Subaru Impreza GC8 & RS Forum & Community banner

1 - 20 of 128 Posts

·
Registered
2000 2.5RS BRP
Joined
·
1,180 Posts
Discussion Starter #1 (Edited)
40-80 5.0s. 93 octane. No boost. No exotic fuels.


Youtube channel: ImprezaRSC
- YouTube
Testing vids, how-to vids, autocrossing, drags, etc.

Instagram: Tinkerfreak (@tinkerfreak) • Instagram photos and videos Road tripping shenanigans

Blog: Tinkerfreak's Tech & Adventure Blog | Shenanigans of a Rabid Inventor

Overview

Welcome to my budget rebuild of the original E85 powered 2.5RS. It has been my trusty steed for 17 years and was due for a rebuild. After 100,000 miles on E85--and picking on turbo cars ;)--I wanted to top that on a reasonable build budget. What fun is just another turbo swapped Subaru? Box 6s? Been done. I love developing new stuff as much as turning old stuff into new stuff. I also love autocross. There's room for both to make a fun and reliable daily that doubles as a weekend track car. What every hobbyist loves are the stories we hear and make for ourselves.

Goal

I am using a classic Impreza and making it look even older, like a "what-if" mashup. What if Subaru had entered the GC chassis into SCCA's heyday to duke it out with the Datsun 510 and BMW 2002 and scrap with the Camaros and Mustangs? It would be styled a little older. You'd see a lot of hand fabricated parts and some battle scars. It wouldn't be turbocharged, likely on a dual carb-like arrangement. It would be stripped for weight. You'd see labor that adds up to less being more. I also love the Harley-like N/A voice that this has. It grunts and snorts like a V-twin bike with pipes. You see and hear that nasty idle. It sounds like it should be a more raw kind of vehicle and ****ing own it at the same time.

Parameters

Fast. Reliable. Cheap. Pick Two...

With any good project it is a good idea to define the parameters and the scope. Reliability was the primary goal. Speed and power were less of a priority. $4k was the financial ballpark to do the entire thing, a relative shoestring budget with some wiggle room. The end result is part junk yard dog, part sleeper, part freakish experiment, part OEM rock, and part troll car :monkey:. This was an incredible amount of labor, too. That was the trade-off of the lower cost; doing or knowing who was doing the labor. Because race car? No. Because I can.

Engine

Some of you are probably asking "so why no boost"? It was a matter of reliability and cost containment, again. Pesky issues, but manageable. This build was stuck on a 4:11 5spd. We know this gearbox doesn't hold up for very long under boost. But, is fun to drive being shorter geared. That meant following other routes to get the speed, like weight reduction. Weight reduction also improves reliability, the other more important goal.

How can we coax more power out of this n/a engine arrangement? Well...there's using the same bolt-ons as everyone else for the last 20 or so years. What fun is that? The same problems. The same limits. The same character. Meh. I'd rather start over with fresh paper and color in all of the details. That can be a badge of honor as much as a cross to bear. I dare to help reinvent the genre, on a small budget. Challenge accepted! Let's start by letting more air and fuel in with multiple throttlebodies. Compress it a little more, and then let it out more easily. Along the way, we'll rev it a lot higher and give it enough air and fuel while it's there. The trade-off is using a slightly smaller stock engine to get it all done. An upshot is 250lbs less of turbo equipment. That's also weight to lose by never installing it to begin with.

Despite the cost constraints, we're aiming for as much power as the 5-8psi intercooled turbo kits from back in the day. 260-280hp from a smaller 2.0L is actually a pretty reasonable goal while n/a, but not with the same off the shelf parts. We've known for a while that these n/a engines can make as much power as stock turbo cars or the bolt-on turbo n/a cars. This build is to push that envelope a little, while still on stock parts. The trade-off is less torque for being able to spin the smaller motor higher. It definitely drives differently, but not unpleasant underfoot. This build was very much a game of compromise.

Exterior

The rest of the car is still getting some TLC and fab along the way. The bullet mirrors were an homage to 60s SCCA cars. In respect to that, I'm "adding lightness". The interior panels are getting stripped and replaced with aluminum skin where needed. The carpet is gone. The sound deadening still needs to come up. The floor will be smoothed and painted over the summer. The sunroof will be treated to a targa-style mod for manual use and removal, a bit more like a Jeep top. I love the roadster feel of this car and the rest of the build is meant to enhance that.

Suspension & Brakes

The modded suspension is unchanged, with Prodrive P1 springs on gr2 struts, 20mm STI rear bar, and your typical anti-lift kit, upgraded mounts and bushings. With the interior stripped, it sits at stock height. The power isn't there to need huge brakes and big tires. It might need some stronger braking after getting the power dialed in, but it should not need Brembos or anything too pricey.
 

Attachments

·
Registered
2000 2.5RS BRP
Joined
·
1,180 Posts
Discussion Starter #2 (Edited)
Engine Swap

In our little world with the 2.5RS, we know the character of the car. Lots of flat torque that does not like to rev very high, like a v-twin bike. That went out the window with the old EJ25D.

Removed:

2.5L
SOHC
10:1 compression
coil pack
non-avcs


Installed:

2.0L
DOHC
11:1 compression
coil-on-plug
avcs (non-functional)

Fuji Factory twin throttle intake and fuel system
Deatschwerks 440cc RS injectors
E85 fuel
ACT Flywheel
PDM TSK kit
Exedy stage 1 clutch
Infernofab 4-1 header
3in catless midpipe

The only thing I kept was:

3in resonated Invidia WRX turboback,
Aeromotive 340 pump
Magnafuel filter.

This is currently running on the stock ECU. It would prefer the 280cc injectors until I get up higher into the revband. That's why 440s and a tuning solution were picked instead. The tuning is going to be highlighted separately. There will be no heroics needed there. Just scaling and mapping changes.

The engine is a smaller 2.0L engine capable of 1,000 more rpm from the factory. I traded low end grunt for top end horsepower. I'm going to be a little dodgy on exact engine specs for now. Just know that up front. I built a one of a kind swap and want to keep it that way for a little while. There is not another one of these, and it is currently a little incomplete without tuning to fully unleash it. I've done what I can mechanically and electrically. That includes engine + driveline removal and installation, fitment, wiring, sensors, parts, debugging, and vehicle assemblies. Tuning is the end of it and then it is what it is.


Driveline

The driveline was the other major limiting factor in the build. The 4.11 5spd is a little dogged in the US market. The gearbox is stock, but with the following repairs:

Rebuilt center diff
Shifter bushings
Output shaft bearing
BG Ultraguard gear oil.

It still has STI Group N engine and trans mounts. Whiteline steering bushings are up front. An STI shifter is topped off by a billet pistol grip wrapped in black leather.
 

·
Registered
2000 2.5RS BRP
Joined
·
1,180 Posts
Discussion Starter #3 (Edited)
Project Lambda Tuning Software

Previous thread: http://www.rs25.com/forums/f5/t221027-n-tuning-again.html for enquiring minds wanting to know.

Overview

ProjectLambda is run by Mike, user 92SS. This is user tuneable software for our ECUs. I do not sell them. I paid and got mine in a few days.

$120 for the Bluetooth device
$280 for 1 vehicle license.

Specs

I do apologize for the lack of screenshots. I was asked to not show anyone, but a review is at least due.

The software gives you just enough controls to be very useful. They are laid out in a simple fashion from the main menu. Auxiliary controls can be turned on/off. DTCs can be deactivated. Available fuel and timing maps can be adjusted.

This is where the weird reading comes into play. You see the X+Y axes, but the numbers you see in the cells make almost no sense. You are adjusting a derived figure from other tables that you don't have access to. It looks like tuning, but feels a little more like tuning a computer instead of a car. You have to use an Excel spreadsheet for now to nail the fuel mapping. Timing is straight forward. Steve is working on improving the usability of it and making the process more straightforward.

Tables available are things like base spark, learned spark, base fuel, power fuel, and o2 sensor scaling. There are a couple of global variables to get right and then you can move on to tuning the cells. The main tables are 16x16. They can be rescaled and it will update all the other tables.

Review

This has been a lot of fun to geek out on and to see it work so seamlessly. I'd suggest taking it to a tuner if you have never done this before. If you have tuned a car before, this is something you can use as long as you accept that it reads just a little differently. I'm glad I took the chance because I could lasso the product and the problems on the car both. This build was a good test of its capabilities as well as the stock ECU. All of the engine fabrication that was done demanded tuning and this software was up to the task. The end result feels as repeatable and consistent as factory.

The product support has been fantastic and the beta software has been quite usable. You do have to be pretty good with computers. You are working on a computer. It does involve a spreadsheet and some basic algebra to nail the fueling. The lingo is a little different, but is easy with practice. I found it to be a very software intensive exercise. Working on this feels more like working on a computer than tuning a car, and I'm just saying that to be fair.

The bluetooth is the biggest reason to get it. No cables are in the way! It can write and log both, but not simultaneously. You do have to be mindful of battery voltage to flash successfully. Putting a charger on the battery is good safety practice.

ECU code AF582 For the record, these cars have very good factory ECUs. Even though we have been unable to reflash as an end user, this ECU lets you get away with murder. E85, turbo, or high compression n/a, you have to admit that they are at least capable. Now, we end users can tune it.

I ended up with a self-made tune that is appropriately rich, with plenty of timing advance in the upper revs. The torque feels upwardly linear like a Honda instead of necessarily table top flat. It winds out instead of grunting all the torque at once, likely due to non-use of AVCS. The lower gears are stupid quick. You have to be committed to using them, looking a lot farther ahead before mashing the gas. I have not raised the redline, yet......
 

·
Registered
2000 2.5RS BRP
Joined
·
1,180 Posts
Discussion Starter #11 (Edited)
It's a combination of things that make it wrong--too big an injector, stock mapping, 2.0 displacement, E85, etc. I could go back to RS' 280cc injectors and it would run fine.

What we're about to see next week is some new tuning software in action to fix it. I like what I see and it looks like a good solution at an affordable price. The car will need 440s for the 7k redline on E85.
 

·
Registered
2000 2.5RS BRP
Joined
·
1,180 Posts
Discussion Starter #16 (Edited)
Your car will use the most fuel at peak torque. If you put 370s in it you should be fine on e85 and it will run like stock
Quite possible. I do agree with that. I always thought 440s were a little big on these cars for n/a, but not by much.

However, I already had the 440s. The new tuning software and dongle will be here in a few days. The system is robust and still simple. I like what I see. There's no extra wiring and even a live tuning option. There's no wastegate mapping. No MAF to rescale. It is as bare bones as it has to be and still focused on the important stuff.

These injectors aren't grossly oversized, so dialing them in should be pretty straight forward. We are confident within our visions that this will work out, on E85 and gas both. I'll post updates and videos in a few days. :bot:
 

·
Experimenter
Tubaru Pickup
Joined
·
3,910 Posts
Scaling injectors is pretty easy. Latencys can be tricky. Biggest thing would be adjusting cranking and tip in stuff thats in milliseconds of pulse width and not effected by scaling. But with 440s it should still be ok and the extra fuel would help e85 cold starts anyway
 

·
Registered
2000 2.5RS BRP
Joined
·
1,180 Posts
Discussion Starter #19

Quick update. I've been puttering with the ECU learning the ProjectLambda software. So far, so good. I'm learning its basics and whittling away at the mapping. It's a little more like speaking "computer-eze" than speaking "tuner". We've achieved a good idle and cruise with decent power. There is more to do with ironing out the AFR and cleaning up the corners of the maps.

It's a helpful tip to have a charger attached to your batter when reflashing. You must have a minimum voltage present for the reflash process to take effect. A charger attached to the battery ensures a solid state of voltage. This ended up needing a warrantied alternator to begin with. It has been reflashing fine ever since.
 
1 - 20 of 128 Posts
Top