Welcome to the North American Subaru Impreza Owners Club Friday July 25, 2014
Home Forums WikiNASIOC Products Store Modifications Upgrade Garage
NASIOC
Here you can view your subscribed threads, work with private messages and edit your profile and preferences Home Registration is free! Visit the NASIOC Store NASIOC Rules Search Find other members Frequently Asked Questions Calendar Archive NASIOC Upgrade Garage Logout
Go Back   NASIOC > NASIOC Technical > Normally Aspirated Powertrain

Welcome to NASIOC - The world's largest online community for Subaru enthusiasts!
Welcome to the NASIOC.com Subaru forum.

You are currently viewing our forum as a guest, which gives you limited access to view most discussions and access our other features. By joining our community, free of charge, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is free, fast and simple, so please join our community today!

If you have any problems with the registration process or your account login, please contact us.
* Registered users of the site do not see these ads.
Reply
 
Thread Tools Display Modes
Old 11-26-2008, 01:10 AM   #1
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default 05 2.5rs @ 124awhp :)



124awhp@5540rpm and 134lbft@3770 and estimated 134mph vehicle max speed. And check out how linear the power is! Woot!

This was on the custom tune that a friend and i made for the 2005 2.5rs. Did some logging to see how it was performing in the winter. Downloaded Dataloglab to take my romraider files. I looked up EVERYTHING as far as info for my car goes, the correct drag coefficient, front area, curb weight, gear ratios, final drive, my weight, estimate weight of the gas, current outside temp and humidity for dataloglab (sound a lot like Bob Lob Law from Arrested development?).. Soooo should be relatively accurate.

Power related mods
- borla cat-back
- intake "snorkel-box-thing" delete

non power related mods
- carbotech panther xp10 brakes f+r
- megan racing street series coilovers (8k/6k)

and a pretty picture of the gal:



Sidenote: 28 degree F intake temps FTW! Although, compared to the logging that I did in the summer with ~75F intake temps it was logging around 120awhp still.

P.S. sorry for the bad iphone pic... couldn't figure out how to take a screen capture in windows on my bootcamped macbook... no prtscrn button :S
* Registered users of the site do not see these ads.

Last edited by rectangular; 11-26-2008 at 02:55 AM.
rectangular is offline   Reply With Quote
Old 11-26-2008, 09:31 AM   #2
Gr3en
Scooby Specialist
 
Member#: 160991
Join Date: Oct 2007
Chapter/Region: MWSOC
Location: Gorillaoffroadcompany.com
Vehicle:
05 Corn Burner
Dirt Silver?

Default

Congrats on the powa!
Gr3en is offline   Reply With Quote
Old 11-26-2008, 09:53 AM   #3
Kevin Thomas
Street Racing Instructor
Moderator
 
Member#: 110
Join Date: Jul 1999
Location: 1997 OBS, 1996 SVX, 1988 RX
Vehicle:
1989 1989 XT6

Default

Nice dyno curve and pic. Thanks for posting the plot. Do you plan on doing more mods and posting up more curves?
Kevin Thomas is offline   Reply With Quote
Old 11-26-2008, 11:17 AM   #4
chazly413
Scooby Specialist
 
Member#: 151079
Join Date: Jun 2007
Chapter/Region: MAIC
Location: College Park, MD
Vehicle:
2005 2.5RS RBP
motec system exhaust-RIP

Default

Do you know what kind of numbers you put down stock?? I like how the horsepower doesn't taper Sweet pic too!!
chazly413 is offline   Reply With Quote
Old 11-26-2008, 12:29 PM   #5
Un0RiGiNaL
Scooby Specialist
 
Member#: 189748
Join Date: Sep 2008
Chapter/Region: MAIC
Default

snorkel delete doesnt do anything to power. but itt sounds vicious i did it to my 2.5i
Un0RiGiNaL is offline   Reply With Quote
Old 11-26-2008, 01:58 PM   #6
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

yeah, the snorkel delete actually lowered intake temps ~3-5 degrees, probably not noticeable, more so for sound Although I did the K&N reusable filter the same time too...

When i flash the stock map on it ends up being around 110-115awhp. I would assume without the cat-back, and K&N filter it would be slightly lower than that as well. I've always been told that the 2.5rs has around 100awhp stock. This seems reasonable from my experience.

If one thing is the most noticeable, I adjusted the CL/OL settings so that after 3700rpm it would always switch over, and anything above 70% throttle it would always switch over. On the stock tune it required 110% throttle to go into OL at any coolant temp :S The throttle response is a million times better now.

I am kindof in the plans for saving for an 09 wrx that would quickly be stage 2 and custom tuned as well.

While the 08 was still disgracing the scene I had plans for cams, headers, lightened pully and some other small stuff for my RS. But I've really grown to love the 09 wrx hatch...

Previously my goals were the same as every 2.5rs owner: wrx power levels... without the turbo lag, and with the shorter gears

This tune has been proven at Grattan track this fall, and did really well. It was a lotus event. In the intermediate run group I was passing some lotus's and porsche's. Does good for driver confidence :P It's super easy and predictable at the track with how linear the power delivery is.

All in all, romraider is usefull even for us NA folks.
rectangular is offline   Reply With Quote
Old 11-26-2008, 04:18 PM   #7
chazly413
Scooby Specialist
 
Member#: 151079
Join Date: Jun 2007
Chapter/Region: MAIC
Location: College Park, MD
Vehicle:
2005 2.5RS RBP
motec system exhaust-RIP

Default

Quote:
Originally Posted by rectangular View Post
yeah, the snorkel delete actually lowered intake temps ~3-5 degrees, probably not noticeable, more so for sound Although I did the K&N reusable filter the same time too...

When i flash the stock map on it ends up being around 110-115awhp. I would assume without the cat-back, and K&N filter it would be slightly lower than that as well. I've always been told that the 2.5rs has around 100awhp stock. This seems reasonable from my experience.

If one thing is the most noticeable, I adjusted the CL/OL settings so that after 3700rpm it would always switch over, and anything above 70% throttle it would always switch over. On the stock tune it required 110% throttle to go into OL at any coolant temp :S The throttle response is a million times better now.

I am kindof in the plans for saving for an 09 wrx that would quickly be stage 2 and custom tuned as well.

While the 08 was still disgracing the scene I had plans for cams, headers, lightened pully and some other small stuff for my RS. But I've really grown to love the 09 wrx hatch...

Previously my goals were the same as every 2.5rs owner: wrx power levels... without the turbo lag, and with the shorter gears

This tune has been proven at Grattan track this fall, and did really well. It was a lotus event. In the intermediate run group I was passing some lotus's and porsche's. Does good for driver confidence :P It's super easy and predictable at the track with how linear the power delivery is.

All in all, romraider is usefull even for us NA folks.
Nice dude, sounds like a very useful modification. How difficult would you say it is?
chazly413 is offline   Reply With Quote
Old 11-26-2008, 06:04 PM   #8
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

How difficult is it to tune the rs? Easy! Pick up a tactrix cable, romraider and ecuflash, pull the rom, do a pull, modify rom, reflash new ROM, do a pull, repeat as necessary.

It's very easy and there are certainly some noticabl gains. The stock tune and timing is retardedly conservative. It basically aims for stoich till 4,000rpm then richens hardcore to 11.5AFR the rest of the way. Definitely not the correct range for max torque/power, but a way to keep it safe for most drivers who just drive this car around town.

There's also some timing gains to be made if you don't mind retuning for 93 octane.
rectangular is offline   Reply With Quote
Old 11-26-2008, 06:30 PM   #9
RaceFaceXC
NASIOC Supporter
 
Member#: 165749
Join Date: Dec 2007
Chapter/Region: South East
Location: stupid Florida
Vehicle:
2007 No YELLING
on the bus!

Default

The stock timing on 07 2.5i's (05RS's too, according to Williaty) is anything but conservative. On 87 oct the ECU isnt even using all of the programmed timing advance. so changing the timing will only cause the ECU to pull more or cause more knock. the mods you have done shouldn't do anything to help gain performance from advanced timing. The biggest gains i have found are from AFR tuning.
RaceFaceXC is online now   Reply With Quote
Old 11-27-2008, 12:36 AM   #10
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by rectangular View Post
Downloaded Dataloglab to take my romraider files. I looked up EVERYTHING as far as info for my car goes, the correct drag coefficient, front area, curb weight, gear ratios, final drive, my weight, estimate weight of the gas, current outside temp and humidity for dataloglab
I've spent a lot of time playing with Data Log Lab. I was as diligent as you in looking up all the specs. I even got my car weighed. However, now that I finally have a real dyno plot for my car, the DLL plots are just badly, badly wrong. Subtract a good 15ftlbs, minimum, from whatever DLL is telling you to get it to match up to a Dynojet.

Speaking of which, does anyone know what a stock, pre-06 2.5RS typically puts down on a Dynojet?


Quote:
Power related mods
- borla cat-back
- intake "snorkel-box-thing" delete
Neither of those are power-related mods. Cat-backs get you nothing. Do you mean the snorkel or the snorkus? Changing the snorkel to the WRX intake trumpet does lower IATs meaningfully. Deleting the snorkus (big plastic gourd thing in the fender) doesn't increase power at all and does cause problems on an 05RS. Did you read the sticky at the top of this forum?

Quote:
P.S. sorry for the bad iphone pic... couldn't figure out how to take a screen capture in windows on my bootcamped macbook... no prtscrn button :S
Try virtualizing Windows instead. Then you can just cmd-shift-3 the screenshot

Quote:
Originally Posted by rectangular View Post
yeah, the snorkel delete actually lowered intake temps ~3-5 degrees, probably not noticeable, more so for sound Although I did the K&N reusable filter the same time too...
See above.

Quote:
When i flash the stock map on it ends up being around 110-115awhp. I would assume without the cat-back, and K&N filter it would be slightly lower than that as well. I've always been told that the 2.5rs has around 100awhp stock. This seems reasonable from my experience.
This makes no sense. Stock calls for more timing that the engine can actually run. So you didn't make any more power from timing. With the ECU holding the engine at stoich for so long and then dumping fuel into it like it's trying to make a puddle once it switches, there's definitely a little power to be found there. However, I can't believe ~10-15whp. I'd have to see eddy-current dyno results before I believe that one.

Quote:
If one thing is the most noticeable, I adjusted the CL/OL settings so that after 3700rpm it would always switch over, and anything above 70% throttle it would always switch over. On the stock tune it required 110% throttle to go into OL at any coolant temp :S The throttle response is a million times better now.
I totally agree with you that altering the stock Closed Loop to Open Loop transition behavior is the biggest single improvement you can make to the stock tune. However, you are mis-understanding how the CL/OL Transition with Delay process works. A number of parameters (such as RPM, engine load, throttle position, etc) can start a timer ticking. When the timer expires, the ECU will transition to Open Loop if the Primary OL Fueling map calls for more than a certain amount of enrichment. The stock tune had 110% in some cells for throttle position to, in effect, disable transition based on throttle position. It still was evaluating all the other parameters even though it can never hit the throttle position trigger. If you zero the delay counters (length of the timers), then the CL/OL Trans Delay logic shortcuts all the way to the OL Fueling table. In other words, with 0 delay, the OL Fueling table, and only the OL fueling table, determine wither the ECU is in CL or OL and it will transition instantly between the two.

Quote:
Originally Posted by rectangular View Post
The stock tune and timing is retardedly conservative.
No, the stock timing is retardedly aggressive. It's pretty obvious that the stock tune is designed to maximize fuel economy and get the car juuuuust out of warranty before the engine burns down. It knocks like a mofo pretty much all the time. On 87*, I've seen 05 RSes pull as much as 12* of timing just at cruise on the freeway. On a stock tune, the engine can't even come close to using all the timing the map calls for on 87 octane gas. In fact, the car even knocks a little on 94 octane gas.

Quote:
It basically aims for stoich till 4,000rpm then richens hardcore to 11.5AFR the rest of the way. Definitely not the correct range for max torque/power, but a way to keep it safe for most drivers who just drive this car around town.
Again, you're not quite understanding how the CL/OL process works. It holds stoich until one of a number of parameters is met and then a timer expires. If you log the parameters that can trigger CL/OL transition and then log the CL/OL status, you'll note that putting your foot all the way down, even at idle, will send the engine load high enough to surpass the threshold for CL/OL transition. This starts the timer. The timer keeps counting for as long as it's set to. During this time, the engine happens to be speeding up. The timer expires and the ECU drops into OL and begins to read the enrichment targets from the OL fueling map. Now, because the delay is time based, you can make the ECU transition at pretty much any RPM you feel like just by altering the conditions. Want to get into OL at a lower RPM? Use a higher gear or drive uphill (engine accelerates slower, therefore timer expires at a lower RPM). Want to get into OL at a higher RPM? Drive downhill or use a lower gear.
williaty is offline   Reply With Quote
Old 11-27-2008, 02:13 AM   #11
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

williaty, i appreciate the very extensive response and many of them are certainly things I took into consideration along the way, but here's my defense :P

Quote:
Deleting the snorkus (big plastic gourd thing in the fender) doesn't increase power at all and does cause problems on an 05RS.
Yes, I ditched the "gourd" in the fender. If you are talking about the turbulence issues causing a rich patch from MAF sensor volt turbulance response, in my experience and from logging this has not happened to me. ALSO the "gourd" I removed in the fender is BEFORE the filter (which would largely reduce turbulence), and BEFORE the helmholtz resonator. Correct me if i'm wrong, but i've been lead to believe the helmholtz is actually the little banana/question mark looking plastic guy sticking out of the side after the filter box... this was not removed/moved and is NOT in the fender, at least on the 05 rs. The whole purpose the helmholtz is to remove turbulence. I actually think the sticky at the top of this forum erroneously identifies its location (although it's been awhile since i've read it start to finish in depth), unless there are two helmholtz resonators, one in the fender and one post air filter box. The "gourd" I removed however doesn't have anything that resembles the one helmholtz resonator I know of.

Everything between air filter box and engine is still stock (aside from the K&N filter). Turbulence has not been an issue in my case.

Removing the "gourd" in the fender lowered my intake temps very slightly (probably not a noticeable amount, but any little bit counts), removes some restriction, and makes it sound better. It's a win in my book so far.

http://blog.autospeed.com/2003/11/30...-into-intakes/


^^ that little guy is the helmholtz resonator I'm referring to (credit to whoever this is image i stole :P )

Quote:
the DLL plots are just badly, badly wrong. Subtract a good 15ftlbs, minimum, from whatever DLL is telling you to get it to match up to a Dynojet.
We all know that even dynos put out different numbers depending on make so they really aren't absolutely perfectly definitive. I'm sure there are plenty of drivers who could drive a 50hp car tons faster than I could a 300hp car anyways. Can only get so good with a couple track days a year :P . So in reality slight hp differences are probably not a big issue. It fluctuates so much depending on temperature, humidity, elevation and other factors anyways.

The original post was mostly me just being an excited idiot since I was expecting it to plot out a lot less. I suppose take these numbers with a grain of salt... and we're arguing over like +/- 15... although since it's almost 15% of my cars power perhaps it is worth debating :P

Quote:
No, the stock timing is retardedly aggressive. It's pretty obvious that the stock tune is designed to maximize fuel economy and get the car juuuuust out of warranty before the engine burns down. It knocks like a mofo pretty much all the time. On 87*, I've seen 05 RSes pull as much as 12* of timing just at cruise on the freeway. On a stock tune, the engine can't even come close to using all the timing the map calls for on 87 octane gas. In fact, the car even knocks a little on 94 octane gas.
Interesting... this certainly wasn't the case in my car. As this is the only 2.5rs I have tuned so I can't really say much other than that this was not so with mine. My car was adding timing all over the map even with 89 octane. I made note worthy additions of timing through a long process of adding some timing, doing some pulls, reading the logs, and advancing more timing. Between 7 to 14 timing advanced in a majority of the cells between 2000rpm and 6400rpm. The car is STILL advancing timing, around 3-9 units added on my logged pulls with 91 octane. Since this is my DD map, I think I will stick with this. The more aggressive map will come later...

Since this was my first time tuning a car, a friend also double checked all my modifications since he is rather experienced.

The car simply just wanted more. Perhaps they changed the tune mid cycle? What year RS did you notice this very aggressive timing issue on? What elevation do you live at? I'm at ~600ft. Are you looking at fine learned knock correction or knock correction advance? What's the ignition total timing in your logs? Mine's at 30 ~4200rpm and climbs upto 35 by 6400rpm.

Quote:
Neither of those are power-related mods. Cat-backs get you nothing.
I suppose that depends on your defintion of power related mods is. Was it a td04 -> vf39 improvement? Far from it. Did i notice a slight improvement when I did the intake/K&N/cat-back modifications. Very slight each time, and the most of all when I did the cat-back. Could it have been the placebo effect? Maybe. Posting the power mods wasn't really for bragging in any way. It was more so to inform people that this wasn't achieved by just tuning a stock car and that other factors may be affecting the final number.

Quote:
Try virtualizing Windows instead. Then you can just cmd-shift-3 the screenshot
I thought about it, but I A. didn't feel like paying $75 for parallels or vmware and B. since it was dealing with reflashing my car, I decided that running windows not virtualized would be the most stable way (almost an oxy-moron :P ) when it comes to the openport drivers and ecuflash. I certainly don't want vmware or parallels to crash during a flash.

RE: OL/CL stuff

I realize that there are multiple values that lead up to the cars final decision. I was simply pointing out in the case of the table regarding CL/OL and coolant temp. Obviously subaru would never put this car on the market with a tune that would never hit OL. Bad 07/08 STI tunes maybe, nonexistant OL, no. :P

Quote:
Quote:
It basically aims for stoich till 4,000rpm then richens hardcore to 11.5AFR the rest of the way. Definitely not the correct range for max torque/power, but a way to keep it safe for most drivers who just drive this car around town.
Again, you're not quite understanding how the CL/OL process works. It holds stoich until one of a number of parameters is met and then a timer expires. If you log the parameters that can trigger CL/OL transition and then log the CL/OL status, you'll note that putting your foot all the way down, even at idle, will send the engine load high enough to surpass the threshold for CL/OL transition. This starts the timer. The timer keeps counting for as long as it's set to. During this time, the engine happens to be speeding up. The timer expires and the ECU drops into OL and begins to read the enrichment targets from the OL fueling map. Now, because the delay is time based, you can make the ECU transition at pretty much any RPM you feel like just by altering the conditions. Want to get into OL at a lower RPM? Use a higher gear or drive uphill (engine accelerates slower, therefore timer expires at a lower RPM). Want to get into OL at a higher RPM? Drive downhill or use a lower gear.
I should clarify my statement. The stock OL fueling tables had the car aim for stoich until 4,000rpm, at which point it would become overly rich. Even aiming for 11.2:1 AFR at 6,000rpm.

Even if the car did switch over to OL pre 4,000rpm it would still be aiming for the stoich 14.7.

For the tables between 3,700rpm and 4,000rpm I aimed it more towards 13.9. Above 4,000 it goes to 13.8 and eventually getting to 13.1 closer to redline. I also made it more gradual instead of 14.7AFR to 12.8AFR instantly that the stock tune came with.

Anyways, I don't want this to end up as a flame war, take this as more of a debate. We've both obviously had different experiences that have led us to our different conclusions. Now we just need to find out why we disagree on many of these points.
rectangular is offline   Reply With Quote
Old 11-27-2008, 02:54 AM   #12
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by rectangular View Post
Yes, I ditched the "gourd" in the fender. If you are talking about the turbulence issues causing a rich patch from MAF sensor volt turbulance response, in my experience and from logging this has not happened to me.
You're probably at least the 25th person who has claimed they have logs proving this issue doesn't affect them. However, literally every single one of the people who have claimed that and then run the test properly and posted logs HAVE had the problem. You could have the one-in-a-million engine that doesn't do it, but the odds are massively against it.

Quote:
Correct me if i'm wrong, but i've been lead to believe the helmholtz is actually the little banana/question mark looking plastic guy sticking out of the side after the filter box...
You're wrong The thing you posted a picture of is a quarter-wave resonator, not a Helmholtz resonator. The Helmholtz resonator is that big plastic gourd in the filter.

Quote:
ALSO the "gourd" I removed in the fender is BEFORE the filter (which would largely reduce turbulence)
Wither the filter reduces turbulence or not is open for debate as it hasn't been studied. However, turbulence is not the problem. Resonance is the problem and the filter will not do anything at all about resonance.

Quote:
The "gourd" I removed however doesn't have anything that resembles the one helmholtz resonator I know of.
Look at the lowest portion of the gourd. There's a kind of box looking structure with narrow necks coming off of it. That's a textbook example of a Helmholtz resonator. Likewise, the question-mark shaped thing between the airbox and the throttle body is a textbook example of a quarter-wave resonator. Both damp resonances but they do it in different ways. In this case, they're also tuned to different frequencies.
my case.

Quote:
removes some restriction
Testing has shown that the gourd doesn't provide a meaningful restriction to airflow.

Quote:

^^ that little guy is the helmholtz resonator I'm referring to
That's a quarter wave resonator. In fact, that resonator (literally the one on an 06-07 2.5i) is used as the example in a tuning textbook I have. BTW, the Autospeed link you posted actually confirms what I'm saying.

Quote:
My car was adding timing all over the map even with 89 octane.
The way you phrase that statement makes me think you weren't looking at the correct values. The "adding timing" rings some warning bells. What were you looking at? Knock Advance? Fine Learning Knock Correction (FLKC)? The reason I ask is that both of those values can be positive and the car can still be pulling timing. To find out if the car is pulling timing or not, you need to add the Base Timing table to the Maximum Advance table. If Base+(IAM*Advance)+FLKC is less than Base+Advance, the car is pulling timing. I'd bet what you were seeing is a low IAM with positive FLKC. That still means the car is pulling timing.


Quote:
Between 7 to 14 timing advanced in a majority of the cells between 2000rpm and 6400rpm. The car is STILL advancing timing, around 3-9 units added on my logged pulls with 91 octane.
Yeah, this basically proves you're not looking at the total picture. It's just not possible to advance 7*, let alone 14*, from the stock tune without running C16 and meth. Also, if the car can add timing when you put 91 in it, it was pulling timing before.

Quote:
Since this was my first time tuning a car, a friend also double checked all my modifications since he is rather experienced.
Thing is, Subaru (well, Denso, really) ECUs use some weird control logic. Someone who's familiar with tuning other brands could easily look at your car, think they know what's going on, but miss some weird twist to the logic. If you watch the os tuning forums for a while, you'll see this happen again and again.

Quote:
What year RS did you notice this very aggressive timing issue on?
Every single RS I've looked at, 05 to 07. I'm not the only one who's noticed this either.

Quote:
What elevation do you live at? I'm at ~600ft.
800ft MSL, so not that different than you

Quote:
Are you looking at fine learned knock correction or knock correction advance?
Initially, I was looking at FBKC, FLKC, and IAM on every single log I did. Used to be that if you didn't monitor all three, you couldn't tell what the engine was really doing. Now, with the addition of Knock Sum, you can watch just Knock Sum and catch every single knock event.

Quote:
What's the ignition total timing in your logs? Mine's at 30 ~4200rpm and climbs upto 35 by 6400rpm.
Hmm... this is the first thing that might mean you actually are more knock resistant than usual. However, without knowing both the engine load and the RPM at those data points, it's hard to compare. On 94 oct, I'm at 10* at 2kRPM and 1.03g/rev, I'm at 19* at 4kRPM and 1.13g/rev, and I'm at 27* at 6.4kRPM and 1.09g/rev.

Note that all of those g/revs are relative to a perfectly scaled MAF. If your MAF scaling is off even a little bit, your reported engine loads mean nothing. As an example, stock MAF scaling, stock intake claims to hit a 1.39g/rev max load. Properly scaling the MAF for the stock intake then shows a max load of 1.14.

Quote:
Could it have been the placebo effect?
Definitely. Well, observer expectancy effect, not placebo effect. I think also that our butt dynos are easily fooled by "Aural Drama". Basically, it's hard to believe it's not fast if it's loud. The cat-back is a total non-starter here. The midpipe and muffler are not the most severe restriction in the stock exhaust setup. Until you do something about the stock cats, and then the stock header, nothing you do with the cat-back matters at all. With the filter, there is a measurable change in airflow. However, the change is so small that it would be impossible to feel. It takes, generally, about a 10% change in power to be definitively perceivable. I haven't read the testing data in a while so I'm a little foggy here, but the flow increase from a K&N is something like 2%.

Quote:
Posting the power mods wasn't really for bragging in any way. It was more so to inform people that this wasn't achieved by just tuning a stock car and that other factors may be affecting the final number.
That's absolutely a critical thing to do. However, from a tuning perspective, you basically have a stock car.

Quote:
I certainly don't want vmware or parallels to crash during a flash.
Just as an aside, I've been running RomRaider/EcuFlash in XP Pro in VMWare Fusion for about a year and a half now and haven't had a single crash. It seems to be stable enough if it's something you're interested in.

Quote:
I should clarify my statement. The stock OL fueling tables had the car aim for stoich until 4,000rpm, at which point it would become overly rich. Even aiming for 11.2:1 AFR at 6,000rpm.
Ah, I see. IIRC, my stock tables tapered away from stoich between 3200 and 3500. Then they tapered to stupid rich land beyond that. The really scary thing is logging a stock car/stock tune with a wideband. It's targeting in the 11s, but the stock MAF scaling is so bad it's actually getting as rich as the 9's. That's just insane.
williaty is offline   Reply With Quote
Old 11-27-2008, 04:06 AM   #13
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

Thanks for the very informative response, it's definitely appreciated especially on an infamous board like nasioc. It appears that I was certainly misinformed regarding the helmholtz resonator. I = schooled :P Regarding the article... I missed the integral paragraph:

Quote:
Normally, a mixture of techniques is used. For example, as shown here, a quarter-wavelength resonator (yellow), a Helmholtz resonator (green) and an expansion chamber (blue).
Would have saved myself some typing if I just read that one small paragraph at the bottom.... :P I also now notice you are the one who wrote the FAQ ...

I'm going to do some logging tomorrow regarding the timing, and log the MAF voltage to double check what's up. I'll post up the log files and my .hex. I'll have a longer response as well, as it's getting pretty late in EST unfortunately. I should have been working on work instead of reading about tuning and knock :P

Quote:
Now, with the addition of Knock Sum, you can watch just Knock Sum and catch every single knock event.
Whoa. Is this a feature in the romraider datalogger? 'Cause it sounds awesome if it's what I think it is. Can't seem to find it.... :S

I also can never seem to find the feed back knock correction to log... Currently (relating to knock) I log ignition total timing, knock correction advance, and learned ignition timing. But I've seen screenshots of logs with feed back knock correction, and I can't find it in the left column. I also have the latest version 0.5.0b of romraider. Needless to say I have more reading to do...

Edit: It appears to be an extended parameter only enabled when plugged into the ECU. I'm definitely going to check this out tomorrow with new logger definition file. I feel as if a whole new world just opened up to me... :P

I'm glad that the vmware fusion works well, I may actually switch over to something similar in the future since that's all I need windows for. I'm a senior graphic design student and only own macs. Although I'm thinking about picking up an asus eeepc 901 just for tuning in the near future. It would be pretty awesome for that dedicated task.

Either way, we have prooved that there is at least some merit to tuning to 05rs. I'm curious how your 2.5rs is setup, as it certainly seems you know what your talking about. You mentioned that it was dyno'd?

Last edited by rectangular; 11-27-2008 at 04:37 AM.
rectangular is offline   Reply With Quote
Old 11-27-2008, 09:50 AM   #14
Matt Monson
Scooby Guru
 
Member#: 832
Join Date: Jan 2000
Chapter/Region: RMIC
Location: Teh Ghetto Garage, CO
Vehicle:
99 2.5RS, '85 911
'73 914 and 2012 BRZ

Default

williaty,
You are a great asset to this forum.
Matt Monson is offline   Reply With Quote
Old 11-27-2008, 10:55 AM   #15
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by rectangular View Post
Thanks for the very informative response, it's definitely appreciated especially on an infamous board like nasioc.
Hey, you weren't asking questions that I'd already answered 5 times this week alone, that's a good way to get me to be civil. Start asking about S.Techs and I'll rip you a new one though

Quote:
Whoa. Is this a feature in the romraider datalogger? 'Cause it sounds awesome if it's what I think it is. Can't seem to find it.... :S
Knock Sum is a brand new parameter, along with a couple of others. Knock Sum is a running count of how many times the knock signal has been set and unset. However, it has a small max value, so you will see it roll over to 0 pretty frequently. The way to track it is to open the log in Excel (or another spreadsheet) and make a new column. Fill that new column with the equation =(KS_column_row_2)-(KS_column_row_1). Then, filter out all the lines where the new column's value is 0 (KS didn't change). The rows you have left (most likely with a value of 1 or -32) will be the rows where the car knocked.

However, if you're going to do a lot of knock analysis, and you should, you'll want to get familiar with MickeyD2005's Knock Counter spreadsheet. It's available in a sticky over on RomRaider.

Quote:
Either way, we have prooved that there is at least some merit to tuning to 05rs. I'm curious how your 2.5rs is setup, as it certainly seems you know what your talking about. You mentioned that it was dyno'd?
Oh definitely. It's worth tuning for the fact that the car feels perkier and smoother. Not to mention the fact that you can keep the engine from destroying itself. You just shouldn't expect to pick up any power on a stock engine.

Yeah, I had it on the dyno Tuesday. The results were not what I expected. The torque fell off at a lower RPM than I had expected and both the torque and RPM were lower than expected. That basically confirms that there's something wrong internal to the engine, which I've suspected for a while now.
williaty is offline   Reply With Quote
Old 11-27-2008, 02:27 PM   #16
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

Alrighty, I have some logs finally.

Wow, I LOVE the extended parameters additions. They pretty much made my day.

The first log I have dips into the supposed MAF v issue area for a couple cells. This was a WOT pull. Probably going to want to right-click-save-as, I left them as csv since I wasn't sure if you would prefer excel or openoffice format.

http://thelocale.org/~geof/car/all-s...127_130058.csv

My knock sum events kept going up one every now and then when a wasn't looking and when I wasn't doing WOT pulls. This seems to be something in CL, perhaps? Just to clarify it ONLY registers on knock sum if it needs to pull from the total timing? I'll need to narrow it down. However, I did log one of the knock sum increases in the feedback knock correction column. It reduced the advance by 1.05*. Not neccesarily a deal breaker.

http://thelocale.org/~geof/car/Knock...127_125906.csv

Correct me if i'm wrong, but the feedback knock correction in this case was the ECU advancing the timing, then a knock event registered so it reduced the amount the ECU was advancing it, since the total timing number is still higher than the base timing. Now this only happened on one of six similar pulls that I did. Although to prevent it in the future, I would be looking at adjusting the timing advance map for those the cells where the engine load and RPM land together and NOT the base timing map, correct? If it happens again in a log I will for sure adjust it.

Either way it sounds as if my car is able to run a good amount more timing than the 05+ rs's that you have dealt with. Probably dumb luck :P

I did a line graph of my MAF v on a pull from 1.3k rpm to redline.



I think I am going to go back out and log only MAFv, engine load, and engine RPM so that I can get more updates per second to get a definite for sure answer since only 4 cells really are in the supposed problem area. All 4 of those cells seem to be perfectly linear and do not match the issue for whatever reason that you noticed in your FAQ. Again, probably dumb luck on my part :P
rectangular is offline   Reply With Quote
Old 11-27-2008, 04:37 PM   #17
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by rectangular View Post
The first log I have dips into the supposed MAF v issue area for a couple cells. This was a WOT pull. Probably going to want to right-click-save-as, I left them as csv since I wasn't sure if you would prefer excel or openoffice format.

http://thelocale.org/~geof/car/all-s...127_130058.csv
Oh my god. That's literally the most unhappy car I've ever seen! Your IAM is 0.375! If you have a normal advance table, that's the car telling you it's pulling as much as 14* of timing!

As far as the MAFv spike, it's hard to prove one way or the other from this log. Do me a favor and run the test this way:
1) Set up the logger to log Engine Speed, MAF Voltage, Throttle Position, and CL/OL Status.
2) Start the car moving on a long flat stretch and get into 3rd gear.
3) Take your foot off the gas completely and let the car slow down to idle, still in 3rd gear (so you'll be lurching along slowly).
4) Start the log
5) Stomp the pedal to the floor as quickly as possible.
6) Let the car accelerate to 3kRPM or so.
7) Stop the log.
8) GOTO Step 3.

Make logs this way and post the results.

Quote:
My knock sum events kept going up one every now and then when a wasn't looking and when I wasn't doing WOT pulls. This seems to be something in CL, perhaps?
No, it's not confined to CL or OL. It just means that, in your case today, the car doesn't knock at WOT.

Quote:
Just to clarify it ONLY registers on knock sum if it needs to pull from the total timing?
No. Knock Sum increments any time the knock signal transitions from unset (no knock) to set (knock). Based on a HUGE number of factors and logic, the ECU will then decide wither to ignore the knock or handle it with IAM, FBKC, or FLKC.

Quote:
Correct me if i'm wrong, but the feedback knock correction in this case was the ECU advancing the timing, then a knock event registered so it reduced the amount the ECU was advancing it, since the total timing number is still higher than the base timing.
Nope, not at all. OK, let's do a bit of definition and then walk through the 2nd log step by step. First of all, the ECU can advance timing in response to not encountering knock and pull timing in response to encountering knock. The crucial thing to understand here is that the ECU can only add timing if it had previously pulled timing. So, let's see how this plays out in your case. I'm going to be a little liberal with explaining when the ECU does something. In other words, I'm going to say it does something when you need to think about it happening, not when the ECU is actually calculating it.

OK, so you turn your car on and the ECU reads the stored IAM to be, in your case, 0.375. Right away, before you even put the car in gear, that causes the ECU to pull a HUGE amount of timing. The ECU multiplies the IAM by the values in the Maximum Advance table (or whatever they're calling it this release ). So lets pretend your Advance table was full of 10s. It's not, but 10s make the math easy. 10*0.375=3.75 degrees. So your ECU just pulled 6.25* of timing without you even moving the car. Now, like I said, your real Advance table is full of all kinds of numbers, some 0, some small, some larger than 10, but the thing to remember is that ALL of those get multiplied by IAM so they all (other than the 0s) just had a huge amount of timing taken from them. Now you put the car in gear and start driving. The ECU looks at the Advance table for that RPM and load and picks out the correct value (which, remember, has been savaged by your low IAM). It then looks at the Fine Learning Knock Correction Table (which is something the ECU creates on its own), finds the correct cell for the current RPM and load, and adds that value to the Advance*IAM value. Then it adds in any FKBC (we're pretending there is none here). Then it looks at the Base Timing table and finds the correct value for the current RPM and load. The Base value gets added to the running total. Then it applies any other timing compensations (which we're ignoring) to the running total. Then it uses that timing.

Now, lets repeat this cycle thousands of times a minute and we've got a running engine. Then knock happens. The ECU detects it, sets the knock signal, the knock stops (micro- or milli- seconds later) and the knock signal is unset. This process caused the Knock Sum to increment by 1. The ECU has 3 possible ways of dealing with this knock event: IAM, FBKC, and FLKC. It's important to remember that the ECU is only evaluating one of these methods at a time, even though the effects of an earlier evaluation may still be influencing the current timing. What do I mean? If the ECU is evaluating FLKC, it's still using the current IAM value. Same goes for FKBC, etc. If the ECU is in IAM evaluation mode, a knock event causes IAM to decrease, thereby pulling timing globally. Likewise, if it's in IAM evaluation mode and it doesn't detect knock, it increases IAM. In FLKC evaluation mode, it pulls or adds timing in the current FLKC Table cell in response to the presence or absence of knock. Note that if IAM is 1, FLKC can never be greater than 0. This is a side effect of the fact that Advance*IAM+FLKC isn't allowed to be greater than Advance. When the ECU is in FBKC evaluation mode, it simply pulls timing in response to knock and then lets that pull "fade away" if the knock goes away.

OK, looking at your log:
1447 RPM: We're in a cell in the FLKC Table where FLKC is 0
1590 RPM: We moved into a FLKC Table cell where FLKC=1.05 OR the absence of knock allowed the ECU to increase the value in the current cell of the table. Without knowing a lot more about some of the values in your rom, I can't tell you which happened here.
2133 RPM: We moved into a FLKC Table cell where FLKC=2.1 OR the absence of knock allowed the ECU to increase the value in the current cell of the table. Without knowing a lot more about some of the values in your rom, I can't tell you which happened here.
2406 RPM: Same as above
2988 RPM: We moved into a FLKC Table cell where FLKC=0.35. You can tell the cell changed because there was a downwards movement without KS changing.
3819 RPM We moved into a FLKC Table cell where FLKC=0.7 OR the absence of knock allowed the ECU to increase the value in the current cell of the table. Without knowing a lot more about some of the values in your rom, I can't tell you which happened here.
5842 RPM: The engine knocked. The ECU was in FBKC evaluation mode at this time and pulled 1.05* of timing in response to the knock event.
6130 RPM: The ECU is still pulling 1.05* of timing via FBKC in response to the previous knock event. The engine has not knocked again.
6314 RPM: You either missed the "fade" of FBKC due to how far apart your samples are or you crossed the FKBC application threshold for RPM. Can't tell which without your ROM.

Quote:
I would be looking at adjusting the timing advance map for those the cells where the engine load and RPM land together and NOT the base timing map, correct?
I personally leave the Advance table alone and adjust the Base table only. Some people do it the other way. Both work. I figure the engineers had a lot more time to figure out the correct shape for the Advance table than I do, so I'll trust their judgment. The shape of the Advance table is a map of how knock-prone the engine is, so it's an important thing to have right. However, with how frighteningly low your IAM is, you need to pull a crap-ton of timing out of your Base table to begin with.

Quote:
Either way it sounds as if my car is able to run a good amount more timing than the 05+ rs's that you have dealt with. Probably dumb luck :P
Well, your engine is very unhappy about the amount of timing you're trying to make it run. Once you get it re-tuned for IAM=1 and only small (around 1*) FLKC in all cells, then we'll see
williaty is offline   Reply With Quote
Old 11-27-2008, 05:01 PM   #18
RaceFaceXC
NASIOC Supporter
 
Member#: 165749
Join Date: Dec 2007
Chapter/Region: South East
Location: stupid Florida
Vehicle:
2007 No YELLING
on the bus!

Default

Williaty, that was the best explanation of the Denso ECU knock control strategy i have read so far. much easier to understand and follow the string of steps than anything i have read in the RR forums.
RaceFaceXC is online now   Reply With Quote
Old 11-27-2008, 05:46 PM   #19
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by RaceFaceXC View Post
Williaty, that was the best explanation of the Denso ECU knock control strategy i have read so far. much easier to understand and follow the string of steps than anything i have read in the RR forums.
It DOES skip or gloss over some important things though. It's definitely NOT complete enough to take as the final word on the subject. The knock control sticky needs to be read too.
williaty is offline   Reply With Quote
Old 11-27-2008, 06:59 PM   #20
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

Thanks for the awesome explanation. I've been reading through this post a couple times as well to make sure I have everything figured out in my head correctly.

http://www.romraider.com/forum/topic1840.html

A must read for anyone tuning with romraider for sure.

Quote:
Total Timing = Base Timing + Knock Correction Advance + Other timing compensations

Other timing compensations = other compensations including those for IAT, ECT, per cylinder, among others.

Knock Correction Advance = (Timing Advance Maximum * (IAM/16)) + feedback knock correction + fine learning knock correction

Note: The IAM (ignition advance multiplier) used in the formula above is the raw value (ranging from 0-16) for the 16-bit ECU. For the 32-bit ECU, the IAM will range from 0 to 1 (substitute "IAM" for "IAM/16").
Correct me if i'm wrong, but since I am running a 32-bit ECU the equation would be thus

Knock Correction Advance = (Timing Advance Maximum * (IAM)) + feedback knock correction + fine learning knock correction

So then the big picture would be:

Total Timing = Base Timing + ((Timing Advance Maximum * (IAM)) + feedback knock correction + fine learning knock correction)

So here's the way I see it. The IAM is a way for the car to scale the Timing Advance Maxium table without actually changing the values in the table. Like you said, it's made by engineers that probably know quite a bit about the engine, more so than us or the ECU for that matter. Thus having a non changing maximum timing advance table would be a great idea. (and i've dually noted not to change it).

If the base timing is super conservative, it will increase the IAM so that it can increase the Timing Advance. If the base timing map is more aggressive, it doesn't need the timing advance numbers to be as large, so it reduces the IAM to scale down the Timing Advance Maximum table.

I suppose then it is possible to have 2 cars, one with the conservative base timing map and one with an aggressive timing map and they would end up with the same total timing since the ECU can scale the amount they are advanced.

So let me give 2 examples of what I am saying to see if I am correct in my understanding.

With a conservative base map timing (note: completely made up numbers):

The equation again:

Quote:
Total Timing = (Base Timing) + ((Timing Advance Maximum * (IAM)) + feedback knock correction + fine learning knock correction)
Let's say this is a conservative base map number, with a maxium advance table of all 10s again, and that the ECU did not detect any knock. The ecu notices that this is a conservative base map through driving and attempts to make it more efficient by having the IAM at it's maxium of 1.

Total timing: 35 = 25 + ( 10 * 1.0 ) + 0.0 + 0.0

Let's say this next example is similar to my car where I increased the base timing map to be more aggressive, and that no knock events were recorded. The ECU will see this and then reduce the IAM to effectively scale down the Timing Advance Maximum table.

Total timing: 35 = 30 + ( 10 * 0.5 ) + 0.0 + 0.0

In this situation both cars ended up at the same total timing number.

The problem it seems with the IAM being set to straight 1 is that the ECU then looses it's ability to scale the advance up. It can only reduce the IAM and only scale the Timing Advance Maximum map down. This is good if you don't trust the ECU to make the timing more agressive. This is bad if you DO trust the ECU to try and make the car run more efficiently. At least how I see it with my current understanding.

My question then is, why is there an advantage of having the IAM end up being 1? It seems that in this case when the IAM is 1, the base timing map is so conservative that the car is trying to advance it the most it possibly can from the Timing Advance Maxium table. It seems to me that the ideal place would be around an IAM of .50 so that there is equal room for it to scale the timing advance map in either direction.

Let's use the previous example again, and say that a Total Timing of 35 is the most efficient timing for the engine to run at in this situation.

Again, the conservative base map first, but this time the conservative base map is SO conservative that even with the IAM at its maximum of 1 it cannot reach the total timing number of 35 that would be the most efficient. Again, no knock events were recorded in either.

Total timing: 25 = 15 + ( 10 * 1.0 ) + 0.0 + 0.0

However, in the next example again, with the more aggressive base map the car's IAM is not 1.0, but instead more in the middle with 0.5 and IS able to reach the most efficient total timing number of 35.

Total timing: 35 = 30 + ( 10 * 0.5 ) + 0.0 + 0.0

So, how does having an IAM of 1 benefit the car? It would seem that having an IAM of .38 does NOT make my car an unhappy car, but rather a car that notices the base map changes and has adjusted the IAM according to scale down the timing advance map as it does not require advancing the map as much.

IAM is still important, but I think more importantly is to monitor Knock Sum, and make sure that the total timing is still more than the base timing and keep an eye on FLKC and FBKC. Since on only 1 out of the 6 runs my car reduced the timing advance because it detected a knock event I think backs up my theory.

Quote:
It's not, but 10s make the math easy. 10*0.375=3.75 degrees. So your ECU just pulled 6.25* of timing without you even moving the car.
My understanding is that it didn't necessarily pull timing since it wasn't in regards to a knock event, instead it reduces the timing advance because of the stored IAM it knows that it is not necessary for the maximum timing advance value from the map.

Also, I'll definitely log later according to your specifications regarding the MAF voltage issue, I'm curious what the result is.

In the end, I guess what's the next step. As from my current conclusion I don't see any benefit from a max IAM stored setting at 1, and possibly not allowing the ECU to advance the total timing to it's most efficient number.

Last edited by rectangular; 11-27-2008 at 07:08 PM.
rectangular is offline   Reply With Quote
Old 11-27-2008, 07:16 PM   #21
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by rectangular View Post
Knock Correction Advance = (Timing Advance Maximum * (IAM)) + feedback knock correction + fine learning knock correction

So then the big picture would be:

Total Timing = Base Timing + ((Timing Advance Maximum * (IAM)) + feedback knock correction + fine learning knock correction)
Yes, that's exactly what I was saying.


Quote:
If the base timing is super conservative, it will increase the IAM so that it can increase the Timing Advance. If the base timing map is more aggressive, it doesn't need the timing advance numbers to be as large, so it reduces the IAM to scale down the Timing Advance Maximum table.
In a way this is true. However, it misses some important considerations. A car running near IAM=1 with minimal FLKC will have two big advantages. First, it won't knock much to "settle" into its final timing state. Second, it'll have the most "safety margin" to deal with a bad tank of gas, altitude change, heat soak, mechanical problem, etc.

Another thing to consider is that when the ECU enters IAM evaluation mode, it instantly sets IAM to the Initial IAM Value from the map. In your case, that means if the knock gets so severe that it forces the ECU back into IAM eval mode, the first thing that's going to happen is that the timing is going to advance considerably. That's instant grenade land.

Quote:
I suppose then it is possible to have 2 cars, one with the conservative base timing map and one with an aggressive timing map and they would end up with the same total timing since the ECU can scale the amount they are advanced.
Yes, that's true. You can get to the same Total Timing via a variety of different combinations. However, IMO, the one with the higher IAM and less-negative FLKC will be easier to tune and safer for the engine.

Quote:
The problem it seems with the IAM being set to straight 1 is that the ECU then looses it's ability to scale the advance up. It can only reduce the IAM and only scale the Timing Advance Maximum map down.
This is true. However, there's about 100 factors that can make the car knock more and about 10 factors that can make the car knock less. It's more important to have the greatest ability to reduce timing to protect the engine than it is to have it able to advance the timing a little if a miracle happens.

Quote:
IAM is still important, but I think more importantly is to monitor Knock Sum, and make sure that the total timing is still more than the base timing and keep an eye on FLKC and FBKC. Since on only 1 out of the 6 runs my car reduced the timing advance because it detected a knock event I think backs up my theory.
Your car is currently in a learned state right now, minimizing the "damage" of the too-aggressive tune. If you reset the ECU or have a large change in environmental conditions, the car will knock like a mofo for a while before settling down again.

Additionally, you have the HUGE advantage of having one of the only ECUs that doesn't reset every time you flash it. Almost every other car out there will totally reset the ECU every time you flash the car, subjecting the engine to a period of intense knocking before things settle down.



Quote:
My understanding is that it didn't necessarily pull timing since it wasn't in regards to a knock event, instead it reduces the timing advance because of the stored IAM it knows that it is not necessary for the maximum timing advance value from the map.
Unless the ECU is in IAM eval mode, the IAM is not representative of current knock it's representative of past knock. If your IAM!=1, then at some time in the past, the engine had to have knocked to push the IAM down. Same thing with negative FLKC values.

Quote:
As from my current conclusion I don't see any benefit from a max IAM stored setting at 1, and possibly not allowing the ECU to advance the total timing to it's most efficient number.
Again, I think you should reconsider as it's hard on your engine until everything settles down. IAM=1 is not mandatory, but an IAM close to 1 is a very good goal to shoot for. Also, because of the coarseness of the FLKC table, if you want to strive for maximum advance at all times, you actually can do a FAR better job than the ECU can.
williaty is offline   Reply With Quote
Old 11-28-2008, 12:42 AM   #22
rectangular
Scooby Newbie
 
Member#: 132839
Join Date: Nov 2006
Chapter/Region: RMIC
Location: Salt Lake City
Vehicle:
2009 WRX
Red

Default

Awesome, thanks for shedding some clarification and adding to my conclusions.

I guess the next question I have is, what is the next step? Should I bring the stock timing map back in and see how the IAM reacts and aim for an IAM of just above .50 so that it has a fair amount of room either direction? It makes sense to have a slightly higher IAM in case of bad gas, altitude change, or heat as you mentioned. These logs were done with 91 octane gas. Is there any advantage to adjusting the timing since the ECU can basically on the fly determine how much it wants to advance it? I suppose if you are maxing out the IAM and the Maximum Timing Advance numbers, then the base map should be increased.

Since the initial IAM is .5 on the stock map (and I haven't touched it). It seems like a good goal. At .38 I'm not far off. Therefore, with a good no-knock timing setup at .5, if the ECU get's reset, it will be safe, and there's plenty of room for expansion for addition or subtraction.

I was browsing around an 09 wrx ROM the other day posted on romraider forums since I eventually plan on getting an 09 and doing some tuning with it, I noticed this entry called Boost Control Disable (IAM). The values disable boost control at less than .2 IAM and will re-enable it at .65 IAM. I think this gives good insight on helping flesh out what subaru thinks are good and bad IAMs.

There is also a value to switch to the failsafe fuel map on the 09 wrx when the IAM drops below .35. I think this definitely motivates me to try and get my IAM up.

In regards to the resonate issue. I was able to reproduce something similar to your chart in the FAQ on one of the three runs. The other two times I couldn't get the car to accelerate through the revs, it would just stay put and the MAF v wouldn't move much at all from where it should have been.

The one time I was able to reproduce it was on a very very gradual downhill, I could feel the car hesitate slightly between 1,200 and 1,350rpm. I don't remember feeling that in the past and new right away I most likely reproduced the resonation you were talking about. Here's a link to the run with a chart in it:

http://thelocale.org/~geof/car/some-resonation.xls

I personally decided that since it's already taken out, I probably won't put the helmholtz resonation "gourd" back in unless I decide to sell or trade in the car. There was absolutely no power there anyways, and I can't remember the last time I was in that rev range at that high of a gear for that long at WOT. In my case it only really occured between 1000rpm and 1400rpm according to my interpretation of the log.

It will certainly be something I consider as far as my driving style is in the near future. Usually when I am in those rev ranges I am in CL. But not enough of an issue for me to worry about at this point.

I am interested in what the AFR values are in the end when that issue occurs. Have you done this test with a wideband?

Dammit, I really need to start actually working on my final graphic design project instead of reading about subaru ecu tuning :P
rectangular is offline   Reply With Quote
Old 11-28-2008, 01:20 AM   #23
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Quote:
Originally Posted by rectangular View Post
I guess the next question I have is, what is the next step?
Well, I'm not sure what you've done already, but the overall workflow is something like:
1) Set up misc. things the way you want them (CL/OL, rev limit, etc)
2) Get CL MAF scaling dead-nuts accurate
3) Get OL MAF scaling dead-nuts accurate (requires wideband. Yes, you have to buy a wideband if you want to tune).
4) Get OL Fueling exactly where you want it
5) Give it all the timing it'll stand.

Quote:
Should I bring the stock timing map back in and see how the IAM reacts and aim for an IAM of just above .50 so that it has a fair amount of room either direction? It makes sense to have a slightly higher IAM in case of bad gas, altitude change, or heat as you mentioned.
I definitely think you need to make the engine happy enough to get the IAM up. A lot. I don't think .5 is high enough. In fact, I don't know anyone who would accept an IAM as low as 0.5.

The quickest way to get in the ballpark is to copy your Base Timing and Timing Advance tables into a spreadsheet. Create a new "table" in the spreadsheet so that each cell=(Advance-Advance*IAM). Basically, you just created a new table that shows how much timing is being pulled because of IAM. Create another new table so that each cell=Base-NewTable1. In other words, if Base is 50, IAM is 0.375, and Advance is 10, the first new table gets 6.25 (Advance-Advance*IAM), then the second new table gets 43.75 (Base-NewTable1). That second table becomes your new Base Timing table, copy it back into RomRaider. Flash it to the car. Reset the ECU following the proper procedure. Then immediately go out raise the IAM using the procedure I've detailed on this forum previously.

What that does, effectively, is to set the Base table to values such that Base+Advance*IAM with an IAM of 1 gets you the same timing values you're using anyway. Using this method, you'll get close to your final timing values pretty quickly. Since IAM is more of a chainsaw than a scalpel, you'll find you still have a few places that knock a little bit and a lot of places where you can add timing to the Base map. However, it's the quickest way I've come up with for getting in the ballpark. After this point, you can tune the rest of the Base Timing map using the Knock Counter spreadsheet (which is a godsend).

Quote:
Is there any advantage to adjusting the timing since the ECU can basically on the fly determine how much it wants to advance it?
Yes, you prevent all the knock that would otherwise happen while the ECU is trying to sort out the timing to begin with. Plus, since the ECU is constantly trying to advance timing, it's constantly going to be triggering new knock as it does so.

Quote:
Since the initial IAM is .5 on the stock map (and I haven't touched it).
Initial IAM is 0.5 because it's very unlikely that this represents enough timing advance to grenade the engine on the first try. IAM eval mode immediately tries to push this number up. From the factory, the IAM is trying to get to 1 but just can't.

Quote:
It seems like a good goal. At .38 I'm not far off. Therefore, with a good no-knock timing setup at .5, if the ECU get's reset, it will be safe, and there's plenty of room for expansion for addition or subtraction.
Except that's missing the point. You're not equally likely to want to add and subtract timing (other than during tuning itself). IIRC, 16* is the largest value in the Advance map on an 05. If you're at IAM=.5, you're running 8*. It's pretty easy to encounter a situation in which the ECU would want to pull 10* of timing but, uh oh, it can't. It's basically impossible that you'd ever encounter a situation where the ECU would want to add 8*, let alone more than 8*. Add to this the fact that not being able to retard timing enough blows the motor and not being able to advance timing enough just costs you gas mileage and you definitely want more downwards freedom than upwards.

I just thought of another issue, really. If you're working at a low IAM, you need to have a VERY coarse FLKC table, like the stock ROM does. If you run a higher IAM, it's safe to alter the FLKC table rows and columns to give the ECU much more fine-grained control over timing. This would result in you actually running more total timing more of the time in the real world.

Quote:
I think this gives good insight on helping flesh out what subaru thinks are good and bad IAMs.
For a turbo car. The turbo tunes knock much, much less than the NA tunes do from the factory. There's a huge difference between the theory for tuning NA and turbos when it comes to safe handling of knock.

Quote:
I am interested in what the AFR values are in the end when that issue occurs. Have you done this test with a wideband?
9 AFR if the car had a stock tune that caused it to still be in CL there. 3.3 to 4 AFR if the car was in OL at that point.

That made me think of something else though. What AFR are you targeting at WOT in OL? That alone could be the reason you're running more timing than I am. I run as lean as I can get away with in OL to make the most power from the least gas.
williaty is offline   Reply With Quote
Old 11-28-2008, 01:48 PM   #24
sc00ter
Scooby Guru
 
Member#: 117767
Join Date: Jun 2006
Chapter/Region: RMIC
Location: Denver, CO
Vehicle:
2011 STI Sedan
DGM

Default

Wow, now I am extremely interested in getting my car to a dyno. I've had a custom enginuity tune on my car for a while now, thanks to Williaty and Master2192.

Current mods, which the car was tuned for:

1) Injen CAI
2) 2.5" custom cat back with mendrel bends
3) Stromung HFC
4) Perrin LW pulley
5) Borla replica UELs

I figured I was putting down about 120-125ish WHP but maybe I've been underestimating.
sc00ter is offline   Reply With Quote
Old 11-28-2008, 01:54 PM   #25
DFL2.5RS
Scooby Specialist
 
Member#: 73228
Join Date: Oct 2004
Chapter/Region: NESIC
Location: Maine
Vehicle:
2004 WRX wagon WRB
2006 Rio Yellow S2000

Default

I'd definitely aim for an IAM of as close to 1 as possible... My car is a 16bit ecu so I always aim for an IAM of 16... When I tune I try to push the timing as far as I can without allowing the IAM to drop below 16.... The way I see it, this allows the car to make as much power as possible, consistently, without frequently encountering knock events which have the potential to damage the engine... If I see my IAM drop below 16 in a log, I alter the involved cells of the timing map....

If you're happy with the power the car is making while you're tuning it there is no point in leaving headroom for the ecu to advance timing beyond where you have it in search of an extra 1-2bhp when the potential cost of that little timing exploration is a new shortblock... besides that, everytime your car encounters a knock event and the IAM decreases you're likely losing more power for the period the IAM is decreased than the car ever gained by allowing the car to advance timing beyond where it was when you were tuning...

It's probably also important to mention that as you approach knock threshold increases in timing don't net you much more power... the biggest gains from timing advance occur when the timing is far more retarded than it could be... you might think of power as asymptotically approaching a maximum value as timing is advanced... and as you approach the limit, changes in timing net very small amounts of power and place you at greatly increased risk of knock...
DFL2.5RS is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -4. The time now is 07:22 PM.


Powered by vBulletin® Version 3.7.0
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Powered by Searchlight © 2014 Axivo Inc.
Copyright ©1999 - 2014, North American Subaru Impreza Owners Club, Inc.