That is interesting. I thought fuel droplets at the intake manifold formed an aerosol mixture with oxygen and were sucked down the intake tract as the valve opened with the partial vacumn in the combustion chamber does the pulling. Depends on rpm, percentage of injector duty cycle?

 

I generally hear the puddling term used when associated with carbs that don't vaporize the fuel very well.

IIRC, there is a function table in the factory PCM for fuel coming into contact with the port walls. I could see this portion causing puddling. Once it returns to a larger liquid formation it is hard to get it to vaporize, even with the heat in the cylinder, as the heat energy doesn't have time to do its thing. Heck, I see a large portion quickly puddling in a liquid state in a testing tube.

A true vapor will not do that unless there is a substantial reduction in heat energy to bring it back to liquid form.

To visualize a true vapor, open a gas can , or while fueling your car, & watch the vapor that escapes. That doesn't quickly form fuel droplets after hitting the atmosphere if the gas is at ambient temp.

You can also do it with a propane tank with a reg. attached. Open the valve & watch what comes out. It is liquid in the tank but comes out vapor.

 

Yes, but do injectors vaporize fuel, or atomize it? To vaporize fuel, you change the physical state of the fuel from liquid to gas. With atomization, you simply reduce the physical size of the liquid fuel droplets, but it is still liquid and not gaseous. Right?

 

You are correct.

The vaporization happens variably within the gasoline because there are light & heavy hydrocarbons in suspension.

The light hydrocarbons are reduced to small enough droplets they can mostly be vaporized by the heat energy. The heavier hydrocarbons are not reduced enough for this to happen.

 

Got it. Thank you.

 

OK, I see where you are going on that now. And yes, under a constant flow scenario, the higher duty cycle will make fore a more even mixture. Here's the problem.

The tune in a third gen LS controls the End Of Injection Timing and then the necessary pulse width for fueling determines the starting point. In a stock tune and most aftermarket tunes, the EOIT point is 60 degrees BTDC on the exhaust stroke - believe it or not, this is the truth. The reason for this has been guessed at many times, and the going theories are that this is done to maximize fuel evaporation prior to the intake charge going into the cylinder, using the natural turbulence of the air flowing through a gap to mix the air and fuel prior to compression. At colder engine coolant temps, the injection timing is moved earlier to give the fuel more evaporation time until the engine warms up.

Now, lets consider a performance cam. The stock cam has something like -34 degrees overlap at 050, so the exhaust valve is basically closed before the intake valve opens, resulting in little if any fuel short-cutting out the exhaust, so the fuel timing isn't that critical. Add 42 degrees of overlap, and now the cam has 8 degrees overlap. In this scenario, the exhaust valve is still open when the intake valve opens, so a bunch of fuel escapes. A common tuning trick is to add 90 degrees to the EOIT so that the fuel spray is timed to stop at 30 degrees ATDC on the intake stroke instead of 60 before on the exhaust stroke. The result is that at idle, far less fuel escapes out the exhaust, so BSFC at idle is reduced. Some people have reported gaining some off idle torque as well from this.

Still with me?

Now, at higher RPM, the whole dynamic changes. 8 degrees of overlap now occurs 6x faster at 6K vs 1K, so there is a lot less time for the charge to short-cut, so the EOIT makes far less difference at high RPM vs low RPM.

Now enter the duty cycle. a 75% duty cycle means that the injector is firing 75% of the time and closed 25% of the time. Put this in terms of a 4-cycle engine, the injector is firing for approximately 540 degrees of crank rotation and is closed for 180 degrees of crank rotation. On my cam, the intake valve is open 281 degrees of crank rotation, and my EOIT is set to stop the spray at 20 degrees ATDC on the intake stroke. This means that more than 3/4 of the fuel spray is happening before the intake valve ever opens and continues a bit after the intake valve opens, but stops while the piston is still pretty close to the beginning of the intake stroke, relying on the intake stroke and then compression to finish the mixing of fuel and air.

At 6000 RPM, the fuel consumption is 27.7 grams / second, and the engine rotates 100 times per second. Since it's a four cycle, that represents 50 fueling events, which means the injector sprays 0.554 grams on a single spray, mostly onto the back of the intake valve, and it sits for at most 30 milliseconds. Hardly a qualifier for puddling.

However, I would think since most of the fuel is on the back of the intake valve, the spray pattern becomes a bit less relevant to mixing effectiveness.

 

Ok, you're alluding to CNG being "dry" vs gasoline being "wet" and so gasoline has an exchange with the puddle on the port wall.

 

A certain portion of it will film & flow on the port walls, vs propane or NG being in a state of vapor at ambient temps above xx. Degrees, yes.

There is obviously a portion that remains in liquid state that doesn't necessarily come into contact with the port walls. This portion will vaporize much easier than
that portion which comes into contact with the port walls. How easily the suspended portion is vaporized depends on its uniform droplet size being under a certain size.

 

I see the timing portion of your comment. That all makes sense.

The portion that puddles isn't really impacted by induction cycle timing. It is in a film on the port walls & not in suspension. The rest of it, based on the uniformity of the droplet size determines how much turns to vapor before the intake valve opens, at a given velocity & AIT, & how much needs a given amount of residual heat energy in the cylinder.

Also consider that if you have a small portion that is being vaporized, being present in the intake tract, vs what is not being vaporized, the heavier fuel molecules possess more inertia & will take longer to accelerate along with the air mass. You end up with a situation where the wanted vapor is expelled & the non-vaporized portion remains. Not the ideal scenario.

As you can see in your example of what is sent out the exhaust when the intake valve opens is determined by overlap. It is easier to look at this in absolute timing rather than using the .050" measure. That is what the A/F mixture actually cares about.

If the vaporization isn't good enough, & compression high enough, you need to artificially enhance VE by overlap. Controlling fuel timing to institute a reduction in net loss in fueling is a necessity of the system as a whole.

 

Lets try and calculate how much fuel comes out of a 50lb/hour fuel injector with 1ms of on-time. My calculations are deriving about 8uL of fuel with 45.5psi of fuel pressure, which comes out to 56uL fuel / 1 second. At 850rpm there are 14rev / 1 second, or 7 revolutions of fuel injection events.

one or eight microliters is a vanishingly small amount of gasoline. If you injected it into a cup at room temperature, it would be gone before you could blink. Now imagine it being injected into an engine- a hot, turbulent engine. None of that gasoline will hit the valve or port walls, its going to vaporize instantly and begin ascending up the intake tract into the plenum. This is where having long runners helps keep fuel "that belongs to that cylinder" in place. I also find it hard to believe my own derivation- I was expecting a larger value. So now I am trying to figure out how to verify my number.


I need more time with it. Anyone else care to try it? Otherwise I will report back with a final solution after I do my real homework.

Also, terminology guys. "puddling" when I said that, I meant "hydrophobic interactions" whereby fuel molecules are allowed to interact, and "collect" (whether in the air, or on the port, or on the cylinder wall, or on the valve doesn't matter) together. Hydrophobic interactions are reduced at higher temperatures and in turbulent conditions. Hydrophobic interactions of gasoline cause poor cold-start performance, since non-vaporized fuel (hydrophobically interacting fuel molecules) will not burn, aka liquid fuel will not burn. A bucket of gasoline on a cold day will put out a cigarette. It might go up in flames on a warm day, since higher temperature = more vaporizing fuel molecules per unit time (more molecules of fuel have high enough escape velocities to escape the "puddle" of fuel in the bucket, and are thus found nearby it in a gaseous state, capable of initiating combustion) This is also called "partial pressure of a liquid".

 

Have you ever seen what the spray pattern is like on an 50lb/hr injector running a very narrow pulse width? How about running injector flow checks on an injector flow bench?

Also, the main issue isn't when looking at it from a low-volume short pulse width issue. It is when the injector is flowing hard & the spray pattern is fully developed. It will indeed make contact with the port wall.

 

Fuel is Atomized by the injectors much better than a carb.

The low pressure in the intake causes the fuel to vaporize quickly. (reverse of a radiator cap)

The higher temps of anything the spray hit accelerates vaporization. (any well atomized fuel that hits the valve or bowl wall boils instantly)(the bowl wall area is above 190°f)

Timed injection only seems to help smooth out idle.

Sequential or individual injection (dampens fuel rail harmonics? by reducing the pressure spikes from 4 injectors at once to 1 or 2)(multiple squirts) and gives the PCM the ability to service cylinders individually.

 

Most carbs, not all.

Homogenization of a pulsed injector, being runner or chamber directed,is nowhere near ideal.

 

I see what you're saying and I agree with you. I have seen injectors fire in a throttle body injection system, and the fuel just sprays all over the manifold surface before it ever atomizes.

 

And the PCM calibration has tables for modelling wall wetting/evaporation.

 

The real proof is if you look at the the port on a motor with few miles you will notice the top portion of the cathedral (injector relief) will be clean & so will the walls just below it on the sides.

 

So I am just kind of picking up pieces of what I just figured out, I am not sure if it is quantifiable (countable, relationship with a graph available).

2-5uL gasoline promptly evaporates, but so does 12uL or 72uL. Even 100uL of fuel is still not really alot of fuel. Think about how long a gallon lasts (25-30 miles?) for thirty minutes at 3000rpm? that would be 1500 injection events per minute for 30 minutes, 450,000 injection events for each injector, and if each was injecting around 1uL that would be 1uL x 450,000 events per injector, 450,000uL or 450mL or 0.45Liters. there are 3.785L per gallon so you used 0.95gallons of gas this way, with just 1uL at a time.

At the time of the firing of the injector on an engine at 3000rpm, the situation is quite different than at home, spraying into a cup and measuring volume. If we apply just 1psi of fuel pressure we should get something, whether it was an injector or a stray, since pressure drives the molecules in the face of an injector opening, there are no applicable harmonic pressure related injector flow dynamics I am aware of, such that any pressure changes in the manifold are also measuring in the fuel injector rail, meaning that when the fuel injector does open, it will be insensitive to the fluctuating pressure in the manifold (these adjustments will not correlate to changes in injector spray except where differentiation models are slightly off but everything is still very closely corelated). The other forces are work are increased temperature, and the velocity of moving air. These are the only two things I can think of that will make a difference, from a hot engine, to a cold engine, as far as NONE of the fuel system is concerned. (in the fuel system, the injector itself, fuel temperature, positioning, pump output, fuel pressure, etc will all affect injector spray)

We should all be familiar with the volatility of gasoline at higher temperatures.
That leaves the effects of velocity which is measuring in volume per time (like CFM). bigger ports lower velocity, which should make it more difficult for fuel to become gaseous. Keeping in mind that, 1uL of fuel is not difficult to send to gaseous state even in a cup at room temperature.


things I still need to know:
1. How many revolutions can occur after removing fuel pulse and still get a combustion event, on each different engine (the type of injector used, the design of the engine, will influence how much available fuel is suspended or collected nearby for use once the injector is turned off) finer sprays from smaller injectors or better designs might result with less "wall wetting" right out of the injector.
2. How much of this contributes to fuel cut related detonation damage, if any
3. When adjusting the pressure of the manifold by moving your gas pedal, what is the fraction of fuel suspended or collected near the valve or in the cylinder capable of acting as a reservour of fuel on a given engine. This is data we cannot possibly expect to extrapolate- the mass of fuel reservoirs on a running engine intake manifold? I could guess 10 or 15% but if I did that, it would be pretty obvious when our out would be, larger injectors and bigger fuel pressures means more extra gas waiting around (10% of the injector duty is alot when the injector is 1600cc/min)
4. To further understand #3, resolution is the key factor in measuring every molecule of fuel. A tiny change to an input number can result with an extremely wacky air fuel ratio numerically. Working the data forward and backwards is fun but ultimately you are going to use the wideband and a quick tap on the keyboard to line the data up for each segmental question to make the engine run how you think it should.


I now also need to relate actual injector sizes to flow rates on one side, then airflow rating of engines converted to moles for air fuel calculations,
and in the center is the BSFC and RPM data to fill in from different applications.
I will post a couple papers of calculations so you can try to find errors if there are some

 

All that "calculation" is fine & dandy, but at the end of the day all you need to ask yourself, as pertaining to vaporization, homogenization (mixing) & compression, is why can't just about any EFI fueled engine (port or chamber injected) run at a stoichiometric A/F ratios at WOT & require so much ignition advance?

The answer is quite simple when you analyze it on it face value of the three primary factors & accept the reality what you think is happening, isn't happening the way it is supposed to. Just because a portion vaporizes off the port walls does not mean it is being homogenized properly. That is a limitation of the system & is a constant due to the constraints of the delivery method. It is what it is.

As it stands now I am not sure how much more I can add to the conversation with regard to the calcs V8R is trying to reach. He is getting closer to being able to extrapolate a given data subset & have a result within a given margin of error. To me, from my perspective, It is fine tuning a pumping loss curve from here on out. Meaning each engine will have its own based on energy balancing & all the other stuff we know comes into play.

Predictions will always be within a given margin of error. How close he is willing to try to get & how close he can get it will be dependent on how much time he is willing to put into it.

I will say this just to throw it out there: What many think is proper in regards to squish & quench is the wrong direction in thinking. The engine wants none of that. It introduces tight spots that create A/F dispersion, & thusly turbulence, that disrupts kernel development & smooth flame propagation as well a creating an environment (think squish) that increases a given fuels propensity for detonation. Unfortunately with a valves engine & the designs that we work with, only so much can be done to achieve an ideal chamber shape.

 

See, I have been seeing this discussion, and the end goal of producing a formula, as a tuning algorithm people who give a **** would apply to their own data.

This work, so far, has been based on an actual engine, with actual empirical data to base the calculations on.

Once the formula is revised, we need to run the model with a new data set, to see if we can maintain the same margin of error across multiple platforms.

The finer points of fuel injection comes down to what a lot of people said at the beginning of this discussion, about be happy with what you can see and measure.

If we do end up with a friction coefficient, that we can manipulate with better bearings... Better bearings that you can go out and buy... We have at least some feasible way to accomplish this. We can exhaust this fuel injector discussion to its bitter end, but unless we have the ability to manipulate the results, it's a fruitless endeavor. Better bearings already exist. If we find out that injectors are actually terrible at their intended purpose, we are screwed unless one of the people involved in this discussion invents a better injection system. Because the injectors we have access to now, are about as good as it gets (to my knowledge).

Very, very interesting conversation about the fuel, though. Lots to think about.

 

I don't mean to sound discouraging. This discussion on fuel is very intriguing. And we may be able to derive a way to calculate fuel mass efficiency of an engine's injection system.

The same way we are looking at the plumbing losses of the engine, we could probably see the intake tract losses of the fuel. Whether or not anyone actually did anything with the data is irrelevant, but it would be very interesting to see if, say, a fast intake provides a column of air that helps homogenization of the fuel, versus a super vic intake. Does the intake manifold design effect it? Who knows?

And, about the quench/squish being the wrong direction...

Can the turbulent air caused by the close proximity of the piston and head surfaces at top dead center cause detonation when the spark plug should already have ignited the mixture by the time the piston reaches top dead center? It was brought up that the turbulent air/fuel mixture is more prone to detonation, but squish only occurs at top dead center, where your spark plugs should already have ignited the mixture, so it being more prone to burning at this point would be a benefit, would it not?