Ideal AFR theory (modern 4cyl engines)

[pickler]

In tuning modern compact high compression NA engines, i have always had trouble dealing with knock/detonation under wide open thorttle. it seems these cars can exceed Minimum Best Timing for torque without knock under partial throttle but nearing full throttle they start to knock severely far from MBT. The cars I tune come fairly lean from factory ranging from 14.5:1 to 12.8:1 AFR based on engine speed under WOT. Enriching the mixture has helped me reduce knock at mid-high RPMs considerably and this also allows me to run more timing for more power (ideal can be more power?). The reason why I went richer rather than touch timing is because timing is retarded enough as it is. Most cars I see are usually pulling 15% timing due to knock. Most of these cars can't handle more than 22-25* timing by peak power.

I know the ideal Air Fuel Ratio varies with many factors. But assuming I retarded some timing due to higher flame front speed and knock, what negative effects are produced going from say a lean 13.5:1 fuel mix to a 12.3:1 rich mix under WOT? less power, reliability, efficiency? Is this mixture even ideal if i make minor gains (brake specific fuel consumption comes to mind)? I read conflicting report about fastest flame front speed AFR and some articles claim that best BSFC comes at ~0.9 lambda which only produces knock for me unless i retard timing another 3-5*. Yet less timing and leaner will surely cost lots of power no? There is also quite a bit of knock remaining at low RPMs, specially in higher gears. Adding fuel and taking away timing does not seem to help lower RPM knock though.

 

[GregLocock]

The production engines I am most familiar with use excess fuel at WOT to cool the exhaust valves and cats. I don't know what AFR they run in that condition.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Lou Scannon]

I've seen as low as 10:1 in papers published by SAE. This is also about what the state-of-the-art aero engines ran at take-off condition in the 40s and 50s.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

 

[bigTomHanks]

A good WOT ratio is 11.5. This may not yield maximum power but is very close for many street engines and is fairly safe to run. I would rather run a little rich and add more timing and also have a little better sense of security. If you have access to a dyno then you can start at 11.5 and walk the timing/afr up to what makes best power without detonation. I think the reason a lot of people run 11.5 is because it makes good power while allowing a little cushion for operating conditions (weather, ect).

 

[Lou Scannon]

Since the OP mentioned NA engines specifically, I should have added, the aero engines were obviously boosted, and so was the 10:1 SAE reference.
I agree with bTH about a cushion - I'd rather leave a few ponies on the table, than have melted piston tops when things go a little off target.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

 

[Tmoose]

" some articles claim that best BSFC comes at ~0.9 lambda which only produces knock for me unless i retard timing another 3-5*."

What throttle setting/BMEP are you talking about?

Are you tailoring your advance curve in specific regions along with throttle opening/manifold vacuum, or just moving the basic "static" timing point.
Are you measuring BSFC, or was that a report of your results tweaking timing at the 0.9 lambda?

 

[pickler]

thanks for the responses. cushion is always a good thing. but I was wondering why these cars come so lean from factory with high timing. I always end up going 1-1.5 ratios richer than factory and also pull a bit of timing. oddly enough this always leads to more torque for me. I guess you can be too lean/advanced. just wasn't sure about being so 'different' from factory tune.

 

[BrianPetersen]

You haven't said exactly which engine. But if the stock setup is pretty mild and doesn't push the thermal limits of pistons, exhaust valves, or catalyst, running at stoichiometric for as much of the speed/load map as possible is better for emissions. I have one that appears to run in closed loop over the entire speed/load regime all the way up to full throttle max revs, at least for the length of time that I've tried it.

Yes, it will be leaving horsepower on the table, but that allows the 3-way catalyst to do its thing practically all the time.

 

[pickler]

I rather change a cat than an engine. The improvement in exhaust gas temperature and reduction in knock counts is pretty great. I just do not understand why an engine should be tuned to opearte at stoich ratios near peak torque. if I remember correctly exhaust gas temp is highest at stoich. its not unusual to hit 1700f EGT on a stock engine. there is about 100-150f improvement by adding just 10% fuel (0.9 lambda). this is not just because of extra fueling cooling the chamber its also because more timing can be run.

I'm just talking strictly WOT here. I have noticed that at part throttle MBT can be exceeded without knock anyway. so stoich with high timing is fine to keep EGT in check.

this particular engine is a 11.5 compression Subaru flat 4 sohc with high lift cams. so far an 11.8:1 AFR has finally allowed us to dial in just 20* of timing at peak power on 94 octane gas. stock it can handle 24* on 91 octane gas with 12.5:1 AFR. I usually deal with Mitsubishi, Honda, Nissan and some Mercedes NA powerplants.

 

[Lou Scannon]

It could be that the actual gas temperature leaving the cylinder is highest at peak torque but not necessarily. At any rate the temperature that the exhaust valves and downstream components experience is usually highest near rated power, since the energy throughput, as a function of time, is highest around there.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

 

[BrianPetersen]

"I just do not understand why an engine should be tuned to opearte at stoich ratios near peak torque."

Two thoughts.

If the engine is smallish for the vehicle that it is installed in, it's quite conceivable that the various emission testing protocols will require operating the engine near peak torque at lower revs during some part of the test cycle. If that is the case, then the manufacturer has essentially no choice but to operate in closed-loop under those conditions in order for the vehicle to pass its emission certification. I'm pretty sure that's the case with my car, because it will pretty easily see 100% load and full atmospheric pressure on MAP according to scangauge, and it also reports closed loop in those conditions.

And even if that's not entirely the case ... One of the ways the manufacturers eke out better fuel consumption figures is to run the engine at as low revs as possible and under high load, as opposed to letting the engine rev but with lower load. In order to do that, it has to simultaneously pass the emission certification tests (the test for emissions and fuel consumption is the same). In order to do THAT, it has to run in closed loop under those conditions.

 

[pickler]

so then I can be convinced the only reason the fuel map on some of these cars are so Lean (specially at midrange) is solely for emission purposes only. Therefore if I'm looking for power and reliability a richer mixture with more timing is the correct way. as to what ratios I think it depends on the engine but if knock is not a factor ~12.8:1 AFR is usually what I would aim for in a modern engine. sometimes going only a bit richer can let you run 3-5 deg of additional timing. If this I'd the case then I would go for it.

however there is also question of flamefront propagation and speed. SAE claims ~12.2:1 AFR causes highest flame front speed but others claim richer 11:1 ratio is the fastest. Either way I would think one need to retard timing at fastest flame front speed no? I wish there was a definite answer to what ratio gives the fastest flame front speed. This could help distinguish between rich knock or high timing knock. Also this can mean that going rich at certain points will not allow you to run more timing so either staying lean or going super rich (like on turbo engines) is the ideal ratio.

 

[pmrobert]

I'm of the opinion that the best way to answer these questions is via empirical methods. It's really difficult to accurately model this stuff as there are way too many variables. Logging the best quality data you can for post analysis, starting in safe operating regimes and going where the data leads you while trying not to melt things is what has worked for many people seeking improved power as they wish to define it. Always remember, you drive your car, you don't drive a dyno although the two are definitely related.