RomRaider

Looks like the newer ROMs for DBW throttle maps has an additional step (Requested Torque Base) between the pedal position and the throttle plate opening. This makes visualizing the overall throttle map (pedal-throttle) even more difficult.


To "see" the overall throttle map, I modified my spreadsheet (again ) to translate/merge these maps together to get a 3D table with Pedal Position and RPM as x/y axis, and Throttle opening as the Z values.


Pretty funky looking...fairly sharp rise from low pedal pos'n to mid, flat mid section and then bam, 100%

If I use a fixed value for Requested Torque Base (325), the resulting map is quite different.

WTF?

The "requested torque base (RPM)" table always existed with past models, but it was always 1.0. With the new logic, it obviously is not. And I've also noticed that this results in throttle mapping that seems too aggressive. There are two possibilities - the new logic also introduces an additive to target throttle plate position which could be negative or the requested torque may be modified somewhere along the way.

I'll do some logging of the variables in question on my 09 fxt and see what I can figure out.

Logged the variables in question with my 09 FXT. The throttle mapping works exactly as I outlined it -> Requested torque ratio = requested torque map value / requested torque base (rpm), which is used as the x-axis input to the target throttle plate position table. The additive I talked about is always zero in the logs and requested torque is not modified from the map value.

The map ratio involved for target throttle plate pos a/b does change - map A is used during cruise/low-throttle and map B is used otherwise with ramping between the two (which takes a little over a second). This is the typical behavior as first explained with other tables for the USDM 07 STi.

Target throttle plate pos map A is more aggressive but portions of that map will never be touched - you should use map B for analysis for a more accurate representation of non-cruise throttle mapping.

I did some logging with an '08 WRX and crunching through the numbers, the Throttle Plate value follows the general shape of the maps but the logged values are a bit higher than the calculated results.



Switching between A/B maps for target throttle plate pos'n had a minor effect at lower rpm area.

You do know that you can log requested torque and target throttle plate position? Try posting the raw logs when you do this. Along with accelerator pedal%, throttle plate opening %, rpm, you should be able to determine if there any other influence. If there is, I can send you some custom logger defs to see what is going on with the 08 WRX. My 09 FXT was spot on.

=

I logged the Accel pedal position, Requested Torque, and Throttle plate position but not the Target Throttle plate. I checked the Requested Torqued data and it followed the map values. Attached is the log file used to generate the plot above.

The throttle plate opening % will not be in sync with the target throttle plate position as it takes a period of time (although small - from previous logs, it appears to be about a 1/4 of second) to hit the target once the target is determined. I would log target throttle plate position - it will be much easier to determine if there's something else going on there.

Logged the '08 WRX some more. This time I included the Target Throttle Plate. The calculated value matches the logged "Target Throttle Plate" but not the "Throttle Plate Opening Angle (4-byte)* (%)" parameter.

Attaching some log files...

Well, I know the target throttle plate position is correct as what you are logging is the OBDII parameter "Target Throttle Opening Angle" which is the same variable as the 'Target Throttle Plate Pos (req torque)' map target.

Although I haven't messed with it much, from what I've seen, I'm guessing determining the actual throttle plate position based on the target throttle plate pos. is handled by the ETC ROM. If you look at your logs, when the throttle is more steady, the difference between the throttle plate opening angle% and target throttle pos is only around 5%. When the target is rapidly changing (i.e. lifting off the throttle suddenly or hitting the throttle quickly) it is greater, however, if you look at the previous line throttle target and the current line throttle opening, the difference is less than 5% (the time it takes to reach the target is greater when the target is rapidly changing).

I was wondering if these requested torque maps have other effects in the ecu beyond controlling throttle position and controlling boost. I find it odd how in previous DBW models generally one value was used at WOT but the CANbus ecu's seem to have different values in WOT that make boost control slightly more challenging than previous throttle based boost maps.

Basically I am wondering why have different values in WOT? IS anything else effected by the requested torque values.

Well, although an arbitrary value, it is supposed to represent torque in n-M. Which is why peak req. torque values are higher on an STi than other models. Remember, if you are talking about later CANbus models, they use a base requested torque that makes up the requested torque ratio as the x-axis of target throttle position. This varies by RPM. Previous models also have this table, however, it is always 1.0, therefore has no effect.

Sure, requested torque is used elsewhere, but I haven't messed with it in these other cases. Not used in anything immediately obvious that I have already worked on.

May I add general comments for all DBW (Dschultz if you hear me ):

All requested torque values are engine torque in Nm, pretty sure.

2D "requested torque Base" map corresponds to the engine torque when the turbo wastegate is closed.

The target boost map X axis is not "requested torque base" but "requested torque value from 3D "requested torque accelerator pedal"



From that, a correct remap should modify all DBW tables values in accordance to the new turbo spool-up capabilities, new wastegate pressure and target torque