Getting the most out of lean burn (Honda)

[greasemonkee]

I've had this idea for a long time that our civic leanburn engines didn't have to be handicapped due to emissions, I believe they have a lot more potential than the EPA allows them to have.

The subject will be the 92-95 Civic VX, D15Z1 or D15 vtec-e equivalent, that is what I'll be doing here. The idea is to maintain a leanburn under all - or almost all - conditions except for mid and max power. Unfortunately, the US P07 ecu limits leanburn to 2600 rpm and the only way out of that is to find a UK version P07 which is extremely rare.

A full retune with another engine management system would be best, but aside from the cost I do foresee some possible implications - a monetary risk would be at stake.




Goal -

1. Allow leanburn to activate at approximately 140 degrees coolant temp and in all gears; 1st gear leanburn is blocked and 2,3, and 4 gears have coolant temp restrictions.

2. Leanburn at idle and when stationary.

3. Allow leanburn to remain engaged at nearly ambient manifold pressures.



To achieve #1 the ecu must see the car as being in neutral and coasting - neutral uses same tables as 5th gear.

To achieve #2 the ecu must see the car as being in neutral and coasting, the TPS sensor must also output a off-idle signal, but must be switchable back to idle values in order for the fuel cut to activate while in gear. Additionally, a greater volume of air must pass through the throttle body to maintain a 500+ rpm idle speed due to the higher manifold pressure required under a leanburn condition.

To achieve #3 a false map sensor signal must be sent to the ecu. The ecu must see a lower manifold pressure than is actually present. The ignition timing will be advanced because of this - every manifold pressure value has a corresponding ignition and fuel value. Distributor compensation may be required and closed loop fueling may need compensation due to potentially exceeding the ecu's limitations on the max percentage of O2 correction, this will be corrected by increasing the fuel pressure. Fueling under full throttle is a lesser concern since the air/fuel ratios are already 12.3-12.6. This engine will produce more power more efficiently at approx 13.5 AFR with its low compression, but with the onboard devices this will be trimmed to its optimal value further lowering the BSFC.


Just for the record - here's a quick and crude vid I shot on the route to establish a baseline for the STOCK leanburn limits. EDIT - I lost this video...



UPDATE EDIT - Here were the results -->



Achieved:

*Leanburn in neutral and stationary while throttle is depressed -

http://s12.photobucket.com/user/grea...606d9.mp4.html

*Leanburn in 1st gear

*Leanburn in all gears (rather than 5th only) at cool engine temps - engages about 1 mile sooner than before.

*Leanburn at/near ambient manifold pressure

*More usable torque in leanburn

http://s12.photobucket.com/user/grea...9efec.mp4.html

*VTEC does not engage until 3000 under any condition

*Max power air/fuel ratios can be held at 13.5 for optimal efficiency.


http://s12.photobucket.com/user/grea...d9905.mp4.html


*A solid 2mpg gain.



----------------------------------------

Added parts - Apexi fuel controller, ebay square wave generator

[greasemonkee]

Beginning with #1 - to make the ECU think the car is coasting we need to feed a false vehicle speed signal to the ecu. The signal is a pulse type wave or half of a square wave. Here is a variable frequency and variable duty cycle square wave generator with a diode in line to block half of the wave. Now a 12v+ pulse signal can be sent to the ecu.









EDIT - This particular wave generator isn't really a full wave it's a half wave already, I didn't realize this until I saw it on the oscilloscope. There is no need for a diode.

[turbothrush]

Wow greasemonkee truly great stuff .Thanks

[WD40]

Wow I have to follow this, it could have a dramatic effect on overall VX MPG
Question tho .. why no gas log for your car

[Daox]

I think you're probably going to run into issues doing any of those modifications. There are reasons lean burn isn't used all the time and its not just emissions related. You seem to know what you're doing, but I have a feeling you're going to run into misfiring problems.

[turbovr41991]

There have been many Honda "tuners" that claim to have their cars tuned to run more lean and get 50+ mpg. Not sure how valid these claims are, how reliable the cars are afterward, or what long term effects or problems this might have on the engine. Eccentrically make any car a "lean burn."

One would think if safety and emissions allowed it, all the car manufacturers would be tuning their cars this way.

[greasemonkee]

Spoofing the ecu with a false VSS signal is successful, but crude. For some reason the VSS acts as a current limiter of sorts so the signal had to be routed through that before going to the ecu. I'll be trying to eliminate having to carry around a spare VSS under the dash to make this thing work.

Now we have leanburn in 1st gear and at idle and as low as approx 600 rpm with the throttle slightly depressed.

http://s12.photobucket.com/user/grea...606d9.mp4.html

[serialk11r]

Cable throttle...you're going to be stuck in lean burn the whole time? Or do you plan to have a lean burn off switch? Acceleration with very lean mixtures will be less efficient.

I feel like the best way to do this would be a wideband O2 sensor and standalone, that way you won't need to remember to manually put it on the original tune to make sure the fuel trims stay accurately calibrated. Along with a button for map switching of course, since your power will be quite limited. Lambda at full throttle is usually 0.9 (typically lower though, to protect catalysts and stuff), so if you're going to run lean your lambda should probably be no less than 1.3ish to reduce exhaust temperatures.

[greasemonkee]

Originally Posted by serialk11r

Cable throttle...you're going to be stuck in lean burn the whole time? Or do you plan to have a lean burn off switch? Acceleration with very lean mixtures will be less efficient.

I feel like the best way to do this would be a wideband O2 sensor and standalone, that way you won't need to remember to manually put it on the original tune to make sure the fuel trims stay accurately calibrated. Along with a button for map switching of course, since your power will be quite limited. Lambda at full throttle is usually 0.9 (typically lower though, to protect catalysts and stuff), so if you're going to run lean your lambda should probably be no less than 1.3ish to reduce exhaust temperatures.

The leanburn idle feature will need to have a manual on/off of some type in order for the fuel cut to work. The reason for leanburn at idle and 1st is for city driving at low speeds or briefly stopped. To accelerate most efficiently to 40-50 mph I don't use leanburn anyway and is easily overridden by depressing the throttle more.

The lambda at full load with this stock ecu/engine is 1.0 up to 2700 rpm then it goes rich to .85 lambda. Only rare WOT occasions does rpm exceed this. Leanburn is 1.55 on a stock VX, that's probably about where it needs to be.

Tuning this type of engine with a standalone system will require a minimum of 3 fuel maps and preferably two different closed loop lambda target values aside from having an egr valve to control. I wouldn't say this is insurmountable, but certainly a load of work if there is even a system out there capable of doing this - if so, I'm unaware of it.

[greasemonkee]

Alright, for a brief update -

I hooked up the square wave generator to the oscilloscope and it turns out the device is fully capable of satisfying the needs of the ecu by itself without having to add any components to the board. With a cold engine under 42~ mph the ecu will allow leanburn to activate only in 5th gear in its stock form, now I can hit any gear, even idle, but the ecu is limited to 2000 rpm until full operating temp - nevertheless, enough rpm to get down the road and save fuel until it's warmed up.


I've installed the Apexi VAFC controller and removed 5% fuel for a quick test run. That 5% allowed a 2"Hg manifold pressure, what a great sound through the throat of the microbeast. I may have to adjust the timing later and play with the settings some more. I see no skipping or any adverse effects thus far. I have hope, will report back as the project progresses.