Tuning for Fuel Economy

Most people don't care about gas mileage, I personally find it as a fun challenge trying to achieve great gas mileage out of turbo "gas hogs".

 

This thread is basically a brain dump of my testing results and theories. Since finding ECU based tuning articles/threads is really hard this thread will be mainly about finding ECU tricks to improve gas mileage, but it will have some mechanical portions too.

 

Since this thread is more about the endless testing and gritty details, eventually I plan on building a How To article with all the mods that actually worked.

 

In the next post you will find more detailed findings and tuning items.

 

Quick Links to Posts:

 

Cold Air Intake Testing Round 2

TGV Delete MPG Testing

Spark Plugs: 2005 LGT SILFR6A vs 2010 LGT SILFR6B8

Edited May 6, 2019 by covertrussian

 

Most of the items shown here will be based around 05-06 Legacy GT tunes. Since I also have an 04 FXT I test various tune changes on it too and tend to compare the tune differences to the LGT. Most of the things that I've done to the LGT tune to increase fuel economy came by default on the 04 FXT tune. It feels like Subaru spent more time trying to sqeeze every MPG out of the Forester compared to the Legacy. Even after modding the Legacy tune, my Forester (fairly stock) still gets 1-2mpg more city then Legacy does, but the two cars do have different gear ratios, turbos, etc, thus it's hard to do 1 on 1 comparision.

 

Fueling

Primary Open Loop Fuel Map

Only areas that really need attention are higher load areas that are very rich. Cruising portions of the map are pretty good as is. But I'll test various settings.

 

This graph shows how rich stock tune is and how innacurate the factory wideband is compared to a standalone wideband.

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

 

Closed Loop Target Compensation

05 LGT tends to target 14.2afr at idle and 14.6afr cruising. I noticed that highway AFR range goes from 14.4-14.7 on my UEGO, which is on the richer side. Usually you want to see 14.5-14.9 ranges, constant rich lean helps insure proper catalytic converter function. More details here.

 

05 LGT Stock Closed loop target fueling

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

 

STUB: more research and testing to be done in this area

 

Closed Loop to Open Loop Delay

While calm city driving, CL to OL delay didn't increase my fuel economy at all. But with aggressive driving it does seem to help slightly, around 1mpg, not worth the engine damage keeping it on though.

 

AF3 Correction Disable

Disabling AF3 correction disabling has helped by 1-2mpg increase on my FXT, but nothing noticeable on my LGT (Could be because Legacy sensor was not as bad?). This tune change is a must for catless downpipes.

 

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Rear%20O2/AF3CompsZeroing.png~original

 

Tip-In Enrichment

STUB

 

Cold Start Fueling

These cars tend to not be too rich while cold idling. Within a minute of cold starting at 30*F the car was already in closed loop hitting 14+ AFR's. The stock tune might be richer for longer periods below 30*F, but otherwise so far I'm not seeing much gain in this area.

 

STUB: more research and testing to be done in this area

 

 

Ignition Timing

Timing table

 

05 LGT Stock Total Timing:

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

 

04 FXT Stock Total Timing:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing%20Advance/04FXTStockTotalTiming.png~original

 

 

Highway Range Timing

Increasing timing to 45* at 70mph range +1mpg (STUB)

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing%20Advance/45Deg70-75mph.png~original

 

City Range Timing

05 LGT tune runs very little timing below 1600rpm, could explain why it's such a dog in higher gears in the city. 04 FXT tune tends to run more timing down low, which could be another point that helps with city MPG.

 

STUB: Currently testing more city timing.

 

Base Idle Timing

STUB

 

IAT Timing Compensation

STUB

 

AVCS Tuning

AVCS tuning has been the holy grail of improving fuel economy on my Legacy.

 

05 LGT Stock AVCS

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/AVCS/05LGTStockAVCS.png~original

 

Stock Downpipe best MPG came from AVCS being at 15*. Sadly I ran over a dear before I could do more testing with stock downpipe. I did recently aquire a newer one and plan on eventually installing it to test again.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/AVCS/AVCS15v11.png~original

 

With aftermarket downpipe, the car preferred AVCS to be at 10* at cruise areas. Going from stock AVCS 0*, to 10* improved my winter highway fuel economy by 3mpg!

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/AVCS/AVCS10DPv2.png~original

 

 

04 FXT Stock AVCS:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/AVCS/04FXTStockAVCS.png~original

 

With FXT, since I rarely drive it on the highway, I am able to test city mpg more often. I've played a little more around with AVCS for the FXT in city areas.

 

Stock AVCS has gotten be the best MPG at 19.38, AVCS 15 got me 18.08mpg, and AVCS 10 got me 18.30mpg. Basically the smaller turbo of the FXT is more sensative to overlap (even with a catless STI up pipe).

 

 

Boost & Wastegate

Target Boost

 

Wastegate Duty Cycles

 

Mas Airflow Sensor

MAF IAT Compensation

STUB

 

MAF Sensor Scaling

STUB

 

Drive by Wire Throttle

 

 

Ambient Temperatures

I found Ambient temperatures alone (not even winter gas) have a great effect on fuel economy on this car. When temperatures dip below 50*F, my fuel economy seems to drop by 2-3mpg, when temps dip below 30*F I tend to loose 1-2mpg more. This can be regained partially with higher tire pressures in temps above 30(F, but it doesn't seem like tire pressured helped with gaining back economy below 30*F.

 

Funny thing, I think this is related to night driving and headlights: Temps dip below 50F in the winter, where it gets darker sooner which means more night driving and headlights usage.

 

Night Driving

Over the life of this car I've noticed that my night time gas mileage is always lower then day time gas mileage, by as much as 3mpg even at the same ambient temperatures. One theory is electric parasitic loss from having the headlights on.

 

Headlights

This one sounds funny, but in a back to back city test I lost 1.25mpg by running my headlights (low beam + fogs) all of the time vs no lights at all. Most people might not see as much of a difference due to DLR's being on. My DLR's are disabled thus my headlights and other lights are off unless I explicitly turn them on.

 

Air Conditioning

To my surprise I haven't found any difference between using air conditioning and not, both city and highway! Having headlights on during the day ate more of my gas mileage then having AC on

 

 

Suspension

Tires

Switching the tires to new wider (from 215/45/17 to 225/45/17) tires, did make me loose a negligible .27mpg post tire size correction. There are a ton of variables here though: different tread patterns, different rubber compounds, and new tires will always give you worse MPG because they have more tread flex which increases drag.

 

The reason I say post tire size correction, stock tire size makes your odometer be off by -1.5%, but speedometer reads correctly. With 225/45/17 tires, my speedometer is off by 1mph, but odometer reads on par with my GPS and Google Maps. Odometer is what matters, if you run stock tires, and don't apply the correct tire offset, you will be reading about .40mpg more then you really are getting.

 

Tire Pressures

Tire Pressures does effect gas mileage, more so in the winter then in the summer. This is because colder air doesn't increase the warm tire pressure as much as warmer air does. With my personal experience,the deciding temperature is 50-60F. Below this temperature range, I have to run much higher cold tire pressure to gain back the lost mpg on 225/45/15 tires.

 

I gained ~2mpg highway at 35-44F ambient temps, by simply increasing the cold tire pressure to 42f/40r (warm 45f/43.5r) from 36f/34r (warm 39.5f/38r).

 

My gains were negligible on warmer days though, I believe the warmer days will get 36/34psi closer to 42-45psi warm. Which seems to be the sweet spot for my tires. I do also believe the magic tire pressure number will vary based on your tire width. Wider tires will need more cold tire pressure to have less drag. I highly suggest looking for 50psi max sidewall tires, to give you plenty of playing room.

 

For city driving, higher tire pressures have resulted with non-existing gains. Part of the reason could be due to my short ~5mile trips, where the tires only warm up by 1-2psi tops. Higher tire pressures do make the city NVH (Noise, Vibration, Harshness) much worse.

 

Alignment

To be tested

 

Brakes

Regreasing all of the caliper pins and removing the rubber bushings that expand over time and drag, ~1mpg gain city.

 

Transmission

Drivetrain Fluids

My LGT is running the Smurf Blood mix (3qt Redline Lightweight Shockproof, .8qt Motul 300). I switched to a different fluid and sadly it didn't help city nor highway gas mileage during the summer, winter to be tested

 

 

Engine

Crank Pulley

GrimmSpeed lightweight crank pulley made no noticeable MPG improvement on my LGT. FXT gained .70mpg gain city though.

 

Intake

On FXT switching from aftermarket air filter to OEM one, has yielded a .40mpg gain city.

 

Turbo Inlet

No noticeable improvements going to an aftermarket silicon inlet, even with replacing a pretty torn stock inlet.

 

Spark Plugs

Currently I am running Side Gaped plugs, Unfortunately I do not have much useful information for preside gaping. When I replace the spark plugs with new ones I'll have better comparison. On my Nissan's, side gaped plugs usually increase 1-3mpg though.

 

Exhaust

UpPipe

On FXT going from stock catted up pipe to STI Catless uppipe yielded me a negligible .17mpg gain city.

 

Unfortunately I do not have any numbers from the LGT going to an Invidia Uppipe, I broke the O2 sensor during installation, thus needed a brand new one. This negates any gains/loses becaue it's a major variable change.

 

Downpipe

Stage 2 Downpipe in it self doesn't seem to help fuel economy. In fact it starts hurting it until you disable AF3 corrections, but I also found I still lost fuel economy even post AF3 disabling, due to tune running too much timing or too much AVCS advance.

 

After a retune a better flowing downpipe should in theory help with fuel economy due to reduction of pumping losses. I am planning on going back to stock downpipe for more testing eventually.

 

 

Fuel

Winter vs Summer Gas

It is easy to blame Winter gas for decreased fuel economy. Since I've been religiously adjusting my tire pressures and logging ambient temperatures. I've noticed that my winter fuel gas mileage has been approximately the same as summer fuel.

 

E10 vs 100% Gas

To be tested, all of my current testing is done on E10 Shell 93 gas.

Edited November 6, 2016 by covertrussian

Just got back from a 1200 mile roadtrip, got plenty of tune MPG testing done. Sadly this was a hilly trip (81 south to 40 west), I tried to note down major elevation changes, but still this skews the data as always. I set the cruise control to 70mph, usually the car likes to go +/-1mph from that set speed. Temps hung around 65-85*F for these runs, I tried to fill up around 150miles for these runs. Car had another 500lbs worth of additional weight (ontop of base car weight).

 

Base run

AVCS: 15* @ 2800-4000rpm to 0.90g/rev

Tire Pressure: 47.5f/45r psi

Temps: 65f-75f

Speed: 71mph

Elevation: 1,500ft to 2,800ft to 2,000ft

MPG: 27.75

Notes: was a simple map that I've been running for a while, AVCS set to 15* for 70+mph, stock DBW tables, stock target boost, and WGDC tables just adjusted for downpipe (still produces stock boost). This made 27.75mpg, not great but not bad, I've bee noticing that the car is no longer liking AVCS at 15* with the downpipe.

 

Second Run

AVCS: 15* @ 2800-4000rpm to 0.90g/rev

WGDC: -10 in cruise, 0 in moderate throttle (see picture below

MPG: 30.30mpg

Temps: 75f-85f

Tire Pressure: 47.5f/45r psi

Speed: 70mph

Elevation: 2,000ft to 2,500ft to 1,400ft

Notes I went with using a reduced boost tune, since the car hits 10+ psi by holding the cruise control up button for 1-2 seconds, I wanted something that's less likely to get into boost. Same AVCS as before, etc. With some traffic this tune got me 30.03mpg! I will say this table needs to be fixed up to have smoother drivability. With this if you tap the cruise button, it will increase speed by around 1mph, very gently. If you hold it for 1+ seconds it still will hit 10+ psi.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/MPG%20Tuning/MPG_WGDC_001.png~original

 

Third Run

AVCS: 15* @ 2800-4000rpm to 0.90g/rev

WGDC: Stock

DBW Map: Linear map similar to this one.

MPG: 26.31mpg

Temps: 65f-75f

Tire Pressure: 47.5f/45r psi

Speed: 70mph

Elevation: 1,200ft to 500ft to 900ft to 400ft

Notes: Decided to try my Drive By Wire remapped tables on the stock target boost tune . Surprisingly this one didn't do well at all and got me 26.31mpg.

 

 

Fourth Run

AVCS: 15* @ 2800-4000rpm to 0.90g/rev

WGDC: -10 in cruise, 0 in moderate throttle

MPG: 28.50mpg

Temps: ~70f

Tire Pressure: 45f/45r psi

Speed: 70mph

Elevation: 500ft to 2,000ft

Notes First Fill-up heading back home, since reduced boost at cruise tunes did so well I used them as bases for the rest of the runs. Still using AVCS @ 15*, but changed the timing table. I noticed that at 2800, stock timing runs 40.12* max, while at 3200 it runs 45.04*. I moved the values down to 2800, that way at 70mph it would run the same timing as at 75mph. This got me 28.50mpg, not bad but not as good as before still.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing%20Advance/Timing70mph45deg.png~original

 

 

Fifth Run

AVCS: 10* in cruising

Timing Advance: 3200's timing moved to 2800 at low loads

WGDC: -10 in cruise, 0 in moderate throttle

MPG: 29.91mpg

Temps: ~80f

Tire Pressure: 45f/45r psi

Speed: 70mph

Elevation: 2,000ft to 900ft to 1,800ft to 1,400ft

Notes I saw some knock at 1.15-1.30 range @3200, figured I would play it safe and run AVCS at 10* tops with increased timing. This got me 29.91mpg!

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/AVCS/DPSTG2AVCS10.png~original

 

 

Sixth Run

AFR: ~14.3afr Closed loop

AVCS: 10* in cruising

Timing Advance: Stock

WGDC: -10 in cruise, 0 in moderate throttle

MPG: 27.68mpg

Temps: ~70f

Tire Pressure: 45f/45r psi

Speed: 71mph

Elevation: 1,400ft to 2,600ft to 2,100ft

Notes With this one I wanted to try running the car richer (14.4afr). Reason for wanting to run it richer is, with e10 stoichmotic is no longer 14.7, it's actually closer to 14.2, but since we can't 100% know how much ethanol each gas station uses it's better to just meet the afr's halfway at 14.4.

I modified the O2 sensor scaling to full it to think, that .10v = 14.70. Sadly the math I did at 2am, got me to around 14.3afr. I also cant read and uploaded the map with stock timing map (not the 45* @ 70mph one). Anyway richer AFR's got me 27.68mpg. I may try this again later, but so far not looking good

 

Seventh Run

AFR: ~14.7afr Closed loop

AVCS: 10* in cruising

Timing Advance: Stock

WGDC: -10 in cruise, 0 in moderate throttle

MPG: 29.57mpg

Temps: ~70f

Tire Pressure: 45f/45r psi

Speed: 71mph

Elevation: 2,100ft to 900ft to 1,800ft to 1,300ft

Notes Since I uploaded the stock timing map, I now needed data with AVCS being at 10* and having stock timing so that I could fully verify that 14.4 AFR's did worse. Same map as eighth fill-up, just with stock 14.7 O2 sensor scaling. This tune got me 29.57mpg.

 

Take Aways from Testing

• Looks like AVCS being at 10 at cruising speeds is a better way to go when your Stage 2. Two weeks ago I even got good results with AVCS being at 0* or 5*.
  
• There is some potential in increasing timing around 2800rpm
  
• There is a lot of potential in messing with wastegate tables
  
• Not much to be gained from cold tire pressures above 36 front/34 rear psi (in warm weather)
  
• AFR's: Even with up to 10% ethanol (e10), 14.7afr's are better for fuel economy.
  
• These cars eat a lot of gas city driving and eat a lot if you don't baby it on the highway. I think this is a tuning issue, and is on my to do list to fix.
  
• With 30mpg within fingers reach, new goal is 35mpg.
  

Edited January 21, 2015 by covertrussian

One thing that I want to mention is ambient temperatures reducing fuel economy. Going below 60*F ambient seems to hurt fuel economy by around 1-2mpg. I think it might be an enrichment table someplace. Testing this side by side is rather difficult too.

Sweet thread! Just wondering what the philosophy behind putting vaccuum values in the target boost table is. I actually just filled my first column in this table to -.5 to try to cure my stumble, but I also read somewhere that the car is essentially running 100% duty cycle untill it reaches the spring pressure anyway...

 

Instead of messing with O2 sensor scaling you should check out the CL fueling comp (load) tables. You can adjust the fueling targets to whatever you want depending on revs and load...

Target boost table and Initial Wastegate Duty Cycle tables work hand in hand. ECU uses the initial WGDC values and sees how close your MAP sensor is to the Target Boost table. When you are not running the exact value in Target Boost, you get boost error (looks to be in PSI). ECU then starts adding duty cycles ontop of the Initial WGDC to get you to Target Boost, it will do this until it hits 0 boost error, or it hits the values in Max WGDC table (hard stop for duty cycles, ecu will not go above these values).

 

While you are cruising you are between 0psi and -25inHg (vacuum). The factory MAP sensor doesn't seem to read below -11inHg. In theory you would be running about -10psi boost error while cruising, but I'm not sure if ECU is able to run less duty cycle then in Initial WGDC table.

 

The goal is to reduce boost error to avoid ECU from overcompensating on duty cycles.

 

Stock Target Boost for reference:

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/MPG%20Tuning/05LGTStockTargetboost.png~original

 

Wastegate:

The 100% duty cycle till it reaches the spring rates is somewhat correct. In the perfect world the wastegate should stay shut until your wastegate spring rate is reached, then it flips open and lets exhaust bypass the turbine. In reality as you start building boost, the boost it starts pushing the wastegate open as soon as you start building boost. You do also have exhaust back pressure that starts pushing the wastegate open too.

 

This is why if you run no boost controllers you always want to get your boost source from as close to the throttle body as possible instead of getting it from the turbo outlet. Turbo outlet sees boost first, and starts opening the wastegate, with pressure drop in intercoolers & piping, by the time you hit the manifold your running 1-2psi less then what your wastegate spring is rated for.

 

Electronic boost controllers work on solving this by bleeding the boost and keep the wastegate mostly shut until target boost is hit. The way the factory WGDC tables are, it seems like it keeps the wastegate shut until you hit your target boost.

 

This is why our cars feel so responsive too, I thought it was because of the tiny turbo, but really it's because of the wastegate being forced shut by ECU even when building boost. I need to do some heavy logging and seeing how the car feels with just wastegate pressures.

 

 

O2 Sensor:

Reason for messing with O2 scaling over CL Fueling table is, CL fueling doesn't have target AFR's, it just has negative values to subtract from main map. Then you would still need to modify a few other tables to make sure they are in check (there is a table that makes you go open loop below 14.41afr). Easiest just to full the ECU and use a real wideband .

wait, can "we" fool the ecu to accept a wideband rather than stock narrow o2?

 

fantastic thread. subb'd.

wait, can "we" fool the ecu to accept a wideband rather than stock narrow o2?

 

fantastic thread. subb'd.

 

On 32bit cars front O2 sensor is a wideband, by definition wideband is able to read wider range of AFR's then just 14.7. Stock front O2 reads 20afr-11.1afr.

 

I'm also using an external wideband, AEM UEGO, it is not tied into the ECU though (probably is possible to physically wire it in).

about the O2, yeah, it's not narrowband, but it's not as good as "aftermarket" wideband.

 

I was really poking at the ability to put in a Bosch (true) wideband sensor as the front O2, where your ECU could not only have a better view of the world, but you wouldn't necessarily need a second O2 sensor/bung but still get all the information you'd want for tuning. you could still even take the raw voltage and put it thru an aftermarket box to get the same results you have now (kind of a double check, ECU vs. aftermarket wideband, but the same sensor). hard part would probably be in tricking either your aftermarket sensor "module" or the ECU into thinking it was heating the sensor up, since both would "want" to be doing it.

 

As for AVCS, I was thinking about following your lead and seeing what results I could get, but my Stock and Stage 2 tune have a very different look than your Stock AVCS tables. not that I'm very surprised, but we do generally have the same intake/exhaust systems, but these are pretty different.

 

And what about tuning individual AVCS cells for V.E. (volumetric efficiency)? so instead of a guess-and-check method, actually tune "each" cell for efficiency (you could tune a few spots and "infer" between them). not that it would give you best MPG, but it's be a great start, and then try bumping it up/down by 5* across the board and see what happens.

 

For a DBW car, it'd be hard to run within the right cell for an extended period of time to collect data, but not impossible...

Edited September 18, 2014 by Flinkly

also, to the point of tricking the ECU to burn at a different AFR than 14.7, wouldn't you want to go higher? say 15 or 15.5? Edited September 18, 2014 by Flinkly

about the O2, yeah, it's not narrowband, but it's not as good as "aftermarket" wideband.

 

I was really poking at the ability to put in a Bosch (true) wideband sensor as the front O2, where your ECU could not only have a better view of the world, but you wouldn't necessarily need a second O2 sensor/bung but still get all the information you'd want for tuning. you could still even take the raw voltage and put it thru an aftermarket box to get the same results you have now (kind of a double check, ECU vs. aftermarket wideband, but the same sensor). hard part would probably be in tricking either your aftermarket sensor "module" or the ECU into thinking it was heating the sensor up, since both would "want" to be doing it.

 

Widebands generally have an output wire, MTX-L has a wideband and narrowband outputs. What you do is you splice the wideband wire into the signal wire of your stock o2 sensor (There are 4 usually, 2 heater wires, ground and signal). If your car has a Narrowband then it's fairly easily to do, but if your trying to splice in Wideband, you need to corelated voltages to AFR's. The table I modified is where you corelate voltages to AFR's, thus in theory you could use that. You will need to make sure you also remove the lower limit on AFR readings (which is set to 11.1).

 

So yeah in theory it should be as simple as running a wire from your wideband's 0-5v output to your factory WBO2's signal wire and updating the voltage to AFR table. As for heater wires either leave the sensor in (cut the signal wire), or remove the factory sensor altogether and disable heater CEL's. Your wideband control unit should be the only one providing heating to the aftermarket wideband!

 

also, to the point of tricking the ECU to burn at a different AFR than 14.7, wouldn't you want to go higher? say 15 or 15.5?

 

100% fuel has stoichiometric of 14.7afr. E10 has stoichiometric of 14.2afr. Running 14.7 on E10 is the same as running 15.2 on 100% gas. Running 15.5 on E10 would be the same as running ~16afr on 100% gas if the math in my brain is workign corectly.

 

Different engines respond differently to running leaner then 14.7afr, some gain MPG, but most loose too much torque thus require more throttle to maintain the same speed, more throttle = more air = more fuel.

 

Problem is we are not 100% garaunteed to run e10, some gas stations say contains 10%, some say contains up to 10%. Because of that a lot of people try to meet the AFR's halfway at 14.4afr, esentially 14.9afr on 100% gas.

 

Continued in next post for AVCS...

As for AVCS, I was thinking about following your lead and seeing what results I could get, but my Stock and Stage 2 tune have a very different look than your Stock AVCS tables. not that I'm very surprised, but we do generally have the same intake/exhaust systems, but these are pretty different.

 

And what about tuning individual AVCS cells for V.E. (volumetric efficiency)? so instead of a guess-and-check method, actually tune "each" cell for efficiency (you could tune a few spots and "infer" between them). not that it would give you best MPG, but it's be a great start, and then try bumping it up/down by 5* across the board and see what happens.

 

For a DBW car, it'd be hard to run within the right cell for an extended period of time to collect data, but not impossible...

 

You have an 07+ car, which has better flowing heads and no cat in the up pipe. Since you also have 4 different maps it does make life difficult on figuring out what is where. Luckily you have Single AVCS right? Dual AVCS is where it gets much more difficult to figure out how much actual advance you have.

 

I personally think 40* advance is way too much, on my 05 going to 30* hurt my gas mileage. 20* was the max on stock downpipe, with a STG2 downpipe the car didn't like even 20*. I would try to model your AVCS map to resemble and older styled one, basically 10* advance everywhere cruise and see if it helps, if it doesn't try increasing it to 20*.

 

Now as for tuning each cell for VE, that's pretty hard to do, but you can try trusting the calculated VE. I've heard people saying that they made the best MPG slightly below maxium VE (the whole concept of tuning up to MBT but not over).

 

The only reason it's hard to keep in the right cell is becauase of the jumpy DBW tables, if you switch to a linear table that issue is resolved. It's really hard to tune it well without an Access Port, with an AP you can make real time changes, while with Tactrix you have to reflash it every time.

Edited September 18, 2014 by covertrussian

Hmm. May look at my tune's tables, just for fun, and see what they look like..

(I have e-tunes from 2 different vendors)

Target boost table and Initial Wastegate Duty Cycle tables work hand in hand. ECU uses the initial WGDC values and sees how close your MAP sensor is to the Target Boost table. When you are not running the exact value in Target Boost, you get boost error (looks to be in PSI). ECU then starts adding duty cycles ontop of the Initial WGDC to get you to Target Boost, it will do this until it hits 0 boost error, or it hits the values in Max WGDC table (hard stop for duty cycles, ecu will not go above these values).

 

While you are cruising you are between 0psi and -25inHg (vacuum). The factory MAP sensor doesn't seem to read below -11inHg. In theory you would be running about -10psi boost error while cruising, but I'm not sure if ECU is able to run less duty cycle then in Initial WGDC table.

 

The goal is to reduce boost error to avoid ECU from overcompensating on duty cycles.

You have the concept correct but it’s important for you and others to understand this:

 

Your boost control system has absolutely NO ability to open your wastegate until you are actually generating boost equal to or greater than the spring pressure. What this means is that below about 7 or 8 PSI, you really can’t prevent your turbo from spooling. What is also means is that your targets below 7 or 8 PSI don’t really mean anything. You will get to 7 PSI at 0WGDC just as fast as you would get there at 100 WGDC or 1000 WGDC because at below 7 PSI the pressure is not high enough to affect the actuator.

 

If you properly adjust the boost control activation thresholds, then you won’t get annoying issues with Turbo Dynamics Integral and Proportional trying to compensate by needlessly adding or pulling WGDC. This will especially correct issues with TD Integral 'Windup'.

 

My activation threshold is currently at Disable Below 6.95 and Enable Above 7.50. When I am cruising, I have a huge boost error, but that doesn’t matter because boost control isn’t enabled. And there’s nothing it could do even if it were enabled anyway. Putting -10 in there likely won’t make a difference though anyway because stock threshold is still around target 2 PSI.

 

Wastegate:

The 100% duty cycle till it reaches the spring rates is somewhat correct. In the perfect world the wastegate should stay shut until your wastegate spring rate is reached, then it flips open and lets exhaust bypass the turbine. In reality as you start building boost, the boost it starts pushing the wastegate open as soon as you start building boost. You do also have exhaust back pressure that starts pushing the wastegate open too.

 

This is why if you run no boost controllers you always want to get your boost source from as close to the throttle body as possible instead of getting it from the turbo outlet. Turbo outlet sees boost first, and starts opening the wastegate, with pressure drop in intercoolers & piping, by the time you hit the manifold your running 1-2psi less then what your wastegate spring is rated for.

 

Electronic boost controllers work on solving this by bleeding the boost and keep the wastegate mostly shut until target boost is hit. The way the factory WGDC tables are, it seems like it keeps the wastegate shut until you hit your target boost.

2 port (bleed type) boost controllers bleed. 3 ports (interrupt) type don’t bleed: they completely divert, hence the name ‘interrupt’. When a 2 port bleeds pressure away from the WGDC, it’s just opening a value and allowing boost to ‘bleed’ back to the intake tract. This means that no matter what, the WG actuator will still see a small amount of pressure coming. “Bleeding” is basically creating a leak for the pressure to escape back to the intake.

 

A 3 port, on the other hand, actually redirects pressure away from the WG completely. This gives you more fine and immediate control over the actuator.

 

This is why our cars feel so responsive too, I thought it was because of the tiny turbo, but really it's because of the wastegate being forced shut by ECU even when building boost. I need to do some heavy logging and seeing how the car feels with just wastegate pressures.

In theory, it’ll feel exactly the same while spooling, IF you have an interrupt boost controller (3 port) and the WGDC values and targets are properly set up.

I should clarify, you are 100% correct on ECU cannot open the wastegate. Btw I love these kind of discussions, just to add more info onto my theory.

 

In a very simple setup without boost controllers, your wastegate feed comes from your intercooler piping track (right before throttle body). As you start building boost, the boost will start slowly overcoming the spring inside the wastegate and it will start opening.

 

Just for reference a cutaway at an internal wastegate actuator:

http://www.turbosmartusa.com/wp-content/uploads/2012/12/IWG_01.jpg

 

Now to throw some funny numbers, say you have a 8psi wastegate spring, at 7psi of boost, wastegate will be 100% open. At 4psi it should be around 50% open (assuming a WG spring is linear). That means that by 4 psi, you are already diverting exhaust gasses from the turbine, which means building the last 4psi will take longer then building the first 4psi (unless you throw more exhaust gasses at it with more RPM's). This does make a fairly laggy setup that most people wont be happy with.

 

 

Boost controllers simply bleed (2 port, like the factory one) the air in the wastegate line to fool the wastegate actuator into thinking its building less boost then it really is. With our stock setups, ECUs are smart enough to know how much to bleed to get to desired boost levels.

 

To go back on the 8psi example, our ECU's will bleed the wastegate feed line so that the wastegate stays shut until target boost is hit. Which means the wastegate should stay close to 100% closed until you hit desired boost, and then ECU will stop the bleeding of the wastegate feed line, which will allow the actuator to open to bypass the turbine.

 

 

I personally don't think setting target boost to -10 is as important as simply zeroing out the WGDC tables, but knowing that turbo dynamics can add WGDC's, I wanted to stay safe in logic for the testing. I'm planning on doing some logs to see how the ECU reacts to different target boost though

Hmm, I did some more research and neither of us is 100% right about this....

 

The strength of a spring is measured using a spring constant K = F/L (or D) which is Force divided by Length/Distance stretched. Not PSI. The problem is that there is no way to directly convert the amount of boost seen (PSI) to the amount of force the spring sees without knowing the amount of surface area the inside of the actuator exposes to the boost pressure. We need that number to figure out how many actual pounds of force is actually applied to the spring by a certain amount of PSI boost.

 

But you're absolutely right, if it's operates linearly, it'll actually start opening as soon as ANY pressure (even a small amount) is applied to it so it seems I was mistaken about this.

 

As far as the wastegate being open 100% at 7PSI or 8PSI, I'm actually not sure if this is the case. This would imply that the turbo's WG boost number is being determined by the size of the WG exhaust opening.

 

MY feeling on the matter is that the WG opens up more and more as the turbo attempts to generate more boost, and so it keeps it in check at around the 7-8PSI range.

 

A way to test this would be to completely disconnect the WG actuator and leave the WG door free swinging therefore allowing as much exhaust to bypass the turbine as possible. I wonder how it would produce.

Keep in mind that the exhaust gas back pressure will also swing the wastegate door open with such a light spring (7-8psi). I've had it happen to me on my Nissan, with ~6psi spring, even with the wastegate feed line fully disconnected I still couldn't build above 10 psi.

 

As for disconnecting the actuator, I've tried it before actually. I would build about 1-2psi by redline, it was very laggy ha. It didn't gain me any gas mileage on the stock tune, I think it could be because some boost does help with wind resistance/overcome drag.

 

I finally filled up my car after a week of calm city driving with the -10target boost, 0wg tune. Got 17.25mpg, my usual MPG for calm city driving is around 18-19mpg. It does seem like, less boost could actually be hurting city mpg due to lack of power to get out of it's own way.

As for disconnecting the actuator, I've tried it before actually. I would build about 1-2psi by redline, it was very laggy ha.

Ah, well, there you have it. So 7-8PSI doesn't open the WG door 100%... it just opens it enough to keep boost in check at that amount. So in theory, you could get a weaker actuator and reduce WG boost down even more, lol.

Ah, well, there you have it. So 7-8PSI doesn't open the WG door 100%... it just opens it enough to keep boost in check at that amount. So in theory, you could get a weaker actuator and reduce WG boost down even more, lol.

 

Yup you are correct. You could get a weaker actuator, but it will probably start flipping open due to back pressure at 13psi . 6-8 psi on this car feels stupid slow though, the turbo is too small and inefficient for such low boost levels. Personally I think zeroing the tables is more then enough. I'm trying to review my logs to see how different target boosts effect the duty cycles.

 

Something to consider with stock turbos is the length of travel for wastegate arm (it is limited on how far it will go) and also the stock downpipe has a blank plate in the way that prevents the wastegate from opening too far. Which means with wastegate fully disconnected, your still not necessarily testing a real life limit of the current wastegate actuator.

I discovered that my tire pressure gauge went bad at some point (months ago), thus my my cold tire pressure was actually 45-47psi when it was supposed to be 34-36psi. My warm pressure increased by 5psi on a 60*F day, after driving on the highway for 1-2hrs.

 

In recent testing reducing the tire pressure to 36 front 34 rear actual reduced my gas mileage from 30.25mpg to 27.36mpg. Not ECU related I know, but just shows that tire pressures are important. With colder temperatures the cold psi can be down, but also highway psi might not increase by as much.

 

I think I'm gonna try running 40psi front and 38psi rear and see how it does. That way it doesn't go above 45psi when warmed up.

45-47 psi? wow about 3 mpg loss huh..

I'm gonna have to see what the middle ground is. What I really need to find out is what is the optimum warm tire pressure for best gas mileage. Then adjust the cold pressure accordingly. In the winter the warm pressures wont be as high as they are on a 60*F day. I think the sweet spot will be 45psi warm, or 40psi cold. Before my 30mpg run was with something like 50psi warm... definitely not safe, but surprisingly the car still handled well with those pressures!

 

Anyway, testing a new timing map, this one has a TON more timing down low.

http://i160.photobucket.com/albums/t188/covertrussian/Cars/05%20LGT/ECU/Timing%20Advance/05LGTCruiseTimingV1.png~original

 

Compared to stock map:

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

 

This got me 26.88mpg, but I went to a different city, had more standing on stop lights etc. Plus the changes were mainly around city cells and not highway ones. Learning views look clean, but will monitor it for knock and see if my calm driving city MPG can improve from 18mpg.

Edited October 13, 2014 by covertrussian

I know that on my car it can go as high as 49 deg of timing while cruising. stage 2.5.

Good to know, I wasn't sure how much is too much. Will try going even higher, might need to reduce AVCS from 10* to 5* at that point.

I rebuilt all of the calipers yesterday due to a stuck piston. Had to go to DC area, around 130miles each direction. I usually fill up once I get into the area.

 

Going there I got 29.03mpg with average ambient temp being 60*F, this is up compared to last couple highway trips. Coming back I got 26.67mpg with average ambient temp being 50*F. Which is odd because returning I usually get better mpg. Tire pressures pressures were within .5psi between cold and 1psi warm, thus I don't think they are a factor.

 

How does this tie into ecu tuning?

Every time it would get below 60*F my gas mileage would drop, come to think of it I've never gotten more then 27mpg when temps are below 60*F. Car felt a lot more powerful just cruising at 70mph and would lunge forward when accelerating with cruise.

 

I got some logs to parse, but will have to do a warm day log with same parameters to see how different the load looks.