Tuning for Fuel Economy

[covertrussian]

It's been a while since last update, I've been having some issues with the fueling (mainly fuel pump assembly). So two weeks ago I swapped in the stock 100k mile fuel pump and which got my WOT happy, but cruising learning was still around +10%. Adding the +10% MAF scale, got the cruising down to +1-3% learning but idle is now at +15% (I think the 100k fuel filter is getting clogged).

 

Anyway with that setup (100k fuel pump/filter and +10% maf scale, AVCS 5 and 40* ign highway map) went for a highway trip on a pretty nice overall day and got only 29.56mpg.

 

Last 4 trips to this place, with the AVCS 5 and IGN 40* map has been right under 30mpg. I went to this map because I got 31.25mpg 5 trips ago, so maybe that was a lucky fluke and I just need to go back to the AVCS 10* and 38* IGN hwy map.

 

Also I'm getting to the point where I'm just dancing around the 30mpg mark, but not really getting much further. I think it's time I start messing with the intake side, since exhaust side is not really showing any benefits (still bummed at the Equal length header not getting me any gains).

 

Reason for starting to mess with intake side is, I've been getting the same peak MPG as the VF46 turbo. Thus I think something else is holding it back, thus these changes didn't help:

• Bigger turbo didn't help, actually lost MPG until timing reduced but that only got me back to VF46 MPG and not higher.
  
• Taller Tires - Reduced the RPM, didn't seem to help
  
• 5lbs Lighter WRX Flywheel/Clutch
  
• Tried making the exhaust happier with an Equal length Header
  
• Different AVCS angles
  
• Different timing ranges
  

 

Plans for the intake side:

- 3" Hard tube inlet (currently have the STI Perrin one).

- 3" Intake with Flow Stack

- TGV Deletes

 

I might try AVCS 10* one last time next week, after that I think I'll move on to the intake side.

Edited June 26, 2017 by covertrussian

[lagacyforce]

You may want to take a look at this story too, discussing 5 ways Subaru improves fuel mileage on Impreza.

 

https://www.torquenews.com/1084/5-ways-subaru-improves-fuel-mileage-new-2015-impreza

[covertrussian]

You may want to take a look at this story too, discussing 5 ways Subaru improves fuel mileage on Impreza.

 

https://www.torquenews.com/1084/5-ways-subaru-improves-fuel-mileage-new-2015-impreza

 

Sadly none of those really apply to us, we don't have CVT, PZEV actually hurts fuel economy (dumps lots of fuel on cold starts). We can't really reduce internal friction (and using 0w-20 water is dangerous for our high power motors).

 

This is actually why I started this thread, to find realistic, non-hypermiling, gas mileage gains for our cars . My car is already getting 30mpg, which is what the 2.5i's get, but I like to set my goals high

[utc_pyro]

Is there really much potential for gas mileage savings on the intake side? The throttle restricts intake flow for the sake of setting the given output power. Thus if you try and reduce plumbing losses you'd have to reintroducing them at the throttle to keep the same power level. Also the TGV's create turbulence and thus improve fuel mixing, I'd suspect delaying those might have a slight negative impact on economy. On the newer cars they actually run the TGV's full time below ~40g/sec for emissions reason, but probably also helps economy.

 

If you really want to reduce pumping losses on the intake side we have a few tricks in the Aviation world that would be more helpful. Fuel burn is range for us, and the leaded crap we use is $5-6/gallon so saving every drop is important.

 

First in flight we can open up the carburetor heat (secondary air intake that goes around the exhaust) to reduce intake air density and thus power output. Then we can open up the throttle more to compensate reducing pumping losses. A warm air intake ( maybe selectively so?) would have the same function on our cars.

 

Second is on injectred motors is to put it into a fairly extreme lean burn state. At cruse you set the AFR manually so the EGT is ~50-100f beyond peak on the lean side. This significantly reduces power, and you again open the throttle more to compensate. To do this effectively you need to monitor EGT and CHT (cylinder head temperature) closely and for each cylinder to make sure things aren't overheating. These are ancient air cooled pushrod motors with mechanical fuel injections, it's be a lot harder to roast one of our water cooled motors with sodium filled exhaust valves.

[Flinkly]

Is there really much potential for gas mileage savings on the intake side? The throttle restricts intake flow for the sake of setting the given output power. Thus if you try and reduce plumbing losses you'd have to reintroducing them at the throttle to keep the same power level. Also the TGV's create turbulence and thus improve fuel mixing, I'd suspect delaying those might have a slight negative impact on economy. On the newer cars they actually run the TGV's full time below ~40g/sec for emissions reason, but probably also helps economy.

 

If you really want to reduce pumping losses on the intake side we have a few tricks in the Aviation world that would be more helpful. Fuel burn is range for us, and the leaded crap we use is $5-6/gallon so saving every drop is important.

 

First in flight we can open up the carburetor heat (secondary air intake that goes around the exhaust) to reduce intake air density and thus power output. Then we can open up the throttle more to compensate reducing pumping losses. A warm air intake ( maybe selectively so?) would have the same function on our cars.

 

I've always considered a WAI for the OBXT, but for an afternoon where i just wanted to zoom around, always seemed a little detrimental (probably not enough to really care about, honestly). but being able to switch between CAI (or OEM intake) and WAI... that would be slick. thinking an actuated flapper like some do for exhausts to switch between straight pipe to atmosphere and regular muffler. I'd also never thought about why a WAI is better, but it makes sense that you lower the density and so have to increase the throttle opening, reducing restriction...

 

which also makes me think about the lofty goal of installing an alternate exhaust route to heat the coolant from a cold start for reduced warm-up time (like the newer chevy malibu).

 

lastly, as for the TGV, i've always wanted to know how much they do on the newer engines for MPG. I've got "deletes", but have been hesistant to install for any tiny gain they'd provide. and i really only want to delete them for the 2x ECU inputs for more sensors.

[covertrussian]

Is there really much potential for gas mileage savings on the intake side? The throttle restricts intake flow for the sake of setting the given output power. Thus if you try and reduce plumbing losses you'd have to reintroducing them at the throttle to keep the same power level. Also the TGV's create turbulence and thus improve fuel mixing, I'd suspect delaying those might have a slight negative impact on economy. On the newer cars they actually run the TGV's full time below ~40g/sec for emissions reason, but probably also helps economy.

 

You know, this one is hard to say. A couple years ago when I built my DIY intake, I installed it and adjusted my MAF scales (and that's it since I used OEM maf housing). I gained about .75 of an MPG, but I dismissed it because a weeks back I got got a little better. But I wasn't keeping track of wind data at that time. Looking at weather underground's historical data, I saw that it was 2x as windy when I tested the intake. Thus I'm at a point where I want to test it out all over again.

 

On my car, as you know, TGV's are only working during a cold start, but then are open all the time. If my car, with TGV's open all the time, can get better gas mileage then 07+ which has TGV's closed during cruising, it makes me really question the efficacy of them. That's what actually really gets me about 07+ LGT's, they have higher compression, variable TGV's, different tuning/WGDC structures, or Spec B's with super low reving 6 speeds (at 70mph), yet they still get the same gas mileage as 05-06's.

 

Since I'm running out of things to test tune wise, I think the next step would be to let her breathe easier. It really worked well on my G20, which has a 3" exhaust, higher flowing intake manifold, 3" turbo inlet/wai. After I learned to tune it right (for all the airflow modifications) she started getting 35+mpg and still will smoke tires in first gear.

 

If you really want to reduce pumping losses on the intake side we have a few tricks in the Aviation world that would be more helpful. Fuel burn is range for us, and the leaded crap we use is $5-6/gallon so saving every drop is important.

 

First in flight we can open up the carburetor heat (secondary air intake that goes around the exhaust) to reduce intake air density and thus power output. Then we can open up the throttle more to compensate reducing pumping losses. A warm air intake ( maybe selectively so?) would have the same function on our cars.

 

Second is on injectred motors is to put it into a fairly extreme lean burn state. At cruse you set the AFR manually so the EGT is ~50-100f beyond peak on the lean side. This significantly reduces power, and you again open the throttle more to compensate. To do this effectively you need to monitor EGT and CHT (cylinder head temperature) closely and for each cylinder to make sure things aren't overheating. These are ancient air cooled pushrod motors with mechanical fuel injections, it's be a lot harder to roast one of our water cooled motors with sodium filled exhaust valves.

 

That's pretty cool stuff, I appreciate the aviation engines and pilots a lot. A friend of mine is a pilot too, we took his plane out once and he was telling me how he was adjusting the AFR's in real time. I would be a little more scared of doing car tricks on plane engines, where engine stalling is far more devastating.

 

About the carburetor heat, from what I remember reading that's how the 300mpg carburetor came about, it was evaporating the fuel before hand. It's hard to say if fuel injected motors can benefit from the same. Our returning fuel systems technically heat up the fuel, while returnless systems don't (though my returnless 2.5i Outback gets 33mpg pretty often).

 

Problem with leaning out the engine is, you quickly lose the gas mileage gains as soon as you start going up a hill. This is why I haven't really bothered with going leaner. Plus with E10, stoich is actually lower, I believe 14.1 AFR, which means ECU keeping the engine at 14.7 is already leaning it out.

 

I've always considered a WAI for the OBXT, but for an afternoon where i just wanted to zoom around, always seemed a little detrimental (probably not enough to really care about, honestly). but being able to switch between CAI (or OEM intake) and WAI... that would be slick. thinking an actuated flapper like some do for exhausts to switch between straight pipe to atmosphere and regular muffler. I'd also never thought about why a WAI is better, but it makes sense that you lower the density and so have to increase the throttle opening, reducing restriction...

 

which also makes me think about the lofty goal of installing an alternate exhaust route to heat the coolant from a cold start for reduced warm-up time (like the newer chevy malibu).

 

The idea behind WAI is actually two fold...

1. Cold air is denser, thus contains more oxygen, thus needs more fuel to stay at 14.7 cruising. Warm Air Intake sucks in hot air, which is less dense and contains less oxygen thus in theory uses less fuel to maintain the car at 14.7afr. This also results in torque loss, which leads us to the next bullet point...

2. Because we have less torque, we now need to open the throttle more to maintain the same speed, this translates to reduction in manifold vacuum, which translates to reduced pumping losses.

 

lastly, as for the TGV, i've always wanted to know how much they do on the newer engines for MPG. I've got "deletes", but have been hesistant to install for any tiny gain they'd provide. and i really only want to delete them for the 2x ECU inputs for more sensors.

 

I'm TGV's because I spent two weeks of sweat porting the damn things, so now I gotta run them . But in all seriousness, they provide about 150CFM of flow (from 330 cfm to 480 cfm). That's a pretty darn good bump in airflow for power (I like power gains as much as I like gas mileage gains ).

 

I'm also curious on how much power I could gain on my stock boost setup basically.

[solidxsnake]

Problem with leaning out the engine is, you quickly lose the gas mileage gains as soon as you start going up a hill. This is why I haven't really bothered with going leaner. Plus with E10, stoich is actually lower, I believe 14.1 AFR, which means ECU keeping the engine at 14.7 is already leaning it out.

 

Just a quick correction... the ECU isn't keeping the engine at 14.7 AFR, it is in fact targeting lambda = 1.0. Remember, the O2 sensors in the car are lambda sensors, meaning they are measuring lambda, not AFR, regardless of the fuel-type used. The ECU only sees and calculates fueling using lambda. The numbers you see in your fueling tables and logger is scaled to gasoline, but what's actually in the ECU is lambda. You could just as easily scale the values to report AFR for E10, (or E85, etc.), but in the end the value being used is lambda.

[covertrussian]

Just a quick correction... the ECU isn't keeping the engine at 14.7 AFR, it is in fact targeting lambda = 1.0. Remember, the O2 sensors in the car are lambda sensors, meaning they are measuring lambda, not AFR, regardless of the fuel-type used. The ECU only sees and calculates fueling using lambda. The numbers you see in your fueling tables and logger is scaled to gasoline, but what's actually in the ECU is lambda. You could just as easily scale the values to report AFR for E10, (or E85, etc.), but in the end the value being used is lambda.

 

I would still stick with ECU is adjusting to AFR's based on it's mA(milliamp?) to AFR scales, otherwise it would automatically keep you at 14.1 AFR when you ran E10, or 9.0 on E85. With stock ECU and wideband, I think the display unit itself is irrelevant. It's like saying ECU doesn't adjust your boost based on PSI, it adjusts it to KPA, when in reality both measure the same exact thing, just display it differently.

EDIT: This paragraph is based on wrong assumptions of how ECU reads air to fuel ratios. solidxsnake and Flinkly point out in future posts that ECU does indeed use Lambda and it's just Romraider and ECUFlash configs that convert it to AFR for us, ECU does not use those AFR numbers.

 

Now while mauling and thinking about your post I did come across some interesting things....

- Stock WBO scaling is not Voltage based like most WBO's are, it's mA based, which if I had to venture a guess is milliamp based?

 

This thread has some really good info and math related to E10...

- E10 stoichiometric ratio is indeed 14.13 AFR aka 1.00 lambda. That means our ECU targeting 14.7 on E10 is indeed always running lean, 1.05 lambda lean. That would be the same as running 15.44 AFR on 100% gas (I made a spreadsheet for this...) EDIT: This point would be correct if ECU adjusted based on AFR's, but it adjusts based on Lambda. Thus we are no leaner from stoich then we would be on E0.

 

This brings the next slew of questions....

- Could this be the reason why E10 gas mileage loses are actually not that high? All pre 2008 cars are effectively at lean burn mode now?

- Are we killing our catalytic converters because we are always feeding them these lean mixtures?

EDIT: These point would be correct if ECU adjusted based on AFR's, but it adjusts based on Lambda. Thus we are no leaner from stoich then we would be on E0.

 

- As a generally excepted rule for EJ255/7 we tune our engines to 11.0:1 AFR for safety, did this rule come about pre or post E10 revolution?

 

E0 11.0 AFR = 0.75 Lambda

E10 11.0 AFR = 0.78 Lambda

E10 0.75 Lambda = 10.60 AFR

 

Basically this means that we need to run 10.60 AFR with E10 to get 0.75 lambda (11.0 AFR with E0 aka 100% gas).

 

When we run 11.0 AFR with E10, we would be running the equivalent of 11.47 AFR on E0. 11.47AFR is still considered safe in most boosted applications, but as we know Subaru's are more sensitive so it might not be so safe...

 

This is why I'm asking about the 11.0 rule, if it came after ethanol then we are still in the safety zone, but if it came after then we might need to run richer.

Edited June 28, 2017 by covertrussian

[Flinkly]

it is milliamp. ECUFlash shows it un-edited in raw ECU value, which looks to be lambda. Romraider converts it to AFR for the user.

 

All the other tables i quickly looked at in ECUFlash are listed in AFR though, so would need to dig into where the conversion happens (ECUFlash or ECU) and what the conversion is. or maybe it's all in Lambda, and ECUFlash just kept that one table as-is.

 

makes me wish it was all in Lambda now, even if that would take a few extra seconds to understand.

Edited June 27, 2017 by Flinkly

[covertrussian]

That's weird, I've never seen it display lambda, what config are you using for ECU flash?

[solidxsnake]

The mA vs Voltage based calibration table is irrelevant. All sensors are lambda sensors, they physically cannot determine if you're running E0, E10, E85, pure acetone, 0w20 oil (), they can only sense oxygen content in the exhaust gas. Whether the sensor reports AFR or lambda is secondary, but they physically only measure lambda. In the case of our cars, the upstream lambda sensor sensing circuit senses a current, and the ECU contains a calibration table correlating the current (in mA) to a specific lambda (unitless). If you see the table listed with units in AFR in ECUFlash/RomRaider, it's because that was the scaling chosen for that table (by default I believe that's the case in all of the different unit systems).

 

In the ECU, all of the tables that involve "AFR" are actually stored as lambda. Take your primary open-loop fueling table, for example. In the ECU, the closed-loop portion is stored as "1.00" and not "14.7." Meaning that the ECU will be targeting 1.0 lambda for its closed-loop target regardless of the fuel type used. Whatever you're reading out in your logger is scaled after the fact.

 

As for the scaling, you can choose different scaling factors in ECUFlash or RomRaider. Look for the table scaling in the settings. You can even scale the tables to display AFR for E10, E85, etc., by just making your own scaling equation. When I get a chance, though, I'll show you what the raw data in the ECU looks like.

[Flinkly]

That's weird, I've never seen it display lambda, what config are you using for ECU flash?

 

i'm using an ALPHA release from a couple months back.

 

the ALPHA has (some) experimental tables that the normal release doesn't. Back when i pulled it i was after the two tables to effectively disable my rear O2 (which i did, and haven't had it installed since).

 

Recently heard about some TGV control tables i also don't have, even though i've got the ALPHA release.

[solidxsnake]

Here's what I'm talking about. At the top is the Primary Open Loop Fueling table definition in the ROM (formatted as a structure). The first number is the X-axis dimension (15), the second is the Y-axis dimension (18), the next three values are addresses in the ROM which are aliased by the program (hence the yellow font color). The next is an alias for the data type (8-bit values). After that are a couple zeros, followed by a scale factor (circled, 0.0078125) and an offset (0.0).

 

Below all of that is the table as stored in the ROM. I forgot that it's actually stored as an additive in the ROM (so everywhere the stock tune is targeting stoichiometric, the value in the ROM is "0") Right next to that is the table's actual values after converting them with the scale factor of 0.0078125 found in the table definition. Note that the "0.52" in the table corresponds to the "9.65" AFR value found in the table when using the "AFR" scaling. How do you arrive at this number? If you dig into the ECUFlash scaling definition, the formula is "AFR = 14.7/(1+x*0.0078125)" where "x" is the raw 8-bit integer value in the ECU's table. Notice that I can change whatever AFR scale I want to display in this formula by simply substituting the "14.7" with the stoichiometric AFR for whatever fuel I'm looking at. If I just substitute that "14.7" with "1.0," I'm left with lambda.

 

Below this is the definition for the upstream lambda sensor scaling. Note that the values are not AFR, but are lambda. Compare them to the values in the ECUFlash table to the right (which is scaled to display AFR) and you'll realize they're identical.

 

 

Note that since these values are stored as unsigned integers in the ROM, it is actually impossible for the ECU to command fueling that's leaner than stoichiometric, at least as the code was written from the factory.

Edited June 27, 2017 by solidxsnake

[covertrussian]

After much more reading and thought, I have a much deeper understanding of Lambda and how it all works, thus I want to thank you, solidxsnake (or shall I call you David , really you should change your username to Hal ), for opening that can of worms.

 

I didn't realize that it doesn't matter what fuel you run, stoich is always going to be 1.00 lambda. I thought you had to adjust the wideband to the fuel for it to show stoich correctly (outside of the eye friendly AFR multiplier). This is actually really cool! Also for this reason, I think I'm done with AFR's myself and gonna switch to lambda.

 

Since ECU uses Lambda and not AFR, that means there is really nothing to adjust from fueling perspective with E10, I got my hopes up yesterday thinking maybe I could gain some MPG from mocking with the AFR's

 

Also added some edit's to my other post to point out that information is wrong.

[Flinkly]

i only ask cause i'm not so in touch with raw ECU things as the rest of you, but what are the true values of the closed loop target comp tables? are they truly a value to be subtracted from stoich (lambda of 1) to reach the target AFR for those conditions (RPM/load)? are we again unable to go above a lambda value of 1 in closed loop?

 

makes sense that you can't in open loop, since that is coded to be when boost/power occurs, essentially.

 

And i agree with Covert. after this discussion, i'm done with AFR. too much work to make sure that all your AFR numbers (gauge, romraider, ecuflash, etc.) are calculated as you think the are or should be...

 

Lambda from now on. will take some getting used to.

[covertrussian]

On a bright side that math is not too hard if you still stay with AFR (Just subtract .5afr), and if your tuning car by web suggested AFR's it still scales well to E10. If I tune for 11.0 wideband displayed AFR, my real AFR is 10.60 on E10, which is still 0.75 lambda. Now if I ever use 100% gasoline, she's gonna be pig rich.

 

If I understand Ethanol correctly, it's oxygen rich, that's why lambda AFR is 14.1 (the other .56 afr is located in fuel). Which kind of cool if you think about it, if you have pure enough ethanol you could get away without having a fresh air source.

 

Also, I tried switching my fuel table in ECUFlash to raw ECU value and it didn't display right at all. Switching to Lambda didn't work right either. There one that show's 14.7's as 0's, but the rest are 4 digit negative numbers = too big to fit into cells and color code doesn't work at all. I think I'm just gotta leave it at AFR's, just knowing to subtract 0.5afr in my head. Since Everything scales pretty good even with the 14.7 calculation, I'm not sure it's worth my effort to worry about it.

[solidxsnake]

On a bright side that math is not too hard if you still stay with AFR (Just subtract .5afr), and if your tuning car by web suggested AFR's it still scales well to E10. If I tune for 11.0 wideband displayed AFR, my real AFR is 10.60 on E10, which is still 0.75 lambda. Now if I ever use 100% gasoline, she's gonna be pig rich.

 

If I understand Ethanol correctly, it's oxygen rich, that's why lambda AFR is 14.1 (the other .56 afr is located in fuel). Which kind of cool if you think about it, if you have pure enough ethanol you could get away without having a fresh air source.

 

Also, I tried switching my fuel table in ECUFlash to raw ECU value and it didn't display right at all. Switching to Lambda didn't work right either. There one that show's 14.7's as 0's, but the rest are 4 digit negative numbers = too big to fit into cells and color code doesn't work at all. I think I'm just gotta leave it at AFR's, just knowing to subtract 0.5afr in my head. Since Everything scales pretty good even with the 14.7 calculation, I'm not sure it's worth my effort to worry about it.

 

You'll probably need to write your own scaling equation (I don't remember if there is already a lambda scaling, or what it's named if it exists... the list of scalings are horrifically unorganized). The equation is just "Lambda = 1.0/(1 + x*0.0078125)" where x is the raw ECU value. You'll also need to write the reverse equation (solve for x in terms of lambda... you do remember your algebra, right? ). When creating your own scaling, you can choose how many significant digits to display so it'll fit within the confines of the table.

 

Flinkly: i'll look into the closed-loop compensation tables. My gut says it'll be similar to the OL fueling table (I think all of the fueling related tables store additive lambdas).

 

Edit: Sure enough, the table is stored as additive lambda in the ECU. This is what the stock CL Fueling Target Compensation A table looks like in the USDM '07 spec.B ROM in AFR and in lambda:

 

 

In this case, it does look possible to compensate positively. The values are stored as 16-bit integers, with a scale factor of 0.000015258789 (which is the reciprocal of 65536, i.e. 2^-16) and an offset of -0.5. Meaning that if I have an integer greater than 32768 (2^15), then the additive number will end up positive. It looks like it's possible to have lambda compensations on the range (-0.5, 0.5).

Edited June 29, 2017 by solidxsnake

[covertrussian]

Thinking more on this, and through every scenario that I've ran to, as long as everything is based around 14.7 scale, I'm not sure that switching to lambda readings really matters.

 

We already know that primary open loop table is not a target AFR table anyway, which would be the only place it would really matter, and only if ECU didn't read lambda.

 

About the closed loop fueling target table, thanks for confirming that is displayed in AFR's (and lambda), I wasn't sure if it was an arbitrary number or AFR's before. Also I remember folks on romraider saying they had issues if compensation was 0 or positive. See this thread.

[solidxsnake]

Thinking more on this, and through every scenario that I've ran to, as long as everything is based around 14.7 scale, I'm not sure that switching to lambda readings really matters.

 

We already know that primary open loop table is not a target AFR table anyway, which would be the only place it would really matter, and only if ECU didn't read lambda.

 

About the closed loop fueling target table, thanks for confirming that is displayed in AFR's (and lambda), I wasn't sure if it was an arbitrary number or AFR's before. Also I remember folks on romraider saying they had issues if compensation was 0 or positive. See this thread.

 

That's correct. If you're using only one type of fuel, the scale is completely arbitrary. Just need to know what numbers you're targeting and you're all set. That said, the only absolutely correct value is lambda, because when you're running E85 and your ECU is reading "14.7AFR" it's technically incorrect. The nice thing about lambda is that you only have to remember a couple values to target, regardless of whatever fuel you're using, and those numbers will be absolutely correct (if you're interested in the actual chemistry). Otherwise, there's really no benefit to switching.

Edited June 29, 2017 by solidxsnake

[covertrussian]

Right, if I was running E85 or even a mix between E0 and E10 I would worry about it more. But since I'm always using E10, it's just not worth the hours to redo the way Romraider/ECUFlash display things.

[covertrussian]

I was going to test AVCS at 10*, as I was mentioning last week, but after some thought I would rather test my custom intake with AVCS at 5*, since free flowing header and intake could increase the chances of fuel escaping during overlap phase. Round 1 of intake testing can be found here.

 

Intake it self is pretty simple, 3" aluminum pipe to a MAF housing that I cut out from a stock airbox, this insures same MAF housing and has a flow straightener.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/WAI%20v1/WAIv1_001.jpg~original

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/Custom%20Intake/CAIv1_001.jpg~original

 

To help increase airflow I like to use flow/velocity stacks. This reduces turbulence inside the pipe, which should lead to more airflow and more accurate readings.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/WAI%20v1/WAIv1_007.jpg~original

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/WAI%20v1/WAIv1_006.jpg~original

 

This intake sits pretty low in the fender well, my final version will sit higher and probably will have an airbox.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/Engine/Intake/Custom%20Intake/CAIv1_004.jpg~original

 

 

Tune updates and Highway Testing Results

Since I've used this intake in the past, I already new what tune changes to make (detailed post here). Since am running without an EBC, all I had to do is add 25% to MAF scales on my last weeks tune.

 

With the only change to the tune being the MAF scales, weather/traffic was also very similar to last week too. The intake got 30.97mpg!, with last week's trip getting 29.56mpg. Looks like intake increased gas mileage by 1.41mpg from last week, that's a tremendous improvement! Now I know I've gotten 31mpg once before to this city, but I was by myself and didn't do as many stops as I did today and last week (extra weight + store stops = more fuel used).

 

Here is the learning view for last week's trip with stock intake and +10% to MAF scales:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/CAI/201706_StockIntake_MAF%2010.jpg~original

 

Here the the learning view for today, with CAI and MAF scale increased by another 25% (+35% from stock tune).

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Intake/CAI/201707_CustomCAI_MAF35.png~original

 

Intake increased the low end airflow by a lot, requiring ECU to add 13% more fuel in 6-10g/s range. I would say the 10-35g/s range is within margin of error and not intake related.

 

Another thing I noticed today, while going up the mountain at -5inHg to 0psi my AFR's were already in 13's, this is due to me removing 1.0+g/s from closed loop range. This was also the case last week, and I believe that's why last week's MPG was lower then one that is slightly leaner at that range, I believe my city MPG suffered because of these changes too (I got 18.67mpg city only last week, where before I would get 19-20).

 

Thus I will update my tune to be 14.7AFR up to 2-3psi at lighter loads for next highway testing trip.

Edited July 4, 2017 by covertrussian

[covertrussian]

I remember why I removed the intake last time... we had a monsoon rainstorm this morning, I went to check the oil level and didn't feel comfortable with driving the car to work, water ontop of the flow stack and filter media was slightly wet.

 

Looks like I'll be raising it up to be more in the engine bay (which might help fuel economy due to being a WAI).

[covertrussian]

I tried out running my WAI again, it was getting hot way too quick without a heatshield, thus I figured I would give K&N Typhoon a shot again. It was also prone to rising in heat much faster then stock, but not quite as fast as my WAI, but I'm not concerned with heat and power as much as I'm concerned with gas mileage in this case.

 

http://legacygt.com/forums/attachment.php?attachmentid=252705&stc=1&d=1500436568

 

I fine tuned for the intake, it needed +15% more MAF scale (totaling 25% more then stock, since my stock intake has been happier with +10% MAF scale). I fine tuned the WOT timing and fuel, cruising timing did stay the same.

 

For my city test, I drove the car really nicely, the whole time it seemed like the car would do better because the needle didn't move as quickly. That got me 17.92mpg city, while stock intake got 18.57mpg, both with no AC or headlight usage. While the tests were similar, they were not 100% identical (really hard to have even 90% identical variables in the city though), thus the only conclusion that I can come up with is, K&N Typhoon got about the same gas mileage, but not really better.

 

After a week of slow driving, I took it on a highway, this time I went alone (~200lbs less weight), once again the needle didn't seem to move as quickly down. While I did hit some crash traffic (requiring a local road detour for 5 miles), the car got 29.17mpg highway, which is a far cry from 30.97mpg I got two weeks ago with extra weight.

 

One interesting thing that I did notice is the car felt more powerful closer to 0 psi. I wonder if it was making too much power, thus using less throttle and increasing pumping losses.

 

While I think it might be possible to squeeze an MPG out of this intake, I really didn't like driving with it, the IAT's climbed to 140+*F all the time, I just don't think it's worth the effort to fine tune it, when my custom intake gained 1mpg instantly (which I'll be testing more after I build a heatshield). Plus my power testing also didn't show any real gains (~5whp in some parts of the curve).

http://legacygt.com/forums/attachment.php?attachmentid=252706&stc=1&d=1500436568

[covertrussian]

Recently have been thinking if going to a bigger turbo actually hurt my highway gas mileage (off boost of course). The possible reason is, smaller turbo spools easier so starts building boost sooner, helping overcome pumping loss inducing manifold vacuum.

 

I've personally seen this twice already: On my old Nissan that had a ball bearing T25 (spooled real fast), going to a journaled T28 cost me gas mileage. Then switching from VF46 to my Big16g, granted reducing timing by 7* helped me regain the losses, but not necessarily go above.

 

Reduction of pumping losses is ones of the bullet point pros for twin scroll manifolds: twin scroll vs single scroll comparison by dsportmag.

 

Now in the city, I can see a bigger turbo helping since it's a lot easier to overthrottle while just daily driving.

 

What are your thoughts? Just something I've been thinking about lately

Edited August 11, 2017 by covertrussian

[Flinkly]

Recently have been thinking if going to a bigger turbo actually hurt my highway gas mileage (off boost of course).

 

Blah, blah, blah...

 

Reduction of pumping losses is ones of the bullet point pros for twin scroll manifolds: twin scroll vs single scroll comparison by dsportmag.

 

Blah, blah, blah...

 

What are your thoughts? Just something I've been thinking about lately

 

well, while it reads like you're only talking about one thing, you're really talking about two different things.

 

1. "bigger" turbo vs. "smaller" turbo (vs. efficiency, obviously)

2. the increased "exhaust" efficiency of a proper TS system

 

your first point is more of a thought experiment. could come up with hypothesis both for and against. most are probably right (to an extent), but with your current system it appears that the "against" column carries more weight.

 

and the second point is "fact", at least for the Subaru OEM SS vs. TS systems, although the amount of savings to be had is still under debate.

 

all in all, there are just so many variables.

 

 

in other news, with this 100+ degree weather we've been getting here, i've essentially got a mild WAI going on and have seen a small increase in mpg.

Edited August 11, 2017 by Flinkly