AFR - What is Optimal ?
 

I've searched using 'optimal AFR', 'optimum AFR', best AFR', and many other permutations - I haven't found anything close to definitive.

So, for an EFI controlled non-turbo 911, for perhaps 3 ranges:
idle, acceleration (strong), and cruise - what AFR's should be targeted?

If someone has links or referrals to either pelican threads that deal with optimal AFR's or technical discussions relevant to 911's, whether opinion or theory, I would appreciate it.

The goal criteria here is max horsepower within reasonable engine potentials - not fuel economy nor testing breaking limits.

thanks,
Jim

chart here may be of interest-->

http://forums.pelicanparts.com/showthread.php?s=&threadid=246163&highlight=air+fu el+ratio

- Wil

stochiometric on idle and half-load, fuel-cut at high-vakum (engine braking), 12.5-12.8 at WOT.

AFR's at WOT are only one part of equation. There are lot of ponnies left in ignition timing once AFR's are squared.

A bump - Checking for any other opinions or comments, please.

Thanks,
Jim

Re: AFR - What is Optimal ?
 

Jim,
I think we can ignore max hp at idle, however can you define or clarify what you mean by "within reasonable engine potentials". Do you mean what we could reasonably expect in terms of horse power? Do you mean before something breaks, such as within the normal mean time between failures (if an automotible OEM would ever go so far as to state it clearly in a warranty).

Also could you define "acceleration (strong)". Do you mean from a rolling start on a flat surface in first gear for example?

Though a few hours on a good chassis dyno with an experienced tech would answer your question quickly, there are some good posts by Mr. Steve under the subject of chips and MAPS.

Hi Charles

By 'reasonable egnine potentials' I mean - probably in my inexperienced lame way - as an example: sustained WOT at redline +/- 5%. What AFR is considered optimum for that condition without testing the boundries of 'too lean'. In my case with 10.5:1 3.2 p&c's with high compression (around 185), 3.0 heads and carrera intake manifolds.

By 'acceleration (strong)' - I should just have said WOT acceleration. As to when - I was thinking the WOT acceleration AFR consideration would be the same whether from stop or upon up shift.

As to idle: there I was looking for a consensus or feedback on what a best AFR might be at a typical 1100 rpm idle. Assume for the sake of discussion a warm idle.

Finally, in regards to 'cruise AFR' - I should have defined that as typical highway cruising - say 80mph for discussion purposes. I would think the stochiatic (sp) would be targeted there?

Thanks in advance,
Jim

Lambda for NA:

Lean 1.2-1.15
Idle, part throttle, cruise, and economy 1.1-0.95
Under load/power 0.9-0.85
Air cooled full load 0.8

Hmmm.

Does this car have a cat? This could conflict with your perfect world ideal h.p.

Stoichiometric or 14.7 is usually thought to be a "perfect" burn with solely n-heptane and iso-octane, however you have to wonder what blend of gasoline we should be comparing "perfect" to today ..... with the various detergents, 'tane's', and simply our farm support alcohol blends.

At idle I would imagine I would only be interested in maintaining something that does not load up the plugs nor runs too hot, say 14.5-14.7 which seems to work just fine on our 87 3.2.

I think if I wanted to fine tune any car for it's max power, again I would find a chassis dyno and pay for a couple of hours of fun time. That will give you real results with real facts.... which always beats hearsay.

No cat: exit is SSI's with a Triad 2 in - 2 out.

So far, I can only get a decent idle at 12 to 13...

It's too late at night for me, "p&c's" are, pickles and cumbers?

How, where and what are you using to measure your AFR?

p&c's = pistons and cylinders
Electromotive TEC3r EFI for tuning including AFR

Thank you for key to that acronym. Though I was not too far off with pickles and cucumbers, it was much easier than
"NIIOMTPLABOPARMBETZHELBETRABSBOMONIMONKONOTDTEKHS TROMONT"

On your initial post I assumed you were looking for an optimal AFR for a "stock" engine, which is a broad range to begin with. Since you are running a higher compression with your new pickles and cucumbers you gotta different salad, something needing a stronger dressing for sure.

Logically when it comes to tuning you want to start off safe and simple, and maybe the new TEC3r provided you with that with a base fuel curve.

1) "Assuming" atypical 3.2 would idle at 950 rpms and you indicated you are at 1100 rpms and "rough", is there a reason why you are above the norm?

2) Have you determined why you are running rough at idle?

3) What type of range can you see with your AFR display and what model AFR display are you using?

4) Have you had this set up on a chassis dyno yet?

5) Do you have any charts yet with AFR's vs. RPM's or AFR's vs. HP and torque, and how were or what generated the charts?

6) What other ignition and fuel management components are not stock?

7) Where is your O2 sensor located and which model are you using?

12.8 under WOT.

Hello.

Let's just say a cat-equipped car should always run stochiometric except while engine-bbraking (it should shut off the injectors) or WOT (it should runt rich, lambda 0.83-0.85 is nice for aircooled cars, slightly higher for watercoolers with pentroof combustion chambers).

There is no reason to run lean, it will only produce extra NOx. There is no reason to injecto fuel into engine while engine-braking, it will just backfire.

While WOT, AFR's should be as near stochiometric as possible w/o risking pre-ignition or too high EGT's. Extra fuel while running WOT is only there for cooling. Aircooled engines (due to their thermal shortcomings) nneed slightly richer WOT AFR's than it's watercooled contemporarys.

Certain cars are able to run lambda 1.0 even at WOT. Aircooled ones won't though.

Stochiometric mixture means that there is just enough oxygen to burn all fuel. Actual AFR for lambda 1.0 will vary depending on brew, ethanol-content etc. Therefore, it's much better to just talk in normalized lamda lingo than assume that 14.7:1 always means lambda 1.0.

Lambda-sensor won't magically lock AFR's at 14.7:1. It actually doesn't sense AFR's at all. It only senses oxygen surplus and goes high/low depending on which side of stochiometric you are.

That's why it works regardless of fuel. If you use E85 (which has stochiometric AFR of ~9.0:1), lambda sensor will sense it and (assuming that ECU has wide enough margins ) adjust accordingly.

To round it up: unless you are running endurance race with state-of-the-art watercooled engine and have fuel penalty, you never run lean. You map it to run lambda 1.0 unless you are running WOT. It's as easy as that.

That way, a catalyst will work at it's best. Even if you don't have a catalyst, running lambda 1.0 will be a good tradeoff between NOx, CO and HC.

Running rich at half-load just wastes fuel.

Goran - thanks for the information and education. First relevant and important observation that the ethanol mixes lower the stochiometric somehow escaped me during this process. Shouldn't have, but did.
We're currently at the 10% ethanol mix (E10) in my location, rather than the 85% of the E85.

I don't know (can't find) what the stochiometric AFR is for 10% ethanol, but considering what you note (approx. 9.0:f`) for the E85, the stoch AFR for the 10% ethanol must be a whole lot less than the 14.7:1

It would make sense then, I think, for my measured AFR's to be substantially richer - especially at idle. Maybe 12 to 13 is ok with 10% ethanol...

Hello!

Few hints:

1. There is no other way of measuring >real< AFR's but by comparing actual air drawn in with accurate measurments of fuel injected. That's nothing a amateur can do.

2. Narrow-band lambda cannot tell you AFR's or actual lambda, no matter how fancy multi-colour instrument you conect to it. It's just a crude sender that goes from low to high when you pass lambda 1.0. It cannot tell you if you are running lambda 0.8 or 0.9. It doesn't have that kind of resolution.

3. Only way to measure lambda is to use wide-band lambda logger. Note that such device can only measure lambda, not AFR's. AFR's are just generated by multiplying lambda index with assumed stochiometric AFR of fuel used.


To round it off:

Forget AFR's. They don't mean much. It's only lambda that matters. AFR's are artificially generated.

With other words:

- If your wide-band lambda logger tells you you are running AFR of 12:1 it just means that it detected lambda 0.81 and multiplied it with what it assumes is actual AFR for fuel used (usually 14.7). Which means you're still running rich, regardless of ethanol content. Reconfigure it to show actual lambda numbers.

- If your narrow-band lambda tells you you are running AFR 12:1 just laugh at it. It can only tell you if you are running rich or lean. What is displayed as 12 can as well be 9 or 13.


- Actual AFR's vary for different fuels. You don't really need to know actual AFR for your fuel unless you are dimensioning your fuel system. You only need to know lambda.