2005 Legacy GT Tuning Guide

[covertrussian]

2005-2006 Legacy GT Tuning Guide

 

Figured I would combine all of the tuning posts that I've made in various threads for an easy to search one stop location. I will edit this post with links and new additions as I test more items.

 

Safety/Reliability

Fueling:

Closed Loop to Open Loop Delay Disable - Disabling the delay that causes leanness during boost, hesitation, and burnt valves.

Disabling AF Learning for Open Loop - Disabling Closed loop learned corrections to open loop fueling

Max Airflow Learning Restricting - Lowering the max airflow to which closed loop tries to learn fuel adjustments at (which later get applied to open loop)

Fixing the Tau Tables - Related to Rich dip on high loads

 

Timing:

Timing Compensation Zeroing - Removing the *2 advance for two cylinders, which could cause detonation in those two cylinders first.

 

Power

Turbo: VF46

 

Turbo: Big16G

 

Fuel Economy

Turbo: VF46

 

Turbo: Big16G

[covertrussian]

Closed Loop to Open Loop Delay Disable

 

This is the very first tune change that you should do on any turbo Subaru. If you don't disable or greatly reduce this delay you run the risk of killing your motor with detonation, burning the valves, and just having poor driveability (very hesitant in boost).

 

Closed Loop to Open Loop Delay is found on USDM's mainly (it's believed that it's there for emissions reasons), JDM tunes that I've looked at had it disabled. What it does is keep the ECU in closed loop while you are flooring the car for set amount of time, aka lean while in boost! This leads to burnt valves on bone stock cars. Going to any stage 1 tune should disable CL to OL Delay (by zeroing out the values).

 

Lets take a look at the VD graph, look at AFR's.

http://legacygt.com/forums/attachment.php?attachmentid=258216&stc=1&d=1510676384

 

As you can see with the 100% stock tune, you are staying at high 13 AFR's while pushing 12ish PSI. This is also happening while you are at peak torque, where the engine is the most susceptible to knock.

 

 

http://legacygt.com/forums/attachment.php?attachmentid=258217&stc=1&d=1510676384

 

By disabling Closed Loop to Open Loop Delay (by setting it to all 0's), the AFR's just about instantly dip to 11's, now stock O2 sensor is very inaccurate below 12:1 AFR, thus take it with a grain of salt. But as you can see the engine welcomes the richer AFR's by gaining you 20ft-lbs of torque.

 

Do notice the peak horsepower is "down", which I would blame a slight road inaccuracy for, thus nothing to sweat about.

 

Moral of the story: ALL TURBO SUBARU's should have closed loop to open loop delay disabled if they want to save their valves and gain some power while they are at it!

[covertrussian]

Disabling AF Learning for Open Loop

 

I feel like this is the second most important item to disable in your tune. For what ever reason Subaru thought that it was a good idea to take closed loop AF Learning and to apply it to Open Loop fueling. This means the open loop section could be much richer or worse, much leaner (what I experienced on my car) then what the map calls for.

 

 

I recently discovered why my WOT AFR's have been every inconsistent between virtual dyno pulls. The issue is with AF learning, where column D affects open loop fueling, which is not good for a turbo car, or any car really.

 

Stock AF Learning values:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/AF%20Learning%20Stock.png~original

 

Stock AF Learning after a highway trip:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/D%20Learning%20Stock.jpg~original

 

D Learning can be disabled by setting the C to D range to be extraordinary high. Guys on Romraider found that 80g/s should be sufficient.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/D%20Learning%2080.png~original

 

AF Learning with D disabled after a highway trip:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/D%20Learning%20Disabled.jpg~original

 

Here are two side by side runs, you can see how much it leans the AFR's at full throttle, on stock fuel map this might not be bad, but on a fine tuned fuel map this could cost you an engine.

(Don't mind the low power numbers, the road I did the pulls on were going up hill)

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/LGT%20D%20Learning%20Enabled%20vs%20Disabled.jpg~original

 

 

I've been tuning D learning for couple months, every time it has resulted an MPG loss. If only I could have another column, then could retain stock learning and still have D disabled, but we can't have our cake and eat it. I just got back from a NoVA trip and got 30mpg going there and 29.8mpg coming back (both at the pump not gauge which was way off), which is some of the better MPG's that I've gotten going there and back. This gives me hope that this AF learning values might be good long term .

[covertrussian]

Max Airflow Learning Restricting

 

If you want to take a step further from disabling D learning, you can lower the airflow point at which the ECU will stop learning and applying the learned values to.

 

As Mr. Flinkly found, there seems to be a table that can restrict the max mass airflow (g/s) that closed loop will be calculated for.

 

In ECU Flash there is an experimental table named: AF#1 Learning Max Threshold, description says "This determines the maximum mass airflow that AF Learning values will be applied to. By setting this lower than D range, no fuel learnings will be carried into open loop fueling." By default on 05 LGT it's set to 80.00. Which explains why we always had to go to at least 80 on D learning to prevent any learning.

 

I reduced the AF#1 Learning Max Threshold to 0.39g/s and went back to stock AF learning scales, after 10 miles, nothing was learned in D column, which means that this table does indeed stop the learning at the set g/s!

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/LV_Stock%20AF%20Learning_Maxlearn-40gs.jpg~original

 

 

While you could just leave it at that and call it a day, but you would be cutting the closed loop learning range by about half. Here's a highway log with 70mph to 75mph driving, graphed it out using awesome little little tool by RomRaider user vgi, MAFScaling.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/70-75mph_MassAirflow.png~original

 

Gotta compare that against the fuel map:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/STG2%20Fuel%20v1.0.1.png~original

 

As you can see, you hit 40g/s as low as 0.75g/rev, closed loop cells are until 1.40g/rev at 70-75mph (2800-3200rpm ranges). As you can also see 80g/s goes into 1.60g/rev which is already in open loop (cells that are less then 14.7 are considered open loop). For this reason I decided to set my max learning to be around 60g/s (though reviewing the data now going up to 65-70g/s might be better).

 

Changed the learning view ranges to be 5.60, 30, 60 and max learning switchover to be 59g/s.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/AF%20LearningV1.6%205-30-60.png~original

 

Here is the stock AF learning ranges learning view from 122m highway trip last week, I got 30.80mpg mostly highway, with pretty heavy traffic.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Fueling/FPR/05STI-FPR_StockAFLearning_30.8mpg.jpg~original

 

And finally here is the new AF learning ranges, 120m highway (to the same city as above), moderate traffic, got 30.03mpg, which is 2.5% less then last week, but I would say within margin of error.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Fueling/D-Learning/LV_AFL1.6_5-30-60.jpg~original

 

 

TL;DR: Setting D learning to 80g/s is too high, you might be learning open loop AFR's. New AF ranges yielded 2.5% less mpg over stock AF ranges.

[covertrussian]

Timing Compensation Zeroing

 

This compensation adds additional timing ontop of the base maps to two cylinders. This might be fine for stock cars, but becomes fairly dangerous for Stage 1+ setups.

 

Good Samaritan Max Capacity pointed me towards another flaw in the stock Legacy Tune. 05-06 Legacy's add 2* to Cylinders 2 and 4 which starts being a real big issue once you start increasing the power. Subaru saw this as an issue and zeroed out the tables stock on 07+ Legacy's.

 

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing/Timing%20Compensation/TimingComp-Stock.png~original

 

As you can see stock tune adds 2.11* starting at 3600rpm.

 

Fix is simple, just zero it all out!

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing/Timing%20Compensation/TimingComp-Zeroed.png~original

 

 

I wanted to do test and see if any power was lost due to less timing. I did two side by side runs on the same stretch of road. To avoid inconsistencies with self learning I set IAM to 1.0 as the default and did the run right after flashing the tune.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing/Timing%20Compensation/TimingComp-StockvsZero.jpg~original

 

As always take these tests with a grain of salt, I was surprised that I didn't loose any power, even more surprised it gained anything.

[hadvw]

Thanks so much! A few questions.. First, I'm not as RomRaider savvy as some. I've used it to log a bit, send those logs to my tuner, and flash new tunes.

 

I have a 2-3 year old installation of RR - from what I can tell, there have been more or less NO updates since then? The cars_def.xml on romraider.com seems to be dated from 2010/2012? Am I missing some more recent updates/experimental defs? Where to find them? The "subscribe here" thread seems to link to itself, to 4-5 year old defs..

 

Having said that, just looking at my tune:

1) CL/OL delay - all 0. Excellent!

 

2) AF Learning for Open Loop. Older revisions had 5.6/21/80.10. Latest revision has 5.6/21/45 - seems ok? I asked tuner for help correcting some misfires.. Seems better now, with new coil. Will probably go back to older tune..

 

3) Can't seem to find the AF #1 Learning Control parameter - this is why I asked about updated definitions. Where is it? (ROM is '05.5 5EAT, A2WC510C)

 

4) Timing Compensation. When I open this in RR, I see RPM ranges (800/1200/1600/2000...) but only a single row - no Engine Load rows, just one row per cylinder. Cyl B/C are all 0. Cyl A/D have some NEGATIVE timing - possible correction by tuner due to logs/learning view?

 

Thanks!

[covertrussian]

Since Romraider Def's are in a single file for all cars, it becomes very hard/time consuming to maintain all cars. ECU flash uses a better system where each car & ecu has it's own XML file. That means it's much easier to update one car at a time.

 

Download ECU Flash and use the definitions that came with it, RomRaider ECUFlash definitions are also way out of date. The default ECUFlash definitions will have all of these tables in them.

 

I think RomRaider's color schemes are much easier to tune and compare with, thus I still like to use RR for screenshots when I can, but I do most of my tuning with ECU Flash (mainly because of speed and keyboard shortcuts).

 

2.) Your old version of AF Learning disabled it, while the latest version is more like stock (where it's enabled). Unless your tuner knocked down the max learning range to be ~45 you will still be open loop learning.

 

4.) Timing compensation, you are running much older version of the rom most likely. 2005's used to have just one row, but then by last couple revisions they had a couple rows. I would suggest upgrading to the latest version of the rom for your car, they usually have other fixes too.

[hadvw]

Since Romraider Def's are in a single file for all cars, it becomes very hard/time consuming to maintain all cars. ECU flash uses a better system where each car & ecu has it's own XML file. That means it's much easier to update one car at a time.

 

Download ECU Flash and use the definitions that came with it, RomRaider ECUFlash definitions are also way out of date. The default ECUFlash definitions will have all of these tables in them.

 

I see. Hmm. I have ECUFlash from 2014 (1.44.4234). I downloaded 1.44.4799, will try that. With 1.44.4234, no changes from what I reported earlier. I don't see the AF #1 control, and per-cylinder compensation is still one row.

 

 

2.) Your old version of AF Learning disabled it, while the latest version is more like stock (where it's enabled). Unless your tuner knocked down the max learning range to be ~45 you will still be open loop learning.

 

4.) Timing compensation, you are running much older version of the rom most likely. 2005's used to have just one row, but then by last couple revisions they had a couple rows. I would suggest upgrading to the latest version of the rom for your car, they usually have other fixes too.

 

Is it possible to convert to the newer ROM automagically, or do all the values need to be copied over manually?

[covertrussian]

Your rom is probably old enough not to have these tables defined (no point if newer version has them all already). No way to automatically do it, gotta copy all tables one by one, RomRaider's Copy Table feature is really good for this though.

[NSFW]

Fixing the Tau Tables

 

There are a set of tables in the ECU that add fuel when load is rising, and reduce fuel when load is falling. This may sound similar to tip-in, but recall that tip-in happens in response to throttle changes; these tables respond to load changes. I suspect that these tables were intended by the factory tuners to compensate for fuel that sticks to the intake passages, a phenomenon called "wall wetting" which is denoted "tau" in academic papers.

 

In stock form, these tables are great for basic driveability. If you zero them out completely, the engine stumbles when slowing to idle or when revving up from idle. However the stock tuning of these tables is terrible for high-throttle stuff. Try stomping the gas pedal with the logger running, and you'll see your AFR dip much lower than the fueling table calls for. On my car, AFR would dip below 10:1 even though my fuel table never calls for anything richer than 11:1. That's rich enough to feel power drop during the AFR dip, and then come back when AFR returns to normal. Fixing these tables cut that problem in half.

 

Subaru's tau model is actually more complex than the three tables shown below, and there are a few other tables in the ROM that are not part of the RomRaider/EcuFlash definitions. However, these three tables are sufficient to address the high-load problems with the factory tune.

 

Here is how I have configured mine:

 

http://legacygt.com/forums/attachment.php?attachmentid=240115&stc=1&d=1478054519

 

Tau Input A Rising Load Activation: The key thing here is that once the engine is warmed up, this table tapers to zero loads of 1.2 and higher.

 

Tau Input A Falling Load Activation: This table zeros out once the engine has warmed up.

 

Tau Input B Activation (Engine Load): This table zeros out at loads of 0.8 and higher. The low-load values are stock.

 

Other than the changes called out above, I've left these tables pretty much stock. I have not noticed any problems with the factory tau tuning at low load, so I haven't had any reason to change that stuff.

 

Note that even with the changes shown above, my AFR does still dip richer than my tune calls for, by about 0.75:1. I'm confident that the ECU is not causing that, however. I suspect that the remaining rich dip is caused by a change in the airflow measured at the MAF sensor vs. airflow entering the engine while the intake tract is increasing. Speed-density would theoretically not have that problem, but I haven't tested that theory yet.

 

The rich dip caused by these tables drove me mad for quite some time. I spent a lot of hours investigating the ECU code that uses these tables, and documenting what I was finding, so that they could be added to the ECU definitions for RomRaider and EcuFlash. I did it all for completely selfish reasons, but I still like to think of it as my way of saying thanks to the open source tuning community.

[covertrussian]

Could these tables be related to the infamous FMIC rich dip? Perhaps the FMIC rich dip is there even on TMIC's just not as noticeable due less turbo lag?

[hadvw]

Once again, no luck finding these tables.

 

Are you showing ECUFlash screenshots, or RR? I downloaded the latest ECUFLASH, and I don't see anything labeled "Tau" - I'm assuming it's under fueling? Or does it need to find these IN the tune before it will show them? I thought the .xml defs were just a way to assign labels to different addresses in the ROM? Or can it actually affect the way the ROMs are interpreted by the ECU? Such as the "old" style cylinder compensation - my tune shows one row, "newer" tunes apparently have multiple rows. Does the ECU change its operation if you flash a newer tune? Or am I only seeing partial tables?

 

Edit: yeah, sorry for all the questions.. I guess we could just start another thread for questions?

[covertrussian]

What rom version are you running? It it's possible that your rom has the tables but the definition file (xml) is defined to only show one row.

 

I personally haven't played with Tau tables, thus dunno if I saw them either.

 

On another note, I might ask the mods to move the questions to it's own thread.

[hero2zero]

Once again, no luck finding these tables.

 

FWIW, I had to download the latest version of EcuFlash (1.44.4799) in order to see the Experimental fields. It was not visible on version 1.29. Also, I have an 05 OBXT on A2WC522S which inherits A2WC521N. This is what the Experimental fields look like on mine (attached picture).

 

covert, NSFW,

Thanks for the info. I've been lurking on this and the RR forums and have read your posts many times. This hasn't been terribly easy to learn (tuning), but with your help along with the others here, it has made open source tuning much more enjoyable.

[utc_pyro]

Once again, no luck finding these tables.

 

hadvw, you have an '06 as well correct? I've found some of these in our ROM. Will post the EcuFlash xml for you when I get home.

[covertrussian]

covert, NSFW,

Thanks for the info. I've been lurking on this and the RR forums and have read your posts many times. This hasn't been terribly easy to learn (tuning), but with your help along with the others here, it has made open source tuning much more enjoyable.

 

Thanks bud! I'm glad it's useful. I personally have to thank NSFW for all of his awesome threads/posts learned a lot when from him when I was first starting out with these cars.

 

On another note, I have a bunch of other posts cleanup and post here from my build and fuel economy threads, will be a bit before I get a chance to find them all .

[NSFW]

Could these tables be related to the infamous FMIC rich dip? Perhaps the FMIC rich dip is there even on TMIC's just not as noticeable due less turbo lag?

 

Yes, those tables cause about half of the FMIC rich dip on my car. The other half actually is the FMIC, I think.

 

I'm guessing that the dip that TMIC users see is probably caused almost entirely by those tables, but I haven't seen before-and-after logs to know for sure. I have heard from TMIC users that the dip does happen and tweaking the tables does help, but I didn't dig into this stuff until long after I'd switched to a FMIC.

[need bb]

Yes, those tables cause about half of the FMIC rich dip on my car. The other half actually is the FMIC, I think.

 

I'm guessing that the dip that TMIC users see is probably caused almost entirely by those tables, but I haven't seen before-and-after logs to know for sure. I have heard from TMIC users that the dip does happen and tweaking the tables does help, but I didn't dig into this stuff until long after I'd switched to a FMIC.

 

The TAU tables fixed my rich dip issue and I run a TMIC setup. Almost all the WRX/STI dyno sheets I have seen show this rich dip, and I assume most tuners don't know how to fix it via modifying the TAU tables.

[covertrussian]

Now I wonder if these TAU tables were used to compensate for closed loop to open loop delay leanness?

 

Looking at my AFR graphs and I sometimes have rich dips (10.5ish) and other times I don't. I think it depends on how I roll onto the throttle as I did the log.

[NSFW]

The TAU tables fixed my rich dip issue and I run a TMIC setup. Almost all the WRX/STI dyno sheets I have seen show this rich dip, and I assume most tuners don't know how to fix it via modifying the TAU tables.

 

Cool, I'm glad to hear it helps. How much richer did your AFR get before and after you made the change? With a FMIC, mine was dipping by 1.5:1 before, and 0.75:1 after.

 

And yes I think the knowledge of the Tau tables has been very slow to spread. One of the best known tuners in my area didn't seem to know what I was talking about when I asked him a few months ago if Cobb had these tables.

 

It just hit me that I should probably post on NASIOC to get the word out... I assumed that wouldn't be necessary, but these tables have been documented at RomRaider.com for about 4 years now. And tweaking them will fix a drivability issue that affects just about everybody. RomRaider.com is totally public, but I don't think a lot of people follow it.

[covertrussian]

Tuning for Intakes

 

In this post I will cover my tuning procedure for a proof of concept WAI. These testing and tuning steps should apply to all intakes, but keep in mind not all intakes are created equally, some will require a lot more fiddling with the MAF scales. Since I used OEM MAF housing and flow straightener my adjustments were very simple. Original thread on this can be found here.

 

 

Custom WAI

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/WAI%20v1/WAIv1_013.jpg~original

 

Did a log with the new intake and compared it another log from the same road. As you can see I stopped doing the pull around 4500rpm, I didn't want to risk it going further.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/WAI%20v1/VDyno%20-%20Stock%20Airbox%20vs%20WAIv1%20Intake%20before%20Tune.png~original

 

First thing that really stands out is much leaner AFR's, about 3 points leaner, instead of a good and safe 10.5-11 AFR's, car leaned out to 13-13.5 AFR. Next thing was overboosting, about .5psi more. Running more boost and being lean will absolutely destroy your engine. As you can see my AFR's were in 13's, when max safe is around 11.5 AFR on stock TMIC.

 

Now that we covered the easy to spot things, lets look at the hidden reason behind such drastic AFR changes.

 

Mass Airflow (g/s)

Not exactly sure why, but the less restrictive intake caused the mass airflow reading to be much lower, about 25% lower. The ECU basically thinks that we are receiving less air, thus it adds less fuel through the fuel maps when in fact we are receiving more air then before.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Tune%20-%20WAIv1%20-%20Mass%20Airflow.png~original

 

My thought on this is, it should have been reverse, more airflow from intake should cool the MAF hot wire more, needing more voltage to keep it warm, thus reading higher MAF Voltage, which calculates to a higher grams per second airflow. Maybe someone more knowledgeable on this can chime in?

 

Engine Load (g/rev)

Because the ECU is calculating lower grams per second, our Engine Load grams per rev is now much lower.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Tune%20-%20WAIv1%20-%20Load.png~original

 

This can be visualized through the trace line. We are stuck at 1.80 g/rev instead of 2.40 g/rev. On this map 1.80 g/rev is meant for about 7psi of boost and 2.40 g/rev is meant for about 14psi. New intake overboosts to 14.2psi and stays at 1.80g/rev, this causes the car to run 7psi's fuel values while your actually at 14psi.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Trace%20WBO2%20-%20Fuel%20Map.png~original

 

Timing

Not only are you running leaner, you are also running more timing, this as because we are stuck in a lower column meant for less boost.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Tune%20-%20WAIv1%20-%20Timing.png~original

 

Further visualized in this map trace (ignore the bottom trace for now). As you can see the car is running 21* instead of 13-17*. That alone, even with non lean AFR's can kill a motor.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Trace%20Timing%20-%20Stock%20vs%20WAI%20vs%20WAI%20tuned.png~original

 

 

Tuning for these changes

Finally, if you haven't been pushed away from intakes all together, we can start adjusting the tune for this mod. These are just the basics, keep in mind that every intake is different and might not be as easy to adjust the tune for (especially if the MAF tube diameter and MAF placement is different from stock).

 

MAF Sensor Scaling

First I calculated the difference between the Mass Airflows between the two graphs and came up with about 24%. I increased the MAF Scaling by 25%.

 

Initial & Max Wastegate Duty Cycle Tables

Next I calculated the difference between the amount of boost I was running, came up with 7.7%, removed 8% from all of the values.

 

Here is the boost map trace, we are back to our previous boost numbers. Now you could always keep the added boost and just adjust your timing and fuel for added PSI, but I personally wanted a 1:1 comparison of the power gains from the intake only.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Trace%20Boost%20-%20Stock%20vs%20WAI%20vs%20WAI%20tuned.png~original

 

Primary Open Loop Fueling

I didn't have to make any changes to the fuel table because my AFR's were identical to what they were before.

 

This map trace compares the three runs AFR's, as you can see the most MAF scaling increase and boost decrease AFR's were back to what they were before.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/WAIv1/Trace%20Fuel%20-%20Stock%20vs%20WAI%20vs%20WAI%20tuned.png~original

 

Ignition Base Timing

At first I didn't have to adjust the timing, but driving higher IAT (which a WAI intake is more prone to higher IAT's) I saw some knock and reduced timing up top.

 

Timing Compensation (IAT)

Warm air intakes soak up a lot of heat and hit high IAT's (Intake Air Temperatures) much faster then with stock intake. On a 90*F day sitting at a stop light and slow city driving got me 150*F in just a few minutes. I saw the ECU pulling timing due to knock at fairly low loads (1.00-1.20g/rev).

 

You should probably update the Timing compensation table to retard more timing at higher IAT's to avoid detonation.

 

 

Custom WAI Power Gains

The stock intake is a very good cold air intake, but while it can deliver colder air it's still very restrictive. The common notion that an intake is not needed or doesn't gain any power below 300-350hp has been debunked by GrimmSpeed, with their intake making around 20whp over stock intake at similar boost levels and minimal tuning. Their intake did so well because it has minimal bends and has a very well designed airbox that reduces hot engine air contamination.

 

When my IAT's stayed down (drive around for couple miles before doing the pull) I saw similar gains in Virtual Dyno on the same stretch of road 30 minutes apart.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/WAI%20v1/VDyno%20-%20Stock%20vs%20WAIv1%20BPI%20Flowstack.png~original

 

Custom WAI Conclusion

The biggest issue is with a new intake is MAF voltage being less, which gets calculated to lower engine load, which makes your ECU to run less fuel and more timing for the same amount of boost.

 

Next issue is WAI specific: keeping the IAT's low to avoid timing retard (less timing = less power) and to avoid detonation. Thus I now believe a sealed Airbox is a must, either buy one with it, or make one for your existing intake.

 

I also found that keeping the MAF location and diameter the same as stock (because I'm using the stock maf housing), still requires a retune (now this could be because of the flow stack, if I get a regular cone filter I'll test this out more).

 

 

 

 

K&N Typhoon Intake

 

K&N claims that the Typhoon intake does not need a retune, most of us know not to trust those statements, but new members tend to trust K&N's word over ours. For this reason wanted to test the stock intake and the K&N Typhoon and see for myself if it needs a retune and why it needs a retune.

 

If you want to read the quick explanation, just read the AFR section. After the AFR section I will dig further into the technical details.

 

AFR

Tune stayed the same, the logs were done within an hour of each other.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/Intakes%20AFR-PSI_Stock%20vs%20KampN%20Typhoon.png~original

 

Here's a more detailed MAP trace of AFR averages:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/Trace%20-%20AFR.png~original

 

This is on my Stage 2 tune which is slightly leaner then 100% stock tune (10.00-10.50afr). Stock intake AFR's are close to 10.50, with the K&N they were close to 11.40. That's almost a whole AFR point leaner, TMIC cars need to be 11.00 afr or less to be safe. Even on a 100% stock car you would still be pushing above 11.00 AFR.

 

Next is Boost, this intake cased me to overboost by .5-1psi. This is not a big issue if your tuned for it, but remember K&N says no tune is required.

 

K&N resized the diameter of the whole intake to be smaller in diameter then the stock intake, K&N is 65mm while stock is 70mm, it's a dirty trick to reduce the need for a tune at the cost of less air. As you will see from these graphs, the trick does work though, since it didn't lean out as badly as my custom stock sized intake did.

 

 

Timing

Looks like timing stayed about the same, this is probably because my timing map doesn't change drastically between 2.20-2.70g/s. More on this in Engine Load section.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/Timing.png~original

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/Trace%20-%20Total%20Timing.png~original

 

Airflow

From these two graphs you can see that K&N intake is running less MAF voltage, which calculates to less mass airflow, which calculates less Engine load and ECU runs less fuel and more timing (based on the Fuel and Timing maps). This is why all intakes need a retune.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/MAF%20Voltage.png~original

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/Mass%20Airflow.png~original

 

Engine Load

As mentioned int he airflow, less calculated airflow = less calculated engine load. This part is what determines what column you will use for fuel and timing. With less calculated engine load the ECU puts you in a column that's meant for less boost, which means you run less fuel and more timing.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/Engine%20Load.png~original

 

Now you probably noticed from the traces that I'm running in the same columns, don't let that deceive you! Since we have a limited amount of rows and columns on a map, we have to use big load ranges and let the ECU interpolate (calculate) between our columns. This is why it's important to look at graphs too, and not just map traces.

 

Wastegate Duty Cycles

K&N is actually running less wastegate duty cycles, even though it was overboosting. This is probably Turbo Dynamics at work trying to reduce overboosting.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/WGDC.png~original

 

 

Learning Views

Finally the ECU Learning views, these show the learned airflow adjustments and learned timing adjustments

 

Stock Intake:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/LV_StockIntake.jpg~original

 

K&N Typhoon:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/KN%20Typhoon/LV_KampNTyphoon.jpg~original

 

ECU saw a good bit more knock with the K&N Typhoon, this was a colder day (70's F), on a hotter day I bet the knocks count would have been higher.

 

K&N Typhoon Conclusion

These cars have very sensitive MAF's and ECU's, which makes them very sensitive to any airflow changes, be it intake or exhaust. Even running an aftermarket paper panel filter will mess with your AFR's, and aftermarket high flow filters will be even worse.

 

Moral of the story, either run a stock intake with stock paper media, or get a tune.