To put it into context, the whole thing is about vaporization of fuel in a stock setup, and about not throwing out fuel on the cammed up side.
In a stock rig, the injector sprays onto the back side of an intake valve. Both valves are in contact with the heat from the combustion chamber, but only one valve ever looks like it's been hot, right? That's largely because the intake valve is being cooled off with fuel, and nothing cools the exhaust valve directly.
In stock form, it's about turning raw gasoline (liquid) into burnable vapors (gas). Raw liquid gas doesn't burn, only the vapors do. So we use an injector at 58 psi to turn big gulps of liquid fuel into small droplets of gas. EFI and it's high pressures do this wonderfully to begin with, compared to the raindrops of fuel coming out the bottom side of a carb, right?
Then we spray this fine mist in a cone pattern onto the back side of the intake valve, and we shape the cone to hit near the edge of the valve, by the seat. This way it cools off the hot sharp edge of the valve, and as the valve opens, the air flowing in tends to shear the droplets into even smaller droplets.
Since the cool fuel hits the hot valve, it tends to vaporize and become a gas. So picture it in your head. The intake valve is closed, the motor is idling, the injector PW is maybe 2 - 4 milliseconds at idle, and the injector delay table is programmed to spray onto the valve just before it opens. This way, other cylinders going up and down don't inadvertently steal the air and fuel vapors out of this intake runner.
The injector will have a start time and an end time. The delay table back-calculates the injection timing. So let's say the intake valve is going to open at X degrees of crank rotation in the engine cycle, and the engine is idling so it only needs 3 ms of pulsewidth. The injector would END its spray pattern just before X degrees rolls around. So it would have to START it's spray pattern at 3 ms before X in order to finish by X. If the pulsewidth jumps up to say 15 ms at WOT, then the injector would finish by X still, but start at 15 ms before X.
Here's where it gets interesting. You can calculate the engine speed, the 720 degrees of crankshaft rotation (the cam runs at 1/2 the speed of the crank), and you can figure out exactly how many milliseconds (ms) of time becomes how many degrees of crankshaft rotation. That's just math. So if you're only idling, that start and end of injection time (SOIT and EOIT) will be very close to the time when the intake valve is about to open.
So picture the closed intake valve, the injector opens for a few ms before the intake valve opens, and it closes just before the intake valve opens, making sure that fuel is now vaporized, broke up into little bity pieces, and basically a gas cloud at this point. Then the intake valve opens, that gas cloud immediately gets sucked into the cylinder, and it's ready to be exploded.
Now picture a cam change. We have a cam with a lot of overlap (because it sounds cool at the local drive-in). The intake valve and exhaust valve are now open at the same time, in a way the factory never intended. The intake manifold has vacuum in it (at this point, that vacuum is coming from the other cylinders!), so the intake runner is in a state of low pressure - vacuum. The exhaust system on the other hand, has high pressure in it, from those other cylinders firing and trying to get out thru a small pipe, causing some backpressure.
Now both valves are open at the same time, so the high pressure goes towards the low pressure. That means the exhaust goes in the exhaust port backwards, goes thru the combustion chamber, and exits out the intake port. It pushes some of this nice vaporized fuel cloud backwards, up into the intake manifold. (This is called reversion or standoff in carb speak). But that doesn't last long, and then it reverses direction and the intake charge now gets sucked out the exhaust valve instead.
But our computer calculated just the right amount of fuel, no extra. So we can't afford to lose any fuel out the exhaust valve, or the cylinder runs lean. So we end up richening up idle to compensate for the lost fuel. This is why cammed up cars like to be richened up at idle and low speeds.
We could move the injection timing back some, so that we don't inject fuel into a valve that has already opened. And that might be good, if we have a longer duration cam than stock, but we don't really have any more overlap. So maybe for a mild cam, that works well.
But if we move EOIT back like that, we make it more likely that the other cylinders steal fuel from our intake port. We also still have the problem that as soon as the intake valve opens, some of that fuel is going to go right out the exhaust anyway.
So if the cam is pretty big, we go the opposite way of common sense. We inject the fuel later, onto the back side of an intake valve that's already opened. Why? Well, with longer duration, we can't start the injection any earlier, or we'll end up injecting on the previous engine cycle, where the valve is still open from the last time. So we're going to lose some of that fuel vaporization by doing this, but we are going to be able to time the injection more properly to the longer duration cam.
We're also not building up that cloud of fuel, just so it can easily get pushed back up the intake, or straight out the exhaust, as soon as the valve opens. Better that we spray it in as we go, instead of 'building it up just to get stolen' all at once.
On here, guys are trying to time the injection so that the exhaust valve has already closed. So that means a few degrees after TDC, instead of a few degrees before TDC for EOIT to happen. But what I think is being overlooked is the fact that at WOT, the injector duty cycle is supposed to be roughly 80%, right? So the injector is open 80% of the entire 720 degrees of crankshaft rotation. Which means it must START long before the intake valve opens, and end long after the exhaust valve closes.
Because of this, I theorize that it's not really about the exhaust valve open/close points, as much as it is about starting the injection at the right time (to match the new cam opening/closing points), and about not having the cloud of fuel stolen or completely evacuated out the exhaust valve.
More common thinking is that it's only about the overlap period, and losing so much fuel out the exhaust, and not so much about having fuel stolen from the intake runner by other cylinders (or pushed back up the intake by exhaust pulses during overlap).
At higher speeds, all this stuff sort of works itself out. But at lower engine speeds, the overlap may be the exact same amount of crankshaft degrees, BUT the amount of actual time it takes to pass thru those degrees (a product of RPM) is much longer. So it's sort of like trying to play ping pong in slow motion. It just don't work.
My best guess about what to do here is to experiment with EOIT by slowly adding to it, making the injector start and end it's on-time later in the engine cycle. (Unless you have a mild cam, then maybe go the other direction). I would start by adding 10 degrees of delay at a time, and monitor the results. If you are improving the situation, you'll notice that the engine is getting richer on the wideband.
This is because the correct cylinder is actually GETTING the fuel you injected for it. So an increase in a/f ratio is an indicator that you're improving your situation. Obviously, you'll end up going back to your airflow model and leaning the car out to compensate eventually.
If it liked 10, add another 10 and see what happens. It's tough to say exactly how much a motor would like, since all cam overlap periods are going to be different. But somewhere between 20 and 40 extra degrees sounds about right.
The LS1 PCM doesn't go by actual crankshaft degrees in their delay table though. They go by reference periods. Each reference period is 90 degrees. So if you want to add in 10 degrees, you would add .11 reference periods. Also noteworthy, some clever fellow on here decided to take an engine on the stand, and test it to see where the 720 degree cycle actually starts at, and found that it started at -784 (if I remember correctly). That's just about where the cam sensor is relative to actual TDC on #1, and about actually sensing the crank magnet go by BEFORE the cycle starts, so the PCM has time to get it's act together by the time the cycle actually starts.
That's unimportant if you're just going to add 10 degrees at a time until the motor doesn't reward you any further. But it is important, if you have drawn up a 4 cycle chart on paper, and are trying to actually calculate when the injection timing is going to end, compared to when your cam valve open/close points happen.
And just when you think you've got it all mastered... "did you degree the cam in, and has the chain stretched any since?". heheh
It's not rocket science. Just add a little delay in at a time to the normal table, and probably leave everything else alone. I am a bit less informed of exactly how the makeup thing works, but I'm sure it's about the same way. I believe makeup is made up of things like acceleration enrichment and other last-minute changes to fueling, while normal is the typical fuel injection cycle.
Originally Posted by picnic_george
The Tremor at AIR
Posting Permissions
Reply With Quote