Subject: AirSIG: KEFI Redux in Tux - Part 2
Alright, back to our latest episode of the NeverEndingStory of Kiss EFI
for XA...
We left our intrepid adventurers still groping in the dark about just
how could you make a FI system so that with ANY single point of failure,
the entire system continued at *full/commanded* power without
intervention. This was the holy grail and quest which began this saga
long ago at Arlington '97 (whew, it's really been THAT long!?). An
electronic FI system simple/redundant enough to rival the reliability of
the Continental FI system of bygone days.
Well, remember I said we discovered something heretofore unbeknownst to
Henry or I, about how the Continental FI system actually works? This is
gonna sound like one of those palm-print in forehead kindsOthangs, but
sometimes the 2X4 has to whack you in the brainstem before you wise up.
We had looked at the Continental system, whose mechanical fuel delivery
pump essentially provides a metered positive volume flow (based on
throttle pos and rpm) to the fixed orifice nozzles at the ends of the
fuel distribution lines, and we had seen that the fuel pressure range
used was about 2-20psi. As we pondered this system further, we realized
that this range of pressure simply wasn't nearly LARGE enough a spread
in pressure to provide the approx 10-15:1 ratio of idle to maxpower/WOT
fuel flow. As we *all* know, heh, the flow thru a fixed orifice nozzle
goes up as the square-root of the pressure, so you can see the range of
>from 2 to 20 psi just wasn't anywheres near adequate to explain how they
were achieving the say 10:1 fuel volume delivery range. We KNEW that the
max pressure wasn't above around 20-25 psi on most any of the models of
the Continental FI system (that from the service man excerpts that Kevin
kindly sent us), and the pressure needed at idle for a fixed orifice
would then place that pressure at such a low value, that the puny lil
dribbles couldn't possibly be distributed evenly between the individual
cylinder fuel lines. Sooooo, we took a VERY close look at the
"distribution block", and loNbehold, it dawned on us how it must work;
there's a low pressure valve in the distribution block that's waiting
until the pressure reaches 2psi before it opens up. It's in the pictures
and diagrams; twas there all the time, but the printed narratives never
fully explained what it's there for. They just said it was there to
insure even distribution, but they didn't say HOW it did that.
If the fuel pump is metering wee amounts of fuel, say at idle, the
volume has to build up to where the 2-4psi minimum for proper
distribution is reached, before the valve in the distribution block
opens up and squirts some out. IOW, that valve is a bloody CHOPPER,
similar to the way PWM works, vibrating openNclosed so that no matter
what the volume metered by the pump, there has to be enough pressure
built up in the distribution block, before that pressure squirt is
applied equally to all the distribution lines going to each cylinder.
And in this way, they can assure that the flow is evenly distributed to
each cyl, even during the wee dribble times of idle. We subsequently
discovered that in other forms of continuous-flow fuel injection, this
same sorta chopper valve is used; in the case of the old Bosch
automotive CFI systems (referred to as the K-Jetronic system), it's
actually in each nozzle, right behind the fixed-orifice. Because the
valves are quite small in the Bosch nozzles, they actually make a
buzzing sound at around 1500Hz!
Another way of looking at this is the following: whenever the continuous
fuel delivery volume gets so small, say during idle, that you don't have
adequate pressure head to insure even continuous distribution when you
try to split this fuel up to each cylinder, then the system stops being
"continuous flow", and becomes a pulsed system of intermittent bursts,
but each burst is made to occur at a minimum pressure (the setpoint
pressure of the so-called "fuel manifold valve" in the distribution
block) sufficient to insure even distribution amongst the individual cyl
fuel lines.
ACTIVE REDUNDANT FI
Does this sound at all familiar to you? Does it sound kinda like
conventional pulsed EFI? Heh heh. The fact is, whether you use a fixed
differential fuel pressure across the injector, and then use pulse-width
to modulate the fuel volume (conventional EFI), OR you use a
variable/metered volume and let the pressure come what may (Continental
continuous FI), in EITHER case, there comes a point where the pressure
MUST be chopped or pulsed to insure even distribution and control of
flow. This is how we ended up looking at each other one day, and
confessing we just realized we never actually KNEW until then, how the
Continental FI system actually worked over its complete range.
But as is often the case, "humiliation is a true friend" once again, and
this insight led to the path of XA FI Nirvana. Consider the following:
the reason why you can't just parallel up two injectors per cylinder to
achieve active redundancy is you end up with twice the volume of fuel if
you have them both operating. That's a no-brainer, yeh? Also consider
this: even a conventional EFI injector has a fuel switch/solenoid that
opens the fuel path, which then goes thru a metering orifice in the end
of the injector tip, which is how the different injectors are made/rated
for a certain # of lbs/hr. OK, well suppose you separated the solenoid
>from the orifice by some distance. Would it work the same? Ahh yup, in
fact that's kinda like the Continental system works when it has to go to
"pulsed/chopped" mode at low volume. The pulsating device is in the
distribution block, but the orifices are in the nozzles at the end of
the fuel lines going from the distribution block, out to each cylinder.
OK, now hold that thot for a moment.
We've already observed that paralleling two fuel FLOWS gives you twice
the fuel, but consider this: if you parallel two fuel PRESSURE sources
and the resultant pressure feeds a fixed orifice, will the flow go
up/change/what? NOPE, it will stay the same. You can gang as many equal
*pressure* sources in parallel as you like (assuming they're equal), and
the orifice will still see that one pressure, and still flow the same
amount of fuel. So now, getting back to the thot above I told ya to
hold, what if we took two normal EFI injectors and separated their fuel
solenoids from their orifices (just figuratively speakin, o'course), and
we had these two dismembered injectors plumbed in parallel, so that each
one opened it's fuel flow solenoid at the same time, BUT each of these
flows fed a common outlet line with ONE fixed orifice at the end? Now
assuming you have two separate fuel pressure sources of fairly equal
pressure (i.e. two fuel pumps) each feeding a fuel solenoid, and you
connect the fuel solenoid outputs together into one pipe, then if you
have one or both of the solenoids on, does the pressure in the output
pipe change? Nope, not unless of course you're bleeding so much volume
that the pressure can't be maintained. And even if/when you pulsed these
two solenoids onNoff, as long as you turn them both onNoff at the same
time, you now have active redundant fuel supplied to the nozzle at the
end of the line, but with no volume change. Voila FI Nirvana!
What's the bottom line to all of these mental gymnastics? Well, if you
take TWO VERY LARGE but otherwise normal EFI injectors, plumb their
individual INLETS each to redundant fuel pumps each delivering equal
pressure, and you tie their OUTLETS together, you now have a parallel
source of fuel PRESSURE. If you then plumb the combined OUTLET to a
distribution fuel line with a fixed orifice in it to provide the
metering, then you have a truly active redundant FI system, where a
failure in any of the active parts cannot even ALTER the fuel delivery
(see the one soft exception below), let alone fail it. The trick is that
the pulsed injectors must be WAY larger rated than the actual fixed
orifice needed at the end of the fuel line, so they (the big injectors)
don't actually restrict/meter the flow, just turn the pressure head
onNoff. We have tested our current Walbro pump, and it does indeed stop
flow in the reverse direction when off, so there's no possibility of one
fuel pump failing, and the output of the good one getting shunted
backward thru the injector/solenoid to and back out the failed one. So
check valves don't even appear needed at the outputs of each pump.
The reason I said the realization of how the Continental system was
working when it went to "pulsed" mode during low flow rates, was the
path to FI Nirvana, was because when that light dawned, it was clear the
obvious fact that you could indeed parallel up two *pressure* sources at
the head of a fixed injector, and the flow wouldn't be altered. Indeed,
in this scheme, as long as both fuel pumps combined have enough capacity
to maintain the required head pressure, BOTH pumps could even be
somewhat degraded and the system still not fail. Certainly if one or the
other pump or injector (or electrics behind the injector) completely
fails, the head pressure to the nozzle orifice remains unchanged. And if
the pump pressures are slightly different, when you have them both on,
they divide the flow somewhat unevenly, but the pressure still remains
the higher of the two. Indeed the worst case failure mode is if one of
the two injectors failed OPEN, which would lead to a full-power level of
fuel flow.
Finally, since in this scheme you are always pulsing the injectors,
especially/certainly at idle, you can indeed run the fuel lines for each
cylinder all from one distribution/divider block, which is fed by this
setup of dual whopper injectors (let's just call them the pulsed fuel
*solenoids*, shall we? :). Only requirement is you find a fuel solenoid
large enough compared to the total fuel flow of the combined fixed
orifices, that the fuel solenoid itself doesn't impose a flow
restriction of it's own.
So the system picture is: dual redundant fuel pumps, dual KEFI
controllers, each of the pair supplying & controlling a fuel solenoid,
each solenoid's output combined and plumbed to a distribution block from
whence go the individual cylinder fuel lines and nozzles/orifices.
Voila, run them both actively in parallel during TOL, and you cannot
have a power loss.
And no, I haven't tried this on an engine yet, so don't ask. Would you
have wanted me to wait until I did, to tell you about it? :) Besides,
consider this a "public disclosure" publication to establish prior art,
so we don't have to wake up some morning in the future and find some XA
suckup claiming to have come up with it himself all on his own, and
wanting to become another vendor. Heh. Pre-emptive strike. Hey, when I
worked for IBM long ago, we used to do it all the time; those things
they didn't want to bother patenting/protecting, they disclosed fully in
their monthly IBM Technical Disclosure Bulletin, so nobody ELSE could
come along and patent the thang. Clever, no?
Now I've got to go find me some hughmongous couple of injectors...er, I
mean fuel solenoids. Precision nozzles BTW are easily and cheaply
available; the guys that run NOX have some nice ones with a full range
of sapphire orifices to choose from, and only about $25 per nozzle. Much
cheaper than the ones used on Air Flow Performance setups, too.
BTW, that's another way of referring to what we've just described; a
simple redundant electronic equivalent to the AFP system. Hee.
Gar
__ AirSIG DropBox: ftp://airftp@interstice.com/pub/DropBox/AirSIG __