Lean burn tuning
 

Looking to have my car finished within the next week or so (whew, this is a much longer process than my optimistic self anticipated), at which point I'll have the opportunity to start tuning.

I'll start off with two articles I think everyone should read:

http://www.superstreetonline.com/how...economy-tuning

Honda's K20A Engine


~

My (soon to be) ECU is modified by Hondata with their "KPro" system. They have a wonderful piece of software called the "Hondata K series ECU editor" or just "KManager". The ECU supports live tuning and supports monitoring and modification over bluetooth and USB.

If anyone is interested in downloading it just to play with, it's free:

https://www.hondata.com/kmanager

https://i.imgur.com/0W4sY9d.png

https://i.imgur.com/MqbbWci.png


Anyway, to my point of making this thread:

I'm looking for some tuning advice from anyone who has actually tuned for lean burn. My understanding is that peak NOx (and combustion temperature) happens around 15.7:1. This is also peak efficient use of fuel under most circumstances. My (new) O2 sensor is accurate to around 19:1 AFR.

https://www.researchgate.net/profile...e/cograph2.gif


After observing what Honda did with the Insight's stock engine, which runs at 24:1 AFR sometimes, I've extrapolated a few things:

-When running lean, you need to advance ignition timing or the engine begins to skip. I believe this is because flame speed is lower, so you need to start ignition sooner. In the Insight, advance is 20-30 degrees higher at 24:1 than 14.7:1.

-NOx can foul a catalyst. Honda would very occasionally run rich for a few seconds to "purge" the cat.

From other articles I've read:

-Cam timing seems to on average improve fuel economy when advanced, but can't just be assumed. It appears that any changes to intake or exhaust can change resonance in the engine and shift points of optimal timing around.

-Going leaner may or may not improve economy past a certain point. Pumping losses continue to decrease, but BSFC might get worse (for a specific power output) past around 15.7:1. Or it might not.

The main factors I have to play with include:
-Air fuel ratio
-Ignition timing
-Cam timing (-10° to 40°)
-VTEC changeover (low vs high cams)
-RPM
-Engine load


It looks like the "correct" way to tune for fuel economy is to get a dyno that can simulate a partial load and to map out fuel consumption vs load at various RPM, varying ignition timing for each 10 degrees of cam angle. I don't have any good way to access such a dyno, unfortunately, and the process sounds like it could be very involved.

Luckily I *do* get to start with a base map for the engine. Unfortunately, none of the included base maps have been tuned for operation any leaner than 14.7:1, so I have nothing to extrapolate from. It would be nice to get ahold of a map from one of Honda's lean burn engines.

I'd love to hear others' experience with this. How did you do it? What were your results?

If lean will fowl a converter don't run a converter.
I tuned for lean burn on my 7.4l suburban and it worked great, pretty basic but it got the job done.

That’s part of the reason I’m ditching my cat too... that way I can eventually add lean burn

Antique 2 way cats like a diesel VW runs can’t foul and are cheap

While running lean a 2 way cat is the best way to decrease VOC, CO, particulate, just not NOx


Erm thinking about it further your running a pollution controlled car with pcv and exhaust gas intake intact, that chart is invalid with an operational pcv and egr

In other words My guess is your stock cat will live a long life and you should log the pollution after your tune, results will likely surprise you

Yeah that's true.
My 454 was not running EGR.

I did all my tuning on the road. I think you need to keep the tuning road only because you find out that your aero will play a big part in your maps verse a load type dyno.


Also make sure that you set your eco tuning in open loop if you don't run a cat. When you start running lean burn open loop is your friend. The engine in lean burn is really picky the leaner you go and needs a smooth fuel map. This will keep miss-fire from happening.


Try to keep leaner or richer then 15 to 16:1. Combustion will heat up the exhaust valves in this range. You should be able to be at 19:1+ for light load pretty easily. You will have to calculate anything over 19:1 do to your sensor. Pretty simple airflow verse fuel flow.


You will also learn what rpm the engine likes to be at in lean burn. Every engine has its own rpm that makes its best airflow swirl and tumble. That will also be your highest A/F you can run.


Anyway there is a ton more to it but these are good starting points.

I had some limited experience with my Scion FR-S, the engine seemed to have trouble at 17:1 or so. I didn't really mess with the timing much, it's possible that if I really changed ignition timing it could have worked. I gave up entirely when I realized I needed to cheat the post-cat O2 sensor, as it was correcting my wideband calibration while I was driving, so I could only run lean while idle.

I believe your Honda K24 should do a little better because the low-rpm cam has low lift. Still, I think you'll find diminishing returns near the 18 AFR range given your engine is stock. If I had a car and time/money to test it, my thinking is that a significantly increased compression ratio and decreased squish volume would help with lean ignition (using a PCV catch can can buy some knock resistance, and giving up a little timing and power at max load should buy a bit more).

In piston aircraft this is called Lean Over Peak

They lean out the mixture until they are well past the peak exhaust gas temperature. Running too close and too long to peak EGT will fry the pistons.

I admit I am a total noob to the subject, but I think it would be a good idea to add exhaust gas temperature instrumentation just as an extra engine health safety measure, and another data point to look at, not to mention a way of directly seeing just how lean you are truly running.

Not fair that aircraft get all that wonderful instrumentation, and we don't on modern cars. We claw back some of those gauges with scangauge and the like.

Aircraft run at WOT then use leaning out to set their speed that is very different than lean burn in a car at partial throttle

In Eckies case running at WOT at 15.5 -1 would likely drive him over 100mph

It will also burn and pit the exhaust valves.
I did this to an engine back around 2001.
Pistons were fine. The exhaust valves were fried.
But I made the mistake of lean at wot.

In my 454 suburban it was rich at wot, lean the rest of the time. Tore down the engine after 9 months of running it that way and the Pistons and exhaust valves were fine.

Don't get super tunneled on the AFRs. There's more to the picture than just the AFR that you can tune around.

If the engine is DBW you can rework the throttle maps to allow greater throttle openings while maintaining the fuel flow rate. That will allow you to adjust the throttle (read: airflow / VE) to where the engine will tolerate it.

Advancing the intake cam has less to do with resonance and wave tuning and everything to do with increasing dynamic compression. The intake valve is closing sooner, so you are compressing a larger swept volume into the clearance / chamber volume. This means there will be greater compressive heating, so mind the ignition advance.

Running lean has everything to do with mixture homogeneity and fuel sensitivity. The more heat you can put into the endgas the more likely you are to not misfire, although the closer you are to pre-ignition and detonation. Also, port flow turbidity and swirl will count. This is all dependent on flow velocity, which you'll be able to "see" in the VE map. Cam switching (and somewhat timing) will affect this too.

If your K-series has a full exhaust manifold, you can gain some improvements in scavenging and reduce the valve heating. That will take some effort. Beryllium valve guides and seats are for the bold ... haha.

Hey Ecky, thought I'd ask in this thread to be more on topic, did you manage to make lean burn work with only Hondata Kpro? Is it open loop lean?

I'm still eyeing an S2000, which I think doesn't have Hondata Flashpro available for pre 2006 models but a Kpro might be worth the investment. A wideband O2 emulator will get the job done but it would be nice to keep more of the stock tables.

I'm running closed loop lean, but I'm uncertain my ignition timing is perfect. I need to get either some dyno time or get my cruise working and find some long, flat roads. There's also a bug in KPro which is a bit annoying:

In KPro you need to set a target lambda upper limit - by default this is 14.64:1. You then have a table where you can set lambda per cell, at every given MAP and RPM value. If it worked exactly as it appears, it would be great. However, let's say I set a target lambda limit of 15.7:1, and one row on my table looks like this:

2000RPM
100mbar - 15.7:1
200mbar - 15.7:1
300mbar - 15.7:1
400mbar - 15.7:1
500mbar - 15.5:1
600mbar - 14.7:1
700mbar - 14.6:1
800mbar - 14.2:1
900mbar - 13.5:1
1000mbar - 13.2:1
(Just an example to show the bug, not exactly what I'm running)

I then set it to stay in closed loop until 900mbar.

What I'll actually get is this:

100mbar - 15.7:1
200mbar - 15.7:1
300mbar - 15.7:1
400mbar - 15.7:1
500mbar - 15.5:1
600mbar - 15.7:1
700mbar - 15.7:1
800mbar - 14.2:1
900mbar - 13.5:1
1000mbar - 13.2:1

What it seems to do is take any value on the AFR table around stoich and just use the lambda limit instead. E.g If I set 15.0 in a closed loop cell I'll get 15.0, but try to run 14.5-14.9 and it jumps back up to 15.7. 14.2:1 will, however, still work in closed loop.

My choices then are to:

1) run lean at a higher load than I'd like,
2) go slightly rich at lower loads where I really just want 14.7,
3) go open loop at a lower load than I'd like

I've filed a big report, but the feedback I was given from Hondata was basically "DO NOT DO WHAT YOU'RE TRYING TO DO, YOU'RE GOING TO DESTROY YOUR ENGINE."

Ah okay. I'd just go open loop at a lower MAP if you're worried. 700mbar is not that high though, I think your engine will be fine. I really don't think burning anything up is an issue, because engines running with cooled EGR have even slower burning mixtures. The peak flame temperature theoretically increases, but you have less and slower heat release so there isn't as much heat for the piston to absorb in the first place.

It would help to raise your compression ratio to provide a blanket increase in combustion speed to compensate for the slower lean burn but I suppose that's not an option.

For my future S2k I might just use a narrowband emulator and run modestly lean to make things easy. The issue with that engine is the stock tune runs at ~13 AFR at WOT, so there's not a lot of room to lean it out for cruising, maybe 5% at best, which is incidentally about where ignition timing needs to be changed before efficiency drops.

That's what I was thinking, but it's good to have that reinforced.


[QUOTE=serialk11r;621493]It would help to raise your compression ratio to provide a blanket increase in combustion speed to compensate for the slower lean burn but I suppose that's not an option. [quote]

The easiest way to bump compression would be with a thinner head gasket, but the piston-valve clearance is already extremely close. I'd probably want to replace the pistons with some that have deeper reliefs. So, not an option.


Seems reasonable!

I'd love an S2000 but prices are beyond silly. For similar money, I think I'd take a K-swapped Miata right now. Have your eye on one in particular?

Another way to think about lean burn and piston damage is that the momentary high temperature in the cylinder is higher in a turbocharged engine and MUCH higher in engines running N2O, but they take a long time to melt at WOT. If you're just running at part throttle lean, the peak temperature could be a touch higher, but the cylinder walls/head/piston will quickly bring the temperature down since the lower density gas has less absolute heat capacity.

If the tiny variation in peak combustion temperature mattered so much, then running hydrogen or straight gasoline instead of E10 would melt pistons too. It doesn't. The peak temperature of the flame is always much higher than the melting point of any of the materials in the engine, but the engine is spinning quickly and any part of the engine is only momentarily exposed to the high temperatures, so the average temperature matters more. Lean burn reduces the average temperature since there's less heat in the first place.

People also like to mention oil consumption, but again, IIRC engines running stoich already have ~2% unburned fraction, so there's already some free hot oxygen available to burn oil, you're just making that reaction go a little faster. I don't think combustion of oil is the primary way that engines lose oil anyhow.

Yea S2000 prices are silly, but I really should not be spending time dealing with modified cars, so I just want to find a bone stock car. Last year there were several cars with minor damage going for below 8k, I'm hoping I can snag one like that. With the economic downturn coming, I figure some dents and minor rust would help me blend into the proletariat :P

My experience:

- You need really, really good ECU control and mapping to do lean burn well

- Forget about worrying about EGTs in light load lean cruise - my measurements on the Insight 1 litre show they're nothing to worry about at all.

I would like to add if anyone ever visits here and gets this far...
Lean burn only works in low rpm conditions. if WOT than this is "lugging" in the extreme

the timing must be advanced and the rpms must be low because the flame front is mush slower
but if you are able to "complete" the combustion before the exhaust valve opens you will prevent valve damage and make useful work of energy.

Does the Insight have a ratcheting parking brake? Even if you wear it out, replacement parts are probably still cheaper than dyno time...

Had the opportunity to speak with Bob Krupa a couple times while he was still alive. He told me his '96 T-Bird was leaned out to 30:1 AFR -- EVEN AT WOT!! His Firestorm Plugs were able to ignite the uber lean mixture. Perhaps Pulstar Plugs, or Aquapulser's PDI would facilitate better lean-burn combustion.

It does! And that's a good point.

Enjoying this thread.

On my lean burn Talon I did some load base dyno tuning a few years ago and could never simulate the real world road conditions. It still amazes me how driving is a major factor to getting good FE. I have about ten different test roads with different traffic and terrain, with that said all have of them have different tuning maps.

IMO its just way to difficult to simulate actual road conditions on a dyno. In my case this applies to 20:1+ A/F on my car. I personally don't like being at 16:1 to 17:1 A/F, these A/F ratio's will make heat and I stay away from them. So my map goes from 14:1 to 20:1+ instantly.