2005 WRX tuning
 

Hello All,

I wanted to start discussing my progress in tuning my 2005 WRX wagon. My current mods are:

Engine

- APS SR30 ball bearing turbo
- Titek catted downpipe
- Borla Hush catback
- Modded injectors (740cc)
- GM boost control solenoid
- Opensource tuning using Romraider
- TinyWRX Launch control patch
- Boost Cooler MeOH Injection Kit with 63mL/min nozzle and VC-25 controller
- Innovate Motorsports LC-1 Wideband O2 sensor

Chassis and Suspension

- 22mm rear swaybar
- Perrin anti-lift kit


As far as tuning goes, I am running 20psi of boost. I run 10.9 AFR and 14 degree timing advance at peak VE (3200 RPM.) I don't run at the bleeding edge of knock since this is a daily driver and it's fast enough for me. I have yet to dyno the car, but hopefully I can take it to Road Race Engineering if Santa thinks I was good this year.

The car definitely feels faster than stock. The SR30 turbo flows about 460 cfm at 18 psi, so it is on par with the more common IHI VF22. The GM boost control solenoid works much better at controlling boost spikes and sits rock solid at target boost all day.

I am in the process of tuning for my methanol kit. From the research I have done, (and logging results from my wideband), At full load, 50% water/methanol shows up half a point richer when injected. Anyone have any tuning experience with MeOH?

I will try and post up fuel timing, and boost logs as soon as possible.

And if anyone is wondering this is the ugly silver wagon that shows up at Nick's. :)

Cool.

What does "- Open Source tuned" mean?

I was talking to him about this at the last meet. Basically the software is free and is created by the community. It actually sounded like one particular person was responsible for the version he was using. It's amazing all the stuff he could do with freeware :bigthumbsup:

Hey Sean. Basically, Subaru released the SSM (Subaru Select Monitoring) communications protocol over OBDII interface. People have figured out how use this to read and write tables in the ECU ROMs. Cobb Tuning built the AccessPort around this technique and made it into a very successful product. I think the AccessPort is a nice turnkey solution for most people.

Me, I like to fiddle with stuff...

I'm impressed. That night after the meet getting on the freeway you were keeping up really well even with a couple passengers.

I love ECUflash, along with Evoscan. Ive never worked on a Subie though. Its amazing what you can do just by modifying the stock ECU.

Interesting. I know exactly what Open Source is in the programming world; I wasn't sure it applied here. It's a pretty powerful model for some things. Motorgen.com is run on a lot of open source code.

I love open source, when it is actively developed and maintained. If I had any programming experience, I would love to contribute to Romraider.

BTW, here are some links for anyone who is interested:

Romraider

Learning View (For viewing Learned Knock Correction per Engine Load and RPM)

ECUFlash - Currently works on Subarus and OBDII DSMs

Tactrix OBDII USB Cable

Open Source is not limited to coders. Depending on projects, they need testers, help with documentation, etc.

That's true. And funding. I have donated a few times to Romraider. I figure if I can't give them my time, might as well give them a few bucks to pay for the beer for a long night of ECU reverse-engineering and programming.

Also, for anyone with MeOH tuning experience, I have a few questions I need help with:

#1 Target AFRs.
How should I tune, knowing that my WBO2 reads 0.5 richer when injecting? Should I adjust my fueling targets until my WBO2 indicates my target? My gut tells me that this is not accurate, but if I just ignore my WBO2 I run the risk of cylinder wash.

#2 If I am offsetting my targets, at what boost PSI should my MeOH controller be set to start injecting? My gut tells me to set it to start after closed loop transition and a few hundred RPM before peak VE, based on the spool characteristics. If all goes well, I should be getting a MAF signal based controller soon, and then I believe this will be a moot point.

After establishing the fueling relationship, I'm pretty sure I can get my timing advance dialed in. I just want to establish the right relationship of fuel to MeOH/Water first.

Here are my fueling targets. After becoming obsessed with MAF scaling for about a month (ask my wife, she was totally sick of me pulling over every 20 minutes of driving for a little tweak), my actual AFRs are within 0.2AFR of my targets (+/-2.0%.) :)

After doing some research, I think I have a answer to my own question:

Long Answer

A WBO2 sensor doesn't care what the fuel is, gas, methanol, propane, etc.
It measures the ratio of oxygen left in the unburnt gases exiting the engine; in essence measuring true Lambda instead of Air/Fuel Ratio. The sensor output is then calibrated for AFR based on which fuel is burnt. When replacing fuel with another fuel type, like when injecting water/MeOH, thinking in terms of AFR will not make sense because of the the different stoichiometry of the fuels.

The magical number for turbo gasoline cars is 12.5:1, which works out to around 0.85 Lambda. Now when injecting MeOH, you should still shoot for this target. Including the MeOH, the target fueling should be 15% rich.

Short Answer

Inject MeOH, try to hit target AFR of 12.5 based on WBO2, back off a little at knock.

Sources

Methanol Injection Dilemmas

Gasoline, Methanol, AFR, Lambda, and Stoichiometry.

Awesome. Thanks for the info.

Questions:

Injecting a little MeOH probably doesn't alter the energy content by volume all that much. Now, as I remember, it's a different story with E85. I think they have to target substantially richer AFR because of the lower energy content by volume. :huh:

I guess I'm trying to sus out the difference between tuning for a little MeOH injection and running E85 (85% ethanol).

Right, E85 must run richer due to its lower energy content, I think its 9:1? I ran it for awhile in my Eclipse and loved how aggressive I could get with my tuning.

You completely right. E85 has much less energy density than gasoline. I found a nice table on Wikipedia listing a comparison between Gas, E85 and 100% ethanol:



Gasoline stoichiometric AFR = 14.7 Lambda = 1
Gasoline max power rich AFR = 12.5 Lambda = 0.8503
Gasoline max power lean AFR = 13.23 Lambda = 0.900
E10 stoichiometric AFR = 14.0-14.1 Lambda = 1
E85 stoichiometric AFR = 9.765 Lambda = 1
E85 max power rich AFR = 6.975 Lambda = 0.7143
E85 max power lean AFR = 8.4687 Lambda = 0.8673
E100 stoichiometric AFR = 9.0078 Lambda = 1
E100 max power rich AFR = 6.429 Lambda = 0.714
E100 max power lean AFR = 7.8 Lambda = 0.870

Something else that goes hand in hand with this is the Stoichiometry of Regular E10 fuel used in California.

I have 740cc injectors. My car likes my injector flow scaling to be set about 5% lower than the actual flow (Flow Scaler = 703cc/min, Actual Flow = 740cc/min), which lines up pretty close to the difference between pure gas and E10. I don't know if this is the proper way of handling this, but it seem to work pretty well.

With ECUflash, you never put the exact injector size in there. You have to scale them, which is about 5-15% lower than actual size. But, different fuels can throw off your scaling just by changing your fuel maps. I hate scaling injectors, and I hated doing them again when I ran E85 in my Eclipse. But E85 is great stuff, would be a million times better if California wasnt so anal about storing it at gas stations (cant be stored under ground). But E10 shouldnt affect anything at all, most stations are E10 blend.

Hi Damian,
I am using Romraider for tuning and ECUFlash for writing to the ECU. Supposedly Romraider is better at getting the scaling right. My stock injectors were dead on with the rated flow. I have heard from several people to start scaling low and go up until your fuel trims are pretty close to linear. From your experience, is this the correct method?

Romraider has a nice little tool for scaling injectors and adjusting latency that works pretty well. They also have one for MAF scaling. I really wish they would start working on Evo tuning integration. The possibility is there...

Is there LS1 support for Romraider/ECUFlash?

I wish. Currently it's only Subarus, but the framework is there for any ECU. The coding is done in Java, all the ECU table definitions and logger definitions are all basic XML, and of course, it's all open source. All they need is someone to map out the ECU tables for the ECU you wish to edit. Unfortunately, Romraider is only good for editing the tables. You also need some piece of software to flash the file onto the ECU.

Yup, thats how we do it. I start 15% lower than actual injector size, and slowly work your way up along with adjusting your latencies. I hate doing injectors though, Im not a very patient person, and you need to be patient with injectors. I found a website that has the different latencies for different injectors and Ive found it a good starting point. I use Evoscan, Logworks, Scantech, and ECUflash for everything I do. The new Evoscan 2.6 is pretty good now. It does the same as before (datalogging) and now it does CEL reading and clearing them, and it does the same thing that Logworks does, so I dont even use Logworks anymore.


Sean, I wish I could use this on the LS2, but my current cable doesnt even work on mine for anything. The new cble would work for checking CEL codes and clearing them, but no one has even tried to crack the ECU for this purpose. I really would like someone to give it a shot.

Damian, which cable do you use? I have a Mitsu compatible Tactrix cable if needed.

I have the first Tactrix cable. I want the new 2.0 Tactrix cable, but I cant justify spending $200 on it.