Testing Lean Burn on my RX8.
 

I'm currently working on tuning for lean burn on my Rx8 with the end goal of 40+ MPG from a stock MPG of around 20. Here is a video on why I believe lean burn will work well on the rotary engine, and how you gain efficiency as you go lean.

Yes I could start with a different engine, but where is the fun in that! I'll post more as I get more data.

I done a few different tests so far and I will be posting video and results as I get it edited. So far I've seen a 20% reduction in indicated fuel consumption in idle, but I have not run the test long enough for solid data, nor is data at idle that meaningful anyway. I will be live tuning in cruise when I can get everything set up and recruit a helper.

https://www.youtube.com/watch?v=FnXWxXKPnH4

Those things like to run rich at least under boost.

Great vid!!!

Looking forward to your findings. I have been running lean burn for about the last 10 years on a few of my cars. My most recent is my highly modified Talon. Check it out some time.

Do these things inject oil with the fuel, like a two stroke engine? If that is the case, make sure to modify the oil supply to get the same amount of oil as before lean burn.

No they would never meet emissions even in china if they injected 2 stroke oil into the combustion process.
I have seen RX7 engines that were modified for racing to use 2 stroke oil, they smoked like a gigantic 2 stroke chain saw or weedeater.

They do, but a lot is unknown about tuning. I plan on finding out if there is a range of timing where they like to run lean as well
They don't inject the oil with the fuel, but they do inject it directly into the combustion chamber. In fact some people modify the oiling system to use two stroke (as it burns cleaner).

I have complete control of oil injection with my tuning setup, so it's not a worry.

Do any of these words embarrass you... Wankel Rotary Engine?

Just be carefull these engines don't seem like they stand up to detonation and pre-ignition as well as piston engines.

Rotary engines don't like sustained heavy (audible) detonation, but mild detonation at light loads isn't a big issue. At least I haven't found it to be a big issue.

I am running ion sensing for both detonation detection and PPP detection. And it can easily pick up mild detonation. I'm setting it up to both pull timing and reduce boost pressure when detonation begins, I may set it up to add more water as well.

Here is the actual first lean burn test. As you can see it does reduce fuel consumption at idle, but it does not match up to my mental model. Need more research.
https://www.youtube.com/watch?v=ft2Gn68doRM

Wait, vacuum leak? That shouldn't make your engine run lean, the O2 sensors will detect it and compensate.

You need a wideband O2 and narrowband emulator to trick your engine into running lean, and then redo the fuel map so it's running rich under load so your engine doesn't blow. You probably want to add some extra premix to the fuel to help lubricate the engine.

I would not get too greedy, and try to get it working at something modest like 1.1 lambda. At 17 AFR, my FR-S engine was not running too great, and that thing has 12.5 compression ratio and port + direct injectors. The 22:1 Honda lean burn engines had a lot more tricks up their sleeve to make it work.

I think your idea about the spark plug is interesting though, it had crossed my mind before. Hope you get it working.

What are you using to measure EGTs?

The Vacuum Leak is causing it to run lean because I reprogrammed the engine and forced it into open loop. I'm using the stock WB 02 sensor for both AFR and EGT. The stock NB sensor is disconnected because it would not let me lean burn the engine.

I can reprogram the engine for lean burn in closed loop, however that is limited by the range of the WB 02 (about 20:1 AFR) and I can't currently change the AFR on the fly.

I think it will have drastically different results when cruising under load.

Interesting, the ECU is okay with the NB sensor being disconnected? My car also had WB sensors + a narrowband, and the narrowband would kick in at cruise to correct my lean biased WB, foiling my lean burn attempts.

20 is plenty, don't get greedy. In fact the sensor accuracy at 20:1 is not that good, I wouldn't go over 18. Yea, you don't save as much fuel, but there are rapidly diminishing returns after that point because combustion temp and stability fall.

I had to silence some CEL codes, but otherwise the ECU doesn't care. At about 16-18:1 the flame front begins to break down leading to worse fuel economy as you push further; however, if you can avoid that fuel economy will continue to increase.

I'm assuming the mechanics of the rotary engine will let me get beyond 20:1 without the flame front breaking down, but it remains to be seen.

Also you will need to advance timing as you lean the mixture if you want the most gains.

Under load is where it gets real. I don't even bother with lean burn tuning at a idle.

I run a lean idle. It just kind of harder to start and keep running for the first minute or so on cold mornings until the hot air can kick in.

Idle is where you'll see the biggest improvement % wise if your engine doesn't start stumbling. I saw 10% fuel economy increase from going 10% leaner (MAF reading stayed identical, but AFR was leaner) on my FR-S. The 12.5 compression ratio and stratified charge from the direct injector probably help make that possible.

Turning the air conditioner on in neutral, the MAF reading also barely went up at all, so I was saving another 10% on the marginal fuel usage of the air conditioner.

On my car at idle I save only 2% going from 14.7:1 A/F to lean burn 20:1 A/F.
At idle I can't run any leaner do to the lack of turbulence in air flow.

At 55mph I save 29% going from 14.7:1 A/F to 33:1 A/F. Plus the amount of fuel weight used at 55 mph compared to idle is massive.

My best savings at idle is when I shut the engine off. So I turn my engine off a lot instead of having it burn fuel at idle.

I tried that last night...

Just got done making a chemical based hydrogen generator and popping hydrogen balloons. Yes I still have eyebrows.

Tomorrow I'll be road tuning/testing lean burn. In a few days I'll be testing hydrogen injection.

Road testing was a bust. Could not get the car to run right in lean burn, I think there were other issues. Re-flashed to try some power runs and then blew out the wastegate flange at 8psi.

So... No luck so far.

Here is the video of my attempt at live tuning lean burn. I didn't realize it at the time, but I didn't set lean burn up to a high enough load so it kept kicking to normal AFR and surging the engine. I'll do better next time.
https://www.youtube.com/watch?v=eeH1fzim4Kc

This is a issue I had when first attempting a 25:1 + A/F. I ended building a lean burn hole in my fuel map with my desired 30:1 in the middle and a leaner value on the outer perimeter to keep the engine from hunting to get out of the lean burn hole. One other thing I found out is I could only build one preset light load lean burn value in the map, because of the space it would take up.

My heavy load 10 to 14psi would run at 20:1 A/F with no knock on pump 91 octane.

Here is a pic of a light load 53 mph 30:1 A/F. The A/F shows 18:1 because this is a AEM wide band and the max value is 18:1. The mass air an fuel show around a 30:1 A/F ratio. This screen shot is a earlier version with out my pre heat intake setup and EGR enable. I'm now running as much as 2psi at this same mph.
http://i147.photobucket.com/albums/r...ps6c6c143b.png

I'm gonna have to either go that route for future testing or add a hot air bypass to the MAF and run open loop. I'm actually leaning (no pun intended) toward a hot air bypass because it gives me preheat on the air, but I have to have enough of an air path to make it work.

Any updates?

Still alive, currently working too much to get anything done. (60+hrs/week) Getting a week off soon, so hopefully more then.