Brent did an excellent clarification:

<<Pressure does not cause atomization. Shear causes atomization. Shear
occurs at the interface of two slipstreams of different velocity. Picture a
high wind over a lake, producing whitecaps and atomizing the water into a
fine mist. In the case of an intake runner, injecting fuel near the wall of
the tube (just above the boundary layer) produces better atomization than if
the same injector was located in the center of the tube. High performance
carburetors have a secondary coaxial venturi and the fuel is injected at its
periphery.>>

Exactly.  But the high pressure used by automotive systems improves
atomization by the shear produced from the higher velocity of the fuel being
discharged.  An example is diesel systems that are going to higher and
higher pressures to improve atomization (now approaching 40,000 psi).  Our
air-bleed injectors have a fuel orifice that is coaxial with a larger
downstream orifice with atmospheric air bled in between.  The reason for
doing this is primarily to make the metering system work as injecting fuel
directly into the intake port would create a variable pressure across the
nozzle that isn't detected by the metering system.  A byproduct of doing it
this way is excellent atomization at lower throttle openings.
Unfortunately, this doesn't help during starting or a WOT, the conditions
about which airplane drivers are most concerned.

<<Combustion requires fuel vapor (not liquid) and atomization encourages
vaporization so the more atomization the better, right?  Wrong. Vaporization
requires heat, reducing the temperature and increasing the density of the
fuel air charge. For best  volumetric performance the fuel should vaporize
after it enters the combustion chamber but before the intake valve closes.
The charge is cooled by the vaporization, increasing its density and
providing more "room" for additional charge. This timing requires fuel
droplets of a specific size range, not too big or small so more atomization
is not better, after a point.>>

Correct, but probably not practical.  As Brent says, the vaporization of
fuel, if it could be done by pulling all the heat from the air, produces a
net increase in density.  In other words, while the fuel in vapor form takes
up volume (about 2% at lean mixtures) the evaporation lowers the air
temperature by about 70F for a net 11% increase in the mass of oxygen in a
given air volume.  But to get the fuel to evaporate only during the time
after it is inducted and before the intake valve closes is not realistic.
In theory then you would like to inject well-atomized fuel as far upstream
as possible and then insulate the mixture from the engine for maximum charge
density.  A good carburetor system with excellent atomization and fuel
distribution and no air flow restriction will produce more power than most
fuel injection systems for this reason.  With a downdraft cylinder head like
used by the IO-550 the injectors could be moved upstream with good results
if the fuel could be atomized by the injector.  Without good atomization the
fuel will simply run down the runner until it evaporates, not a good thing.
Cold starting should actually be improved as the fuel would be deposited on
the manifold surface, giving it a chance to evaporate before running out the
port drain.  I would encourage someone to take this approach and report the
results - the increase in power could be as much as 5% and likely about 3%.

The option taken by our systems is to inject fuel onto hot surface for later
vaporization.  I think if you could look inside the intake port the fuel is
mostly deposited directly on the cylinder head surface to be evaporated
during the dwell time between intake strokes.  In this case most of the heat
of vaporization comes from the cylinder head, not from the air.  In the past
I've done lots of studies that show that this is the dominant evaporation
process for injected fuel.  The Lycoming design is better and is similar to
most automotive engines in that the fuel is aimed at the back side of the
intake valve, using that as the heat source to vaporize the fuel.  No, the
idea was NOT to squirt the fuel into the open intake valve as that fuel has
very little time to vaporized before the ignition event.

<<Does any of this matter in our aircraft engines? Not very much, really.
Our aircraft engines are big and slow and use (on injected engines)
continuous injection. In high performance fuel injected racing engines, fuel
injection begins at the start of the intake stroke and stops at the end.
This produces a relatively homogeneous mixture of fuel and air entering the
cylinder. On our aircraft fuel injected engines all the injectors are
flowing all the time. The fuel accumulates in the intake tube for 75% of the
time and is then gulped into the cylinder in one very rich slug followed by
a very lean mixture. Fortunately, because aircraft engines operate at
relatively slow speeds, there is lots of time to vaporize the fuel.  As
George Braly will likely tell you, consistency from cylinder to cylinder and
from cycle to cycle has a greater effect on performance and smoothness than
fuel droplet size.>>

All production automotive systems of which I am aware take pains to inject
fuel when the intake valve is closed, specifically to improve vaporization
and reduce emissions.  Injecting the fuel just after the valve closes
produces the best emission results (most homogenous charge) while injecting
later produces better throttle response.  At full throttle, maximum speed,
most injection systems are set up to have the injector on about 80% of the
time, which means it will be injecting most of its fuel when the intake
valve is closed.  It has been shown many times that when injecting the fuel
on to a closed intake valve improving atomization shows no power increase or
emissions reduction.

What does all this mean to us?  As Brent says, not much, although air
pressure for atomization were increased as much as possible there might be a
slight improvement.  Would I want to use shrouded injectors on my NA (by the
way, "NA" stands for Naturally Aspirated, not Normally Aspirated) engine?
From the posts I am reading I think it will be worth the trouble, but not so
much from the improved atomization as from the improved fuel distribution
since all injectors will have the same reference pressure.  Correct?

Gary Casey