Can't see your image.

je suis charlie

I also can't see your table. But simply stating "... 12.5:1 at any RPM..." with no mention of load value is not very meaningful. At full load a ratio of less than 12:1 would be typical, and at light load stoichiometric (14.7:1) usually works well.

So sorry dropbox failure as usual.



12.5:1 was a general range I have seen for most NA applications. And I was wondering isn't it better to go richer or leaner to run more timing. Seems 12.5:1 gives me the highest flamwfeonr peopegarion speed.

Flamefront propegation *

Here's a related chart, laminar flame speed vs Fuel to Air Ratio (Not AFR):

http://oi59.tinypic.com/7307ph.jpg

From the general description you gave, I'm thinking this is a Subaru flat-four.

This engine was designed in the modern emission-control era, and it's designed to run stoichiometric almost everywhere, except nearing full load where it will likely need to be a little richer - as you have shown in your map.

If going richer in the mid-range induces detonation, don't do that.

If NOx emissions are not a problem for you ... try continuing the trend that you have established by leaning it out a smidge further at light load. Perhaps in the 15.0 - 15.5:1 range. It might want a little more ignition advance.

"Lean cruise" was a common strategy before NOx emission standards put a stop to it.

The part load area will deliver better fuel consumption with leaner mixtures (Lambda 1.1 - 1.2) and additional spark advanvce to suit, but NOx emissions will suffer.

The "diagonal" yellow line on your chart is normally steeper - ie mixture is more "load dependent" than "speed dependent".

je suis charlie

thank you for the replies. unfortunately 1.0 lambda is the leanest I can go. or in other words subaru ecu doesn't allow lean mixtures. rich or stoich only. I found there is quite a good gain of torque if you only run 0.2-0.5 AFR richer than stoich. that's why i never run exactly 1.0 except for cruising speeds (that's done via the closed loop table). I wrote a paper on laminar flame burning velocity which i will summarize below this.

basically what I found is the closer you are to 12.5:1 AFR or 0.85 Lambda the higher the burning velocity and hence more torque. Also loss of burning velocity and torque is non-linear the further you go from 0.85 lambda (straight gas). but at this ratio the tolerance for timing sharply decreases so you end up losing power. I found something like 13.7-13:1 AFR makes for best power since it allows you to run more timing.

---here is the summary of what i found in regards to AFR and laminar burning velocity----------

What is it
Laminar Flame Burning Velocity is the speed at which an un-stretched laminar flame will propagate through a quiescent mixture of unburned reactants (Wikipedia). This is NOT to be confused with stoichiometric fuel mixture.

Background
I have learned so much in past couple years tuning subarus so I thought to share my experience with tuning or dialing in your air fuel mixtures correctly for best performance (or at least to know what changing your AFR exactly does). Note some of these points made may already be known and I'm in a no way an expert, I just have some education in this area and personal hands on experience.

First off, the most important factor in determining your Open Loop fuel table is the fuel you are going to be using. Different fuels have different burning velocities and properties but in this thread I will be strictly talking about Straight Gasoline! Straight Gasoline experiences its fastest laminar burning velocity at roughly ~1.15 equivalence ratio which is about ~12.5:1 AFR.



In SIMPLE terms, what this means for an average tuner is that tolerance for ignition timing falls as laminar velocity increases and more timing is needed as the velocity falls. So one would in theory need the least amount of timing at ~12.5 AFR due to good burning velocity. From the graph above, going leaner or richer than this value starts to drop the velocity.

Tuning for the Perfect AFR
So as a tuner it's your responsibility to determine where your fuel ratios should be (rich of ~12.5 or lean or perhaps at the fastest flame speed itself). From my experience ratios in mid 12s always led me to retarded ignition timing. I have found ways to improve power by slowing the flamefront so i could run more timing or boost and as I said it's either going richer or leaner. Going richer such as 11.8:1 or leaner 13.4:1 allowed me to run more timing knock free. However we all know heat rises considerably as we enlean fuel mixture. In most turbo applications running a 13.4:1 AFR can lead to dangerous Exhaust Gas Tempratures (EGT) under boost. Even still most factory turbo Subaru cars run this type of ratio or leaner at midrange even under peak boost! I can see Subaru doing this for emissions and fuel economy since Gasoline/Petrol's stoich ratio is at 14.68:1. However such lean factory fuel ratios in low 14s can generate too much in cylinder heat and possibly cause cracks inside the engine and even damage the turbo*. So I think it would be best to instead go on the rich side (anything richer than 12.5:1) and still keep the same laminar flame velocity as stock lean fuel ratios.

This is how:

If you look at the laminar table graph above you will see that an eq ratio of 1.05 (14:1 AFR) has nearly same burning velocity as eq ratio of 1.22 (11.5:1 AFR). Therefore by going to a rich AFR of 11.5:1 vs. The factory 14:1 at peak boost you will keep the same laminar velocity that will net you (THIS IS ASSUMING CONSTANT TIMING AND BOOST):

- same torque and ultimately power (roughly)
- allows you to use same ignition timing and boost tables
- much cooler temperatures across the board (egt, in cylinder, turbo, catalyst)
- more reliable

However running richer has its downside too. You increase fuel consumption and hydrocarbon emissions. Your power per amount of fuel (not Power itself) also suffers. This is called Brake Specific Fuel Consumptions and Turbo engines are known to offer poor BSFC compared to NA engines in part due to rich fuel ratios.

Of course I'm not saying 11.5:1 AFR is the perfect air fuel ratio, I just picked this number since it had similar laminar properties as the stock fuel tables which runs lean usually up to 4000 RPM. For lower octane gasoline and higher boost levels than factory one might need to go even richer than this to control detonation as sudden throttle openings and boost can causes temporary leaner conditions than 11.5:1 which could put you in high laminar velocity areas causing knock. Of course in tuning there are various ways to combat this such as tip in enrichment but I'm not writing this thread to get into tuning specifics.

Conclusion
So in short I can sum up this post in to the following:
[List]
~12.5:1 AFR gives fastest Gasoline burning velocity
[/list]
[List]
Rich or Lean side of ~12.5:1 AFR slows burning velocity
[/list]
[List]
Higher burning velocity requires less timing and/or boost
[/list]
[List]
Therefore lower burning velocity requires more timing and/or boost
[/list]
[List]
Going too rich or too lean will cost you power no matter how much timing or boost you add
[/List]
[List]
Perfect Fuel mixture for turbo applications would be the one that let's you run near MBT timing and good boost without diminishing the burning velocity too much (there's a balance)
[/list]

- All these conditions make finding the perfect AFR impossible but you can get real close

For starters I would start with factory fuel table and enrich mixture (reduce burning velocity) according to load (boost) and not engine speed (RPM).
At same boost and ignition timing (two of which have greater impact on power than reasonable AFR levels), if a 2.5 g/rev engine load is targeted with 11.8:1 AFR a 3.2 g/rev load must experience a slower burning velocity of 10.5:1 AFR regardless of engine speed. Of course you could do this on the lean side (14:1 to 13.4:1 AFR) as the factory tune does this (except its more RPM biased than load) but expect dangerous in-cylinder and exhaust temperatures. Personally I target 10.5:1 to 11.8:1 ratios under boost for high octane gasoline and found this to lead good results. Any richer and power starts to drop dramatically and any leaner you will experience detonation.

*note that NA engines can run leaner than 12.5:1 ratios fine under load since they develop much less pressure and heat. In NA tuning practices running relatively lean ratios like these allows for combination of good power and economy or brake specific fuel consumption (BSFC).

Just askin' - what is different about them that the Subaru would prefer to be rich (12.5) at cruise than the more usual 14.7 - 16:1 that many other current engines run?

jack vines

i'm not sure but my and factory subarus are tuned to run at 14.4-14.7 in closed loop as posted above. who does 12.5 for cruise?

Hi Pickler. Have you looked at the current F1 regulations (Fuel-Flow-Limit 100kg/hr) and thought about what mixtures they might be running? Clearly with a fixed fuel rat, best power will coincide with best BSFC.

My hypothesis is that full power AFR is somewhere between lambda 1.2 and 1.5. It depends a little on whether you calculate AFR on the trapped air mass or total air mass since such engines would have a fairly high scavenge ratio.

There is no historical precedent for such lean mixtures but aircraft engines running at best BSFC (but still with substantial boost) have run in the region of 1.2. Current F1 engines are DI so stratified charge is a certainty - allowing overall mixtures that are leaner still.

je suis charlie

@gruntguru

hey thanks for the reply. no i haven't looked at it. I would be super interested to find a F1 fuel and timing table. Of course such info is mostly secretive. My only guess for going lean is to reduce fuel consumption. otherwise rich is just the same strategy albeit more safe and also with less NOx pollution. Unfortunately with most modern cars you can't really lean the ratios that much. I have tried it and it's only suitable for cruising but I honestly didn't see any change in fuel consumption with timing locked to full advance (50* btdc).

For efficient operation significantly lean of stoich, the combustion system basically has to be designed accordingly, from the ground up.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

Depends what you mean by "significantly". Any standard combustion chamber will see best BSFC at approx 10% lean (lambda = 1.1). For most modern PFI engines it is more like 1.2

je suis charlie

I played with this on a motorcycle engine with a DOHC pentroof layout. Stock setup was obnoxiously rich (with air injection into the exhaust, plugged for testing), sometimes showing lambda = 0.75 at steady cruise, and had poor fuel consumption and range (hence, me tinkering). Leaning it out to about lambda = 1.05 to 1.08 at part load gave good driveability and fuel consumption. Any leaner than that wasn't appreciably better for consumption, and lean misfire was near 1.2. The method of adjustment (PowerCommander) doesn't have sufficient throttle position resolution at lower speed and load to accurately maintain stable lambda; I couldn't really set it up for leaner than lambda = 1.08 without having occasional excursions too lean and getting driveability problems (and running too hot).

The problem with the throttle position is that the PowerCommander mapping has map cells for 0%, 2%, 5% throttle position and so on, but with this bike (zx10r, 160-ish hp) 0% throttle according to the TPS could be idle, or no load at 5000 rpm, or 50 km/h steady cruise in higher gears puttering around town, or shut-throttle coasting, and obviously these are different conditions and the real throttle position is some differing fraction of a percent, but the coarse mapping of the PowerCommander doesn't recognize the difference.

My cruising range or any operating range in those red cells above are at near lambda 1.0. I have tried reducing lambda 1.0 to 16.5:1 AFR or this lean operation as mentioned but poor idling and cold start prompted me to go back to stoich 14.7:1 for lambda 1.0. Cruising and generally throttle response was clearly much weaker with this leaner operation despite running 50* of timing advance.

now i'm wondering for going leaner, do i need to do anything else other than add timing? will going leaner increase temperatures unnecessarily? I have a catalyst in the midpipe, will that cause an issue? I'm willing to try this lambda 1.2 under load as done on F1 engines but need to build a list of requirements and some hypothesis firsts.

i tried lean mixtures today by locking afr to 17:1 and advancing timing 3-6* accross the board. I did not suffer from any missfires. However at high RPMS i was experiencing severe timing retard due to knock. i think i'll stay with 0.85-0.9 lambda for WOT.

"..... I have tried reducing lambda 1.0 to 16.5:1 AFR or this lean operation as mentioned but poor idling and cold start prompted me to go back to stoich 14.7:1 for lambda 1.0."

In the early days of Fuel injection ( Bosch K Jetronic ) Bosch said something like this -
" Today’s standard concepts rely exclusively on stoichiometric mixtures for the operation of engines running at idle "

Are you able to assign a "special" AFR at idle ? Or are you forced to apply 1 lambda across the board?

Lean operation at high engine load will NOT be good.

It's pretty common for idle to not tolerate being as lean as the cruise setting. My zx10r won't idle smoothly leaner than around 13.5:1 air/fuel.

Don't try to apply a "blanket" lean setup, it will not work.

Quote (gruntguru)

Depends what you mean by "significantly". Any standard combustion chamber will see best BSFC at approx 10% lean (lambda = 1.1). For most modern PFI engines it is more like 1.2

No, I mean lambda 1.4 or higher.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

Yes I did enriched the idle. However lean mixture just doesn't work under load. Even with more timing than with rich mixture it suffers to produce any power. Lean under light load works but torque falls dramatically as well there. I think 15.5 Afr is best in terms of balance between economy and power for cruising. Since my car is equipped with EGR it does already dilute the mixture under light load and decrease combustion efficiency. Adding lean burn to this seems to 'suffocate' the motor.

Quote (pickler)

Yes I did enriched the idle. However lean mixture just doesn't work under load. Even with more timing than with rich mixture it suffers to produce any power. Lean under light load works but torque falls dramatically as well there. I think 15.5 Afr is best in terms of balance between economy and power for cruising. Since my car is equipped with EGR it does already dilute the mixture under light load and decrease combustion efficiency. Adding lean burn to this seems to 'suffocate' the motor.

See my post at 02:54am earlier today.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

Significant leaning will require significant additional spark advance. More than 50 degrees is not unusual. Most engines will reach the misfire limit at about lambda 1.2. Many will have detonation problems at these extremes. Engines designed for lean burn will have higher ignition energy, additional turbulence and sometimes stratified charge.

je suis charlie

Increased Charge Air Temperature, fuel temperature and smaller fuel droplet size can improve ignitability and combustion stability.

je suis charlie

When you added ignition advance did you add it in stages, like 2 degrees, between tests?
Did you add it to the stock or pre-existing curve, or just in a section of the rev range?

Did you end up with a load/rpm/timing map as varied as this?
http://www.911chips.com/USignst.gif

Yes I added timing in stages and it was load vs rpm based. I added more in lighter loads and slightly less as loads increased. I gave up after first stage due to loss of torque and detonation. I dont think the risk is justifiable.

I think I will stick with stoich or richer AFR tuning...14.5-14.2 AFR seems to give best torque without loss of BSFC for part loads. Above 3000 rpm at full MAP I start to enrich slightly to 14-13.7 AFR and keep it until peak torque. This gave me best BSFC rich of stoich partly because I could advance timing further than before. Also I can run 3-4° more timing than factory under cruising loads without detonation. I just wasn't sure about my method since it's leaner than factory and common practices for NA.


Total Ignition Timing

Factory maps are often richer than optimum at full load. This provides a safety margin for things like - flow variation between injectors, airflow variation between cylinders, variation in operating environment and engine condition over the life of the vehicle, fuel quality variation etc etc.

Some of these conditions can be detected during closed loop operation and compensated by the Engine Management System - some cannot.

je suis charlie

laminar flame speed probably has little to do with actual burn speed in a real engine as there is alot of turbulence in the chamber.

If you look at how long it would take flame traveling at 350mm/s to go from plug to farthest point in the chamber say 40mm in an 80mm bore this would be 0.11 seconds.

Then work out how long it takes for an engine to do a full revolution at 6000rpm (0.01sec) and clearly laminar flame means little

Quote (inline6)

laminar flame speed probably has little to do with actual burn speed in a real engine as there is alot of turbulence in the chamber.

If you look at how long it would take flame traveling at 350mm/s to go from plug to farthest point in the chamber say 40mm in an 80mm bore this would be 0.11 seconds.

Then work out how long it takes for an engine to do a full revolution at 6000rpm (0.01sec) and clearly laminar flame means little

well if it was like that then I could get away with richer AFR and more timing without knock wouldn't I? I'm definitely sure the knock i get approaching 12.5-12.7:1 area is due to higher flame speed. the stock ecu completely avoids this ratio suggesting the engineers who mapped this experienced this knock phenomenon as well.

The stock programming/mapping avoids anything other than stoichiometric everywhere they possibly can, because a 3-way catalyst can only function effectively at that specific condition. Emissions, emissions, emissions.

High load at higher engine RPM will have to be richer, and it can be, because those conditions are not encountered anywhere in any of the official test procedures.

The US RX-8 Mazda rotary is factory mapped to go rather richer than stoich under certain operating conditions that are not high load (such as higher speed cruise) in order to keep the cat cool enough to last 80K miles. My point is that there are sometimes reasons other than optimum engine performance and fuel economy that account for manufacturer's mapping strategies.

Rapid combustion and knock do not normally go hand in hand in fact usually the opposite. Slow combustion increases the time available to heat the end-gas. Rich mixtures suppress detonation by a reduction in the temperature of the mixture and an increase in the auto-ignition temperature, not by a reduction in flame speed.

How did the knock you experienced at 12.5 AFR respond to retarding the timing? 12.5 should really be a sweet spot for power and knock resistance.

Turbulent flame velocity in an engine is 20 to 40 times higher than the laminar flame velocity.

je suis charlie

The best timing at high load at 1000 rpm is 11 degrees AFTER TDC?

And, at 4000 rpm under heavy load it has to be limited to 22 BDC?

Did I miss that this is a turbo engine?

Quote (tmoose)

The best timing at high load at 1000 rpm is 11 degrees AFTER TDC?

And, at 4000 rpm under heavy load it has to be limited to 22 BDC?

Did I miss that this is a turbo engine?

No its NA. But its compact pentroof chamber with high swirl and fast burn plus its 10:1 compression head requires much less advance than your regular big block v8.the closest V8 I think of to this engine is the ls7. Mid 20s in knock free timing makes peak torque with this engine.

Quote (gruntguru)

How did the knock you experienced at 12.5 AFR respond to retarding the timing? 12.5 should really be a sweet spot for power and knock resistance.

I experience 90% of my knock problem at peak torque 3800-4200 rpm and at peak VE ~5300 RPM. At peak torque going any richer than 13.5 AFR gives me knock. The closer I get to 12.5 it seems the worst it gets. You can retard timing as much as 4° and it will still knock. At peak VE its not too sensitive to fuel and more so to timing. So anything higher than 25-26° at peak VE causes knock whatever your fuel ratios.

On the dyno approaching 24° starts to flatten the torque curve and any more timing the torque falls. I can run MBT on pump gas with the AFR table I posted above. With stock furl map which is richer (NOT stoich) I can Not come close to such timings. It's strange but it is what it is.

I said above that 12.5 should be a power "sweet spot". I should have said "13.5" - approx lambda = 0.9 which makes your results a little less surprising. Also unsurprising - this coincides with the highest flame velocity for gasoline.

je suis charlie

getting back on to the cruising and lean burn topic, how much timing should be added if you go leaner from stoich and what AFRs are typical or best? I remember reading somewhere that 15.6:1 AFR gave best BSFC for a PFI engine.

Ie. If I run at 15.6:1 vs 14.7:1 for cruising, how much more timing should be added?

Best BSFC mixture varies enormously from engine to engine. The 1986 Bosch handbook says "approximately 1.1". The 2011 edition says "1.2 - 1.5" (PFI) and "up to 4.0" (DI). The variables involved include engine design, load, air temp, fuel temp, droplet size . . .

MBT ign advance can only be determined on the dyno but leaning by 10% (say from 14.7 -> 16.2) at light load would typically require at least 5 - 10 deg extra advance.

je suis charlie

so it wouldn't be typical to add 10% more timing for 10% leaner mixtures?

I'm currently targeting 16.8:1 AFR which is 14% leaner than before. Adding 14% more timing would roughly equal to ~5* more timing.

New AFR Table:


Timing


note timing may seem to be over advanced at cruising ranges, this is because I performed part load tuning before this lean burn and timing was advanced until best torque. Actual timing added after lean burn was around 3-5* on top of previous tune or 14%.

Sorry if question is dumb, but why higher engine loads needs less timing advance?

Hi Pickler,

" timing was advanced until best torque.."
How was best torque determined?
Was it some kind of a "flash" reading?
Or, was the load maintained for at least several seconds?


What is the make model and year of this car?
What modifications have you done?

Higher load = higher intake manifold pressure = the residual exhaust in the cylinder at the end of the exhaust stroke means less, and there is less "wrong way" dilution of the flow which contaminates intake charge with exhaust during the overlap period. Low load on the engine means more "uncontrolled EGR".

Quote:

Higher load = higher intake manifold pressure = the residual exhaust in the cylinder at the end of the exhaust stroke means less, and there is less "wrong way" dilution of the flow which contaminates intake charge with exhaust during the overlap period. Low load on the engine means more "uncontrolled EGR".

So at high load cylinders have more "fresh" mixture and then why it requires less advance and less "fresh" mixture requires more advance?

Full load ... it is almost all "fresh" mixture in there. The amount of exhaust residual should be only a few percent.

Light load ... it is a mixture of "fresh" mixture and recirculated exhaust, which can be a significant fraction of the charge.

A mixture of "fresh" mixture and recirculated exhaust will slow down the burn rate and require more ignition advance.

There are some other temperature, density, and turbulence related effects, too. Lots of mixture inflow means lots of swirl and/or tumble in the cylinder. Very little mixture inflow will give less charge motion. How much that is meaningful, will depend on the combustion chamber design.

For lean burn I added 10% timing in leaner cells. Before lean burn I added timing until knock and backed off a degree or two.

"So at high load cylinders have more "fresh" mixture and then why it requires less advance and less "fresh" mixture requires more advance?"

Detonation is one pretty strict limit of ignition advance.
Best torque may, or may not, be produced by ignition timing set just before detonation occurs.

There are a dozen or so factors affecting whether detonation will occur.
As far as I know it boils down to whether the end gasses are hot enough to form radicals in a series of pre-flame reactions that lead to autoignition.
Maybe 900 - 1000F.
http://www.factorypipe.com/t_deto.php

I don't think ( the right mount of ) extra advance being best when cruising is just due to lean-ness. Other wise the ignition maps ( and simple vacuum advance) would not dip as the load changes but mixture A/F remains ~ constant.
there is the matter of "compressing" a rarified mixture, so the fuel molecules are 4 times farher apart, gas temperature at TDC being a lot cooler, the exhaust valve running less red hot, the piston dome and cylinder head being cooler, and they all

I started getting knock advancing low load cells to make up for lean mixture. I had to back up to previous timing. Which is still well advanced of stock settings.

I think 16.5 AFR with 40* timing is most efficient it can get.

Also could someone help me understand FUEL air ratio? Not Air Fuel Ratio (AFR). For example from this chart what is 0.07 FAR in AFR for gasoline?

just takeing a guess here as most of this is way over my head but I do enjoy "trying" to understand, wouldn't is be the inverse of the air/fuel ration, i.e. 14.7:1 is .068
???

Looking at your fuel table it appears the cells you have chose to lean out are medium load. One cell is in the 0.6 g/rev column which is almost 50% of the full load figure. Is this proportional?

Ignition advance is not something that lends itself to "percentage" increments. e.g. if timing is at TDC how much is a 20% increase.

je suis charlie

Quote (djhurayt )

yes i believe lambda 1.0 would be around 0.07 FAR if i'm not mistaken.

Quote (gruntguru)

Yes 50% relative engine load is correct. This is the engine load at which my car cruises on the highway. Could be due to the all wheel drive. This is why I don't think lean burn can work for me. At slower speeds my RPMs are too low ~around 1300-1500 and on the highway load is too high.

I'm having difficulty fighting knock with leanburn and making the same amount of power. The lean conditions at low loads 'make their way' to higher loads with sudden opening of throttle cause detonation. I think I'll stick with stoich/rich map for now which allows me to advance timing to MBT everywhere in the powerband excluding at low RPMs (1000-2000 RPM):




VE:


also on a separate note, my engine experiences higher loads than defined in this table. It's not unusual to see 1.7 g/rev at higher RPMs specially with colder temperatures. MODs are mostly on the exhaust and intake side, like headers and higher flowing piping. MAF has been re-scaled as well. I'm just happy I managed to kill detonation with an unusual way of leaning (closer to stoich). As said I experience knock near the VE peaks 4000 and 5500 RPM. Currently I'm not getting any knock.

correct timing table

volumetric efficiency over 100% with a modified street vehicle at 5300 rpm?
How are you calculating that?

============

I guess the load scale units are grams/per revolution passing thru the mass air flow sensor.
Are the values for the whole engine?

regards,

Dan T

I'm not calculating it. The ecu is estimating VE based on airflow from Maf and other sensors. Stock with no mods it was around 95%

MAF = Engine Load * RPM / 60
Example:
at 3534rpm * 2.91g engine load / 60 = 171g/s maf

I don't wanna create new topic, so i decide to ask here: if in cylinder are more air/fuel mixture(when i press accelerator pedal more), peak cylinder pressure also will be higher?

In the absence of complicating factors - Yes. (Why would you think otherwise?)

I don't think otherwise, i just wanna make sure that i thinking correctly :)