Subaru Drive By Wire and How to Improve it

[throttlehappy]

A is the cruise map It wont need more lines as it goes out of closed loop by then and then uses Maximum

* Registered users of the site do not see these ads.

[camronnie]

so if A is for cruise, and I do not have a table for requested torque B or a target throttle plate position B, does that mean I can't change by DBW while just using my foot?

[ArcticWolf]

Bumping this, just did it this morning. First heel-toe in the cruise I took after the flash and I both cannot thank you enough, and am kicking myself for not having done this sooner.

[williaty]

Originally Posted by ArcticWolf

Bumping this, just did it this morning. First heel-toe in the cruise I took after the flash and I both cannot thank you enough, and am kicking myself for not having done this sooner.

Damn, took your sweet time doing it.

[ArcticWolf]

Yeaaaahhh, it did... Spent a weekend shadowing someone who was playing with MegaSquirt stuff on some other folks' cars (some seriously cool stuff, if I had a car old enough to run it...). Finally jammed through the books from Banish and Hartman after that. Also kinda hitting that part of the stride with the car... everything else's more or less sorted out, except the power side.

PM'd you, btw... side question slightly related to the DBW mapping stuff here, but a little off-scope for this thread.

[Booki]

Very interesting read - though I have a question about the values in the tables.

With the older 2005 car, the values made sense to me. 100 was max (considered 100%), but on the newer car the values went up to 325? How do you know what 100% is based on those numbers?

Would 325 be considered 100% and do the math for the rest?

I see he just changed the table to replicate the 05, using 100 as the new max.

I have noticed this numbering system in alot of my ECUs tables (MY09 Forester XT). I assume this is all relevant to turbo cars as well.

[williaty]

Somewhere in this thread is an update addressing exactly what you just brought up. It shows a way to replicate the same effect of throttle linearization while scaling the tables from 0 to the factory max value. Doing it that way significantly improves cruise control performance, suggesting that the CC system has PID logic that calls explicit requested torque values rather than percentages which really is a bad way to code it.


EDIT: Well, I can't find the update to how to handle the 06+ cars that I thought I posted to this thread. Summary is:

1) Treat my "fixed" values for the 05 map I created as percentages.
2) Multiply those percentages you looked up in my fixed 05 maps by the maximum value in your maps
3) Use those values.

[rotaxmax3]

This thread would be great.. If I had a Denso Ecu.. =(

[Booki]

Originally Posted by williaty

Somewhere in this thread is an update addressing exactly what you just brought up. It shows a way to replicate the same effect of throttle linearization while scaling the tables from 0 to the factory max value. Doing it that way significantly improves cruise control performance, suggesting that the CC system has PID logic that calls explicit requested torque values rather than percentages which really is a bad way to code it.


EDIT: Well, I can't find the update to how to handle the 06+ cars that I thought I posted to this thread. Summary is:

1) Treat my "fixed" values for the 05 map I created as percentages.
2) Multiply those percentages you looked up in my fixed 05 maps by the maximum value in your maps
3) Use those values.

I am still a little confused, here is a screenshot of my stock DBW map for my forester.

http://puu.sh/vuJKZ/16aa11cbba.jpg

If I understand correctly, when I have the throttle pressed all the way down, the throttle plate is only open "fully" between 2800rpm and 4400rpm, after that it begins to close??

Why would subaru do this, I don't understand....I mean it will make the car "feel" faster then what it really is, less throttle input causing the throttle to actually open more giving that effect.

None the less, if i was to do the math and make the changes correspond to the top numbers "accelerator pedal angle" to ensure my throttle mimics what the throttle plate is doing, I think I will achieve what I want to do. To keep it linear, as if my throttle was being controlled by a cable.

So my map should look like this? Is it safe to change the values in the white column up the top?
http://puu.sh/vuKet/43cf8fe974.png

[williaty]

Originally Posted by Booki

I am still a little confused, here is a screenshot of my stock DBW map for my forester.

If I understand correctly, when I have the throttle pressed all the way down, the throttle plate is only open "fully" between 2800rpm and 4400rpm, after that it begins to close??

Yes, you're confused. You're only looking at the input map and, from that one map, you can't determine what the throttle plate is doing. You have to look at both maps at the same time.

Why would subaru do this, I don't understand....I mean it will make the car "feel" faster then what it really is, less throttle input causing the throttle to actually open more giving that effect.

Emissions? Thermal management?


I didn't check your math on your second table but the form of the thing looks right.

[deesh]

Just did this in my 2006 Auto Impreza. Holy crap what a difference - it's actually driveable now. it also seems to have cleaned up the shifting a little bit as well, but that could be placebo

[AnthonyDre]

Anybody have the modified tables pictures for the 06 2.5i? Since they aren't showing anymore?

Thanks in advance

[bigstrusk]

In case someone was looking still:

https://web.archive.org/web/20160314....php?t=1537010