatever the tq table is. You can play with this by adding and subtracting from the TQ tables. When you get to an unrealistic high demand the car will oscillate, opening the TB too much and itll buck. too low and itll just die. experiment with this and make your tq max and min unrealistically high and low so you can watch IPC error and how it reacts vs the changes you made in the tables. Also Disable CLIP/ADD or just switch to intervention protection and log it so you can see what its adding or subtracting.

Torque is inferred form Load, Load is inferred from MAP and MAP is inferred form SD model - MAF is just a correction to those models.
SD model must match your Torque model - you can do slight changes to one but if you are going form N/A to F/I (big changes) you have to change both accordingly.

[txtailtorcher]

Torque is inferred form Load, Load is inferred from MAP and MAP is inferred form SD model - MAF is just a correction to those models.
SD model must match your Torque model - you can do slight changes to one but if you are going form N/A to F/I (big changes) you have to change both accordingly.

yes but on the flip side what some have done is disable SD and run off the MAF. I always use both.

[veeefour]

yes but on the flip side what some have done is disable SD and run off the MAF. I always use both.

You can't disable SD on those vehicles - you can disable some corrections related to SD.

[blackbolt22]

My original MAF calibration was causing STFT to pull fuel (-2 to -4%) . If I've calibrated the MAF would it cause Engine Brake Torque to show a lot lower? I'm trying to figure out why I'm showing around 40 less ft. lbs of torque @ peak during a WOT pull. Of course the conditions of the pull in comparison are a lot warmer so that could be part of it too..

I also richened WOT lambda..... Leaving everything else alone except lambda and doing WOT pulls under near identical conditions, could I use Engine Brake Torque to find out the BEST lambda values for the most torque?
Is .820 lambda too much for 5500+rpm - bone stock except mufflers.

[CCS86]

My original MAF calibration was causing STFT to pull fuel (-2 to -4%) . If I've calibrated the MAF would it cause Engine Brake Torque to show a lot lower? I'm trying to figure out why I'm showing around 40 less ft. lbs of torque @ peak during a WOT pull. Of course the conditions of the pull in comparison are a lot warmer so that could be part of it too..

I also richened WOT lambda..... Leaving everything else alone except lambda and doing WOT pulls under near identical conditions, could I use Engine Brake Torque to find out the BEST lambda values for the most torque?
Is .820 lambda too much for 5500+rpm - bone stock except mufflers.

How do you know that the MAF curve was wrong?

What percentage difference in reported torque did you see?

Unless you were at identical temp, baro pressure, spark timing, etc; you can't expect the exact same reported torque.

[blackbolt22]

How do you know that the MAF curve was wrong?

What percentage difference in reported torque did you see?

Unless you were at identical temp, bar pressure, spark timing, etc; you can expect the exact same reported torque.

40 ft. lbs difference around the same RPM (4400ish). I used STFT + LTFT to adjust the WOT MAF down a little since it was showing rich%. Adjusted by special multiply by half %. I could see that lowering the MAF lb/min.. this makes sense as well as the hotter air + engine but 40 ft. lbs?!! The biggest difference is .820 vs .830 lambda between runs.. It seems @ .830 though more timing could be had if I corrected this or upped the advance cap. Not sure why .830 would want more timing then .820 though!

3rd gear pull WOT, June 30th Peak = 436 ft. lbs Torque. (360HP est) , 4350 RPM, 77F Amb. / 192F ECT, ETC Tq Req = 358 ft lbs. , Timing 20 degrees., Lambda .820 , baro 14.1psi, map 15.7psi. , Knock Retard -1 , , MAF 30.52 lb/min, Absolute load 102.4%

3rd gear pull WOT, Sept 16th Peak = 396 ft. lbs Torque. (333HP est) , 4400 RPM, 88F Amb. / 205F ECT, ETC Tq Req = 358 ft lbs. , Timing 21 degrees., Lambda .830 , baro 13.9psi, map 15.4 psi. , Knock Retard -4 (capped) , MAF 29.25 lb/min, Absolute load 99.6%

[CCS86]

40 ft. lbs difference around the same RPM (4400ish). I used STFT + LTFT to adjust the WOT MAF down a little since it was showing rich%. Adjusted by special multiply by half %. I could see that lowering the MAF lb/min.. this makes sense as well as the hotter air + engine but 40 ft. lbs?!! The biggest difference is .820 vs .830 lambda between runs.. It seems @ .830 though more timing could be had if I corrected this or upped the advance cap. Not sure why .830 would want more timing then .820 though!

3rd gear pull WOT, June 30th Peak = 436 ft. lbs Torque. (360HP est) , 4350 RPM, 77F Amb. / 192F ECT, ETC Tq Req = 358 ft lbs. , Timing 20 degrees., Lambda .820 , baro 14.1psi, map 15.7psi. , Knock Retard -1 , , MAF 30.52 lb/min, Absolute load 102.4%

3rd gear pull WOT, Sept 16th Peak = 396 ft. lbs Torque. (333HP est) , 4400 RPM, 88F Amb. / 205F ECT, ETC Tq Req = 358 ft lbs. , Timing 21 degrees., Lambda .830 , baro 13.9psi, map 15.4 psi. , Knock Retard -4 (capped) , MAF 29.25 lb/min, Absolute load 99.6%

That's only a 10% difference in reported torque.

Ambient temp was 11* hotter (less power), ECT was 13* hotter, baro was lower, and the MAF shows 4% less air entering the engine. Nothing sounds out of whack to me.

[blackbolt22]

That's only a 10% difference in reported torque.

Ambient temp was 11* hotter (less power), ECT was 13* hotter, baro was lower, and the MAF shows 4% less air entering the engine. Nothing sounds out of whack to me.

OK so nothing to worry about, thanks! Still I don't believe I have any log showing that much torque after I adjusted the MAF closer. 436 ft. lbs of torque seems very high for a stock car...

[MRRPMBRP]

This is a calculated value, it's not being read at all.

[blackbolt22]

This is a calculated value, it's not being read at all.

Read my earlier posts. I know it's calculated that's why I was asking if adjusting MAF might be the cause for it being lower. I thought that perhaps dialing in the MAF would bring it closer to reality... you know like make the calculation more accurate.

[veeefour]

My original MAF calibration was causing STFT to pull fuel (-2 to -4%) . If I've calibrated the MAF would it cause Engine Brake Torque to show a lot lower? I'm trying to figure out why I'm showing around 40 less ft. lbs of torque @ peak during a WOT pull. Of course the conditions of the pull in comparison are a lot warmer so that could be part of it too..

I also richened WOT lambda..... Leaving everything else alone except lambda and doing WOT pulls under near identical conditions, could I use Engine Brake Torque to find out the BEST lambda values for the most torque?
Is .820 lambda too much for 5500+rpm - bone stock except mufflers.

ECM 6497 - Torque Lambda efficiency