Lean-burning engine... (what vehicles have them?)

[Arragonis]

I think we only got the V8s, certainly in the UK. They were sold on novelty more than anything.

Still interesting gestation of names, ecotec being plastered on everything.

[texanidiot25]

"Instead, Holden manufactured their own 3.8 litre V6 engine based on a Buick design, adapted from front- to rear-wheel drive."

"The latest revision of the Buick 3.8 litre V6 engine was fitted to the VR Commodore, featuring rolling-element bearings in the valve rocker arms and increased compression ratios"

Holden Commodore - Wikipedia, the free encyclopedia

And so on, even a supercharged version of the V6 around the same years that GM used it on the 3800 in the US, along with similar power ratings.

Sho' Nuff, just who ever wrote the 3800 info didn't put the Commodore in, or didn't consider it to be the 3800 due to the holden changes.

[Phantom]

The biggest change on the Holden Commodore from the 3800 Series II is that the Holden ecotech was OBD-I even though GM did not make the Series II with OBD-I. (It is possible that the L67 or 3800 SII supercharged did use OBD1.5 in the 1995 Bonneville change over but I'm not %100 on that.)

I have found the programming for the Ecotech 3.8 v6 but being OBD-I i think will make it harder to locate the lean cruse portion of the code. I guess its time I start to learn assembly.

[mpgx2]

Rather than trying to hack the ECU I suggest that it
might be easier to piggyback it.

If you have an mpguino installed, then you'll already
be reading the injector pulse lengths.

It's my pet theory that if you can read the injector
pulse widths then you can make a daughter board
to intercept them, and then reduce to time of the
injector pulses, thus leaning the motor.

One needs some special circuitry.

But if an injector pulse is 10 milliseconds and you
want to lean by 20%, output a signal of 8 milliseconds
to the injector.

The ECU will never know..

I'm told by my mechanic that LPG gas totally cut
the injector signals entirely leaving the motor to
continue running.

Anyway, I am myself working through the issues.

[texanidiot25]

I've heard of some people running piggy-back'd ECUs with a stand-alone unit (for lack of a better description). Main reason is so that the OBDII computer is still there for when it comes time for emissions testing. Otherwise, the aftermarket unit carries the engine's duties for horse power.

I would figure you could setup the aftermarket unit to run lean, and just have it wired to a switch I suppose. Easier than having to manually adjust the a/f ratio or monitor the mpguino. You could map an entire lean burn mode for the whole RPM range and have all the backup systems needed to protect the motor.

[mpgx2]

That's exactly what I think too...

I'm doing it as an open source project. I have two spare engines
to test on. A fiat and a vw motor.

The project is posted at:

djlyon / mpgx2 Vapour workbench / wiki / Home — bitbucket.org

I've only just started.. so a while before it's going fully..

[Phantom]

I have no issue editing the ECU as I have done that currently using HPtuners and a secondary program TinyTuner that has additional tables mapped that HPtuners does not have. The issue is the lean burn portion has not been mapped if I can locate it I can easily add it to the program to edit it when I want.

The issue with piggy-backing it is that if I would do this the ECU would try to learn out the "trick". So the Fuel Trims would be pegged at +15-20 due to the skewing of the injectors and the ECU trying to get back to 14.7:1 AFR. This could be corrected by also scaling the O2 sensor at the same time so it thinks that its is at 14.7 or replacing the O2 with an O2 simulator so it thinks it is always at 14.7. If I would go the sim route then I will need to get a wideband O2 and change the fuel maps when the weather changes a lot and when gas changes to the winter blend.

I currently can/could set the ECU to do lean cruse by sending it into Speed Density mode (The MAF is not used) and editing the VE table. The biggest problem with this is that the fueling will be based on the RPM and Load only. That could work but I will get burst knock when there are throttle transitions or if I need to accelerate quickly from a cruse cell.

I would like to be able to run it based on TPS% delta, MPH, RPM, Load, and time with an EGR factor. Also to edit the timing when in Lean mode currently I run 45-50 degrees advance when cruising.

[mpgx2]

Phantom, you have an amazing wealth of information.

I've never seen hptuners before but looks incredibly interesting.

Hacking the ECU doesn't work for me personally because my
car is too old and I have too little brand loyalty to make the
learning exercise worthwhile.

btw I am told you can just disconnect O2 sensors and the
car gets easier to work on. I actually figured out the O2
sensor on my car must have died ten years ago.

To run mine lean, I will replace the O2 sensor with an EGT
sensor and then work on leaning out way beyond the 20-30%
boundaries set by the ECU.

My target is to even try down in the vapour range at about
10%. Eventually...

[rmay635703]

Originally Posted by mpgx2

Phantom, you have an amazing wealth of information.

To run mine lean, I will replace the O2 sensor with an EGT
sensor and then work on leaning out way beyond the 20-30%
boundaries set by the ECU.

My target is to even try down in the vapour range at about
10%. Eventually...

If you get too lean and it runs you tend to get detonation, although it may not make tons of heat it will blast little pockets in your pistons

[mpgx2]

I have two old motors to test on. My main car has a 300hp 3SGTE
motor which is good up to about 700hp I'm told. It should be
fairly strong.

Actually, I hate heat already. It damages the shiny paint on the
bonnet.

I'm testing a water injection system. Once the temperatures
and leaning are under control I will switch to vapour.

Not new ideas, just stuff from the 1920's that is fun to
rediscover.