Drive by Wire Throttle tune = ~3 MPG gain & more
 

Background:
For those of you who are unfamiliar with Drive by wire throttles (DBW), they are basically 'electric' throttles. The computer in your machine (ECU) determines how much the throttle plate should be open based on several conditions including the accelerator pedal position. Almost all new cars have a drive by wire system now. DBW systems are often Not linear in function compared to traditional cabled throttles and DBW accelerator pedasl do Not directly control the throttle plate.

Issue:
My factory throttle mapping for my subaru was always a pain in my butt. After buying this car I started to sorely miss my cabled throttle in my old Honda. This Subaru Drive By Wire throttle (DBW) is so sensitive (presumably to attract buyers at testdrives) that I would accidentally go WOT at sometimes only 20% pedal travel. So an example: using scangauge, while accelerating, I would try to maintain 70% engine load at 2000rpm. but when transmission would shift all of a sudden my engine load (manifold pressure) would jump to 99%. This was because Subaru programmed for more throttle opening at lower pedal travels for lower engine speeds. This caused poor gas mileage in the city and I could accidentally use more throttle than I needed when hypermiling in traffic. Unwanted acceleration, a jerky drive and a fatigued foot were causing problems.

Solution:
I was desperate for a drive by wire tune for a more linear feel, at least for the first 1/4 of throttle travel. After much research and reading I learned to use an opensource Subaru tuning software to design my own Maps. I used a guide from a Subaru fan site and adjusted my DBW until I was satisfied.

Here are the drive by wire maps - before and after:
http://cdn.makeagif.com/media/11-12-2013/DgJSyR.gif
*I don't want to bore you with the technicals so skip this paragraph if you are not interested. Basically in the map above lower values 'could' mean less throttle plate opening, but note these values are checked against another table which determines how much requested torque means exactly how much throttle plate opening for every RPM. So >120 requested torque at 1500rpm could mean WOT while you will need at least 200 requested torques to open the throttle plate wide at 2500rpm. It's a 3D map.


Results:
WOW! Not only did my driving experience become silky smooth but engine Load stays very close after shifts. No more spikes in manifold pressure and unwanted accelerations. I really feel sorry for owners who are stuck with crappy factory tuned DBW throttles. Also when cruising it's very easy to feather the throttle to put the engine in its 'sweet spot'. So far for the past month I gained about ~3mpg in my city driving which is a massive improvement to my dismal 23mpg. The only downside is that the torque converter unlocks easier now since there is more pedal travel now. This has to do with TCU which I am not able to manipulate. Also engine load can fall to pretty low levels during acceleration if you're not careful which can induce pumping losses. But for experienced drivers armed with vacuum gauges, this is not an issue.

a graph to illustrate the change (red line is the new requested torque vs. stock black dots):
http://img844.imageshack.us/img844/1440/8tzo.png
please note 'load' in this context is actually "absolute load"

Now at reasonable engine speeds <3000rpm and reasonable accelerator pedal positions <30%, the throttle response is almost linear. Above 30% accelerator pedal things start to pick up quickly. This was kept so to prevent pumping losses and throttle response problems.

I am curious how I could apply this to a Toyota vehicle

Agreed!

And also... subscribed! Very cool concept. I wonder if there is a way to trick the system into thinking it needs to lock up the converter when you want it to, instead of when it thinks it should. That might net you another couple of MPG.

My toyota prius has a throttle with two potentiometers (for redundancy). It would be a fairly simple thing to splice in a microcontroller and two digital potentiometers to create any throttle position to signal map you like, probably something exponential where there is finer control at lower throttle and more rapid change towards full throttle (perhaps with limits to prevent any enrichment modes).

Weigh any such changes against the litigious nature of your locale, as well as actual dangers involved, which I will not go into here because they are largely counterproductive in the innovation space.

But in the case of the prius, it could tie into some feedback for the various efficiency operating modes, such as keeping it in EV mode or coasting mode (or opening it up more when the engine lights to reduce pumping losses). Certainly just having it coast when off the gas could be useful instead of regen.

Ah, folks are doing something similar. Plug and play but expensive, and not quite the feature set I would hope for:
Pedal Commandertoyota Prius 1 8L Hybrid Throttle Controller Faster Acceleration | eBay

Nice project.

Manufacturers are doing this more and more with "Eco" buttons that smooth/calm throttle inputs (among other things).

There was a similar mod done by Julian Edgar (Autospeed) and others for the 1st generation Honda Insight which was intended to smooth out throttle transitions in order to keep from accidentally kicking the engine out of lean burn mode. In that application, a capacitor was used to "buffer" the throttle signal: http://ecomodder.com/forum/showthrea...tor-25641.html

not that I know of...I have heard of commercial softwares for scions like ECUTEK. but i'm not sure about toyotas. I have seen some people doing DBW to cable conversions though. There is growing support for software tunings on vehicles with growing role of computers on cars though. One other thing that can also be tuned is the variable timing and lift systems.

thanks. Yes good thing about electronic systems in cars are that you do not need to have an engineering degree to mess with it.

It's been done by our very own Orange4Boy: http://ecomodder.com/forum/showthrea...tch-10028.html

And for the Subie fans: Bobs Torque converter lockup mod (PhaseII 4eat/01 4eats) - Subaru Impreza GC8 & RS Forum & Community: RS25.com

This last one grabbed me in particular because virtually all of it translates word-for-word to work on my Forester. I can't tell if the 4EAT got bolted into any 09 Imprezas and of course the wiring might be different too, so YMMV. But you can see there are some possibilities.

For my own part, I would try to install a relay that depowered the TC lockup at the 3-2 downshift, so you couldn't accidentally leave the TC locked up at stoplights, for instance. It might become a tad troublesome to keep resetting the lockup, but how much cheaper is that against getting a new lockup clutch, especially when it permits you to skip gassing up for another day?

My Forester on its best day might break 24mpg. It's a great car when the snow flies, but man. It's thirsty.

ya those mods don't seem too easy to do (for me). my engine doesn't like lugging at all. high loads at low RPM for more than a second starts causing weird engine noises and poor mileage. i find its sweet spot to be around 2000rpm-2500rpm or 3000-3500rpm at moderate loads. since the torque converter is 2300rpm stall, it doesn't usually shoot up the rpm's too high. on the highway it will not unlock unless i go WOT. it's in the city that it unlocks easier now...but havent seen any negatives from the unlocking/locking itself yet. maybe an extra 0.5mpg to squeeze?

unless you have installed this with a couple other drivers/posters and done a complete review.............I would not try for these reason.

Ebay seller out of country
reliablility
quality
install issues
not worth the risk of paying 300 bucks and then not working plus having negative impact on my vehicle..................

I have seen quite a few plug and play devices that intercept the pedal position sensors and change the response of DBW throttles. They are aimed at increasing the throttle movement though, the opposite of what is wanted (for economy driving).

One guy did a write up of one such device (in what he called a "White Paper") on it. If I recall it was in a Mercedes-Benz but they'll all operate similarly. His conclusion was that it was likely using Op-Amps to change the two (some do use three) voltage signals out of the pedal position sensors.

P-hack's suggestion of using a microcontroller seems reasonable, if the ECU mapping can't be accessed. The failure modes include dropping the throttle back to idle position if the brake pedal is applied so it won't be particularly dangerous if it did fail.

I located the throttle on the electric "car" I built in high school on the steering column to provide finer control, and reduce spikes in throttle position due to potholes.

I'm surprised a throttle hasn't been incorporated on steering wheels of conventional cars. Not only does it provide a finer level of control, but it encourages drivers to keep their hands on the wheel.

At any rate, I have a "chip" on my diesel that changes the throttle mapping, and cars such as the Prius have increasingly adopted driving modes that re-map throttle response.

P Hack's got me thinking of modifying the signals from the MAP (or T-MAPs) and Engine Temp while retaining factory ECU maps. The nett effect would be similar to remaps. I have actually reduced vacuum along the intake with a secondary intake (as per Autospeed). It affects the T-MAP and therefore you get less fluctuation in the manifold pressure.

I don't totally understand what's being said here, but from what I do understand, I think my 1995 Toyota Camry does the same dang thing. Sometimes in the city I barely put my foot down and it will just spike in RPM's and jolt me (does it in freeway traffic as well when I'm on and off the throttle alot). I wish there was a solution for it but I doubt it as my engine is only OBD-1 (does that make a difference?).

I could just look it up but for the thread and anyone else wondering, Is there a complete list of vehicles with DBW Throttles on the internet at least for so many years? And when were they first incorporated into vehicles?

*EDIT* I read this, Is it true?

"Just look at the accelerator pedal assembly. If you see a cable that moves with the pedal and goes through the firewall then it almost certainly is not DBW.

Read more: http://www.city-data.com/forum/autom...#ixzz2l3rwoHKY "

I read it is true there. On a similar subject, Could I adjust the cable in the Caravan to be looser, For fuel economy? Could it work? *Second edit* It's not a DBW Throttle.

The earliest example of DBW throttle in a car that I can recall is in the mid to late '80s BMW 7-Series, with the then new V12. That used two TB's. DBW throttle became much more widespread around 10-15 years ago, although even now I don't think they are ubiquitous.

Creating slack in the cable won't do much other than prevent full throttle opening. It is possible, in principle at least, depending on skill, to play with the throttle linkage. Unlike DBW, you can't have it change with rpm though.

If there is a cam that the cable wraps around, the shape of that determines how far the throttle opens for a given pedal movement. You might find one from another vehicle that can be adapted.

If there is a simple stud connection, that can be moved.

You might also move the pivot point for the pedal.

The cable travel at the throttle and at the pedal still have to match at wide open throttle and you don't want to go over centre.

You will also alter the pedal spring force. Actually, that might work too; use a stiffer spring.

ever green's screen capture looks very much like OpenEcu.org/Tactrix stuff. They've got software and an OBDII connector/dongle that's capable to reflash most Mitsu's and Subaru's ECU. So it's pretty limited at the moment to those marques.

A few years ago VDO/Siemens would advertise the specs for their ECUs and you could actually find out how fast they work. A better ECU or Microcontroller would then be one that has a faster scan rate. Better scan rates would enable usage of faster sensors like Bosch Motorsports Piezo MAP sensors which would run at 0.2ms instead of the Bosch's standard OEM stuff of 1ms response time. Again it doesn't mean a thing if your ECUs don't run as fast or have big enough look up tables.

Yea IF I am inclined or motivated to this I would not use the above hardware that is out of country however this might be a better alternative
Weapon R I Throttle Controller Booster Boost Selectable | eBay

wow, I can't imagine having a car with this throttle map! my car is DBW, but GM made it more of a linear feel. I can also have mine tuned any way I want, and have the shift points and lockup tuned if I had an automatic.

My 2013 Subaru is having similar issues, now that I've gotten a feel for it. The throttle is either "Off, Normal, or Too Much"

I can't have light throttle. I back my foot off the pedal and nothing changes, so I back it off more and it goes into fuel cut immediately.

I believe hybrid revolt sells something like that for the older gen insight. The 2nd gen insight does this in eco mode.It numbs the throttle and cruse control so it is delayed, but further usage it does pick up its response.

I really need to find someone to tune up my cobalt this way, it has the same exact problem, 20% throttle gives you 95%+. A light touch and the rpms rev up to 2500 in neutral, impossible to rev match correctly due to the poor control.

where are you in WI? I tap my throttle to rev match, and it works great. I also had trifecta remove the throttle lag, so fine control of the throttle is more predictable and actually seems at least more progressive.

I travel between Weston and Oshkosh most every week

one thing I've noticed in general since the move from carburetors to fuel injection is the throttle pedal is more touchy.

There are at least 4 different programs on my Focus. It's hard figuring out which one it will be using.


On some drive by wire systems (chrysler) it's horrible.

Interested thread. I don't seem to have that issue with my '13 RAV4. On a note, it does have the ECO/Sport buttons.

I also had a 2010 Prius. It was very smooth also. I even converted it to a plugin and it performed well in regards to the "surging" and pedal response. hmm, if memory serves me right the Prius also had the ECO/Sport buttons.

Perhaps manufactures, at least Toyota, are paying attention to this. IDK...

Now that the throttle has been tuned to perfection, I have moved on to other things like tuning fuel and timing. I have already increased part throttle timing without issues and now have been looking at idle timing. I was wondering what effects ignition timing has at idle and coasting on fuel economy? I have read lower timing during fuel-cutoff/coasting means less engine breaking. If so this is a good thing for fuel economy purposes? My current timing is set at 15* at idle and up to 30* when coasting at speed. For some reason the manual (5-speed) ECU calls for only 9* of timing at low RPM idle. Why the difference?

http://i.snag.gy/cmTpQ.jpg
Idle Ignition Timing Table

as for fuel I have not been able to induce lean burn. Lowering values above 0 simply just sends the car to open loop. However I have managed to replace it with leaner numbers. This is how the fuel map looks like:

http://i.snag.gy/wp8oc.jpg
closed loop fueling table. To calculate AFR subtract from 14.7. Ie. -0.1 means 14.6:1 AFR. I simply replaced all values at lower RPMs with 0 to run at 14.7.

It been long time since i messed around with timing curves .
things to watch out for If my memory is still good . high timing in low rpm/load can increase emissions . whats in your chart doesn't seem to bad at low speeds (500-1500 ).
What was the stock values .
Manual model will have no load at idle vers auto is always in gear so can handle slightly more timing , I think the curves are mainly influenced by emission output then performance .

There was talk in another thread of having the throttle open when fuel injectors are cut off. Resulting in less pumping losses and able to ride DFCO longer. Would that be something that is programmable?