D Chapman said:
I wish Drew Sobota would chime in here. He would blow your theory to pieces, Rob, with real data. He's been on a dyno binge here lately testing several engine configurations with various cams, crank/rod/piston combos, and heads.
One thing is for sure, though, these engines are not running lean. Not even close.
Why do you think I said they are running lean? I just said they are NOT running lean.
R_Lefebvre said:
Who said the engine runs lean? It runs at 14.7:1 just like every other engine.
Wait, I don't get this part, you are saying the system is accurate, but need to be less accurate to have NOx and CO to pass through the catalyst for proper efficiency? Can you explain that?
Yes. I went through this last year. On my track car running a motorsport ECU that I programmed, Motec PLM wideband, and a high flow cat. I came in with a cruise AFR of 17:1. Works well for fuel efficiency. But, failed the sniffer. Way too much NOx. Richen it up to 14:1, NOx is zero, Hydrocarbons go through the roof. Change it to run EXACTLY 14.7:1, it's easy enough to do with a global fuel multiplier, I can adjust fuel +/-1% increments with the +/- arrows on the keyboard, so it's easy to hit exactly 14.7:1 at steady state. Both NOx and CO are lowish, but not low enough. Change it to run closed loop on the wideband, targetting 14.7:1, but with a messed up PID loop so it dithers just either side of 14.7:1 (ON PURPOSE!) and the NOx and HC are both withing spec, pass the test, get my plates.
The reason is, modern cats reduce the NOx and CO emissions of the engine to way below what is possible on even the perfect engine. But they actually need NOx and CO to do this. If the engine is a bit lean, and is sending lots of NOx, the cat can't do anything with it. It can't chemically reduce NOx into N2 + O2. If the engine is a bit rich, it's bombarded with CO. It can't change the CO to CO2, it doesn't have any oxygen. It needs both NOx and CO, so that it can change 2NO + 2CO = N2 +2CO2. It does this most efficiently when it's being hit with waves of NOx and CO. The reason for that is, cats work best when they're hot, as we all know. Sending waves of CO and NOx down the pipe gives them work to do. The chemical reaction increases the temp of the cat, and therefore, it's overall efficiency.
This is touched upon here in the section for a 3-way cat which is what we're talking about:
http://en.wikipedia.org/wiki/Catalytic_converter
And much more thoroughly here:
ftp://www.eecs.umich.edu/people/grizzle/papers/TWC98.pdf
The systems are fast enough and accurate enough to hold the A/F very close to 1.0 Lambda (~14.7), didn't used to be that way, but now it is. So now they actually have to force the system to dither a couple points either way of 14.7.
PTSchram said:
Never in this thread did detonation come into the discussion.
When discussing block failures and possible lean conditions, what other reason could there be? Surely you're not suggesting that if the engine is running lean, with resultant higher combustion temps but not detonation, that the cylinder walls would fail before the exhaust valves?
What other reason could lean running lead to block failure?
YOU said they run at 14.7 and I stated that they ran close to 14.7 but to say that they all run at 14.7 is silly due to the huge number of variables, not the least of which is the variation in gasoline formulations as gasoline is not a finite mix of compounds, but rather anything that comes out of the cracking tower within a varying range of boiling points.
Who cares about what the refinery did? You fill up with a load of fuel once every 200 miles. Your ECU can adjust the fueling 4 times per second! If the gasoline you just pumped has a different energy content than the tank before, your ECU has adjusted before you're out of the gas station parking lot!
The ECU can adjust the fueling almost instantly. Variabilities in fuel chemistry, air quality, etc.... it's meaningless. The ECU is several orders of magnitude faster at adjusting fueling than the inputs change while driving your vehicle.
You're stuck back in the carburetor world where you have to pull your carb and change jets to change your fueling.
The EFI systems are so fucking fast now, the only thing stopping them from being faster is the transport time between exhaust port and the downstream oxygen sensor. It takes some amount of time for the exhaust gas to actually reach the sensor so the computer can measure it, and adjust fueling.
But the latest systems even have THAT conquered with calculated transport times so accurate that the computer can resolve exhaust pulses coming from individual cylinders, and develop a fuel correction factor for every INDIVIDUAL cylinder!
"This lean pulse came from cylinder #3, so I'll add 1% more fuel to just that one"
Have you ever actually driven a truck and watched the Short Term Fuel Trims? They change once per second, and probably faster than that but once per second is the polling time of the OBD reader you're using. The STFT's have an authority of +/- 25%, and can go from one end to another in seconds if required. The composition of fuel and air do NOT change anywhere near that fast.
Result? Engines cruise at EXACTLY 14.7:1, +/- what the engineers CHOSE to make them dither at.
The only exception to this, is worn O2 sensors giving an inaccurate reading.