[DeTomaso] EFI

[Daniel C Jones]

> Dan
> Do you remember the label on your old post about control schemes?
> dave

No but here is a cut and paste summary from my database:

> The company is called 'Mass-Flo'. The other outfit uses the Mass-Flo system
> in some of their builds.

I looked that system over for some guys on the Pantera forum.  I'll re-post
some of what I posted there:

The Mass-Flo system uses a Ford EEC-IV mass air ECU with a GM LS-6 mass air
meter mounted directly atop a carb intake that has been fitted with injectors
in each runner.  They use the GM sensor for packaging reasons.  A 4 hole
throttle body is used, along with a Ford idle air controller and a box to
convert the GM frequency output to an analog voltage the Ford EEC expects.
Mass-Flo claims their system requires no tuning and will work with any cam.
That is simply not true.  It'll work with cams of sufficiently low overlap.
Cams with large overlap can fool the mass air sensor.  In cases like that,
flow travels in both directions in the intake tract at idle (in plenum V8
intakes where one cylinder is on it's intake stroke and an adjacent cylinder
is not) and at certain resonant conditions at non-idle RPM's.  The mass air
sensor doesn't know where the air is coming from but tells the ECU to react
anyway.  The result is a surge.  This is the primary reason most aftermarket
ECU's are alpha-n or speed density-based.  With the Mass-Flo set-up, the mass
air sensor is located right on top of the plenum which would make it very
susceptible to surging.  One of the tricks to tuning mass air 5.0L's is to
locate the mass air sensor farther from the plenum.  Also, the Ford EEC that
Mass Flo uses assumes a 5.0L engine of a certain volumetric efficiency. These
tables exist since you don't always operate with the MAF sensor in the loop
(or the sensor limits).  Stray too far from the table assumptions and you'll
run into tuning issues.  In short, the Mass-Flo system isn't magic.  It is
susceptible to all the tuning issues the stock Ford EEC and mass air set-up
is susceptible to.  There's a reason things like the TwEECer, EECTuner and
PMS exist.

You can duplicate the Mass-Flo system with a single plane intake (set up for
injection with injector bungs), a 90 degree elbow with throttle body, a Ford
EEC-IV processor, and a mass air meter matched to your injector size.  That
set up would have the added benefit of allowing you to move the mass air
sensor, if necessary, and would not need the extra electronics that conditions
the GM frequency output to work with the Ford analog voltage input.  Also, the
large single disc style throttle bodies usually make a bit more power than
the 4 hole carb style.  A friend runs that set up on his drag race Mustang
with a PMS piggyback controller.  Alternatively, you could do the same thing
with an aftermarket speed density or Alpha-N ECU (without the mass air meter).
That approach will be more flexible but will require additional tuning.

There are three popular EFI control schemes:

 1. mass air flow
 2. speed density
 3. alpha-n

None of these approaches is best for all engines in all applications.  You
need to understand the limitations to understand which is best for your
particular application.  They differ primarily in the way they sense engine
load.  Speed density systems use manifold vacuum via a MAP (manifold absolute
pressure) sensor to sense load.  Fuel is metered using the MAP input, engine
RPM, and volumetric efficiency tables.  Mass air systems use a MAF (Mass Air
Flow) sensor to directly measure the amount of incoming air.  Those sensors
typically use wires that air exposed to the air flow, though there other
approaches (flapper door, venturi, etc.).  As the air flows over the wire,
it changes the voltage drop across the wire.  Tables in the computer convert
the voltage drop to air mass.  Alpha-N systems are the simplest, using only
RPM and throttle position to determine load.  Note that a MAP sensor can be
used with Alpha-N, but it's used as a barometric pressure sensor to detect
altitude changes.  You can also blend the authority of Alpha-N and speed
density by directing the authority between the throttle position sensor and
the MAP (manifold absolute pressure) sensor.

Each of these approaches have pros and cons...

Mass air systems don't handle large overlap cams well.  Reversion pulses
produce an unsteady flow over the mass air meter which can lead to surging
at idle and lower RPM.  A friend runs an EEC-IV mass-air based system in
his drag race Mustang and the engine has a strong surge below 3000 RPM due
to the large overlap of his cam.  Wide lobe centers helps some (most 5.0L
"EFI" cams are ground on 112 degree lobe centers for this reason) and some-
times you can tune out the surge by repositioning the mass air meter relative
to the intake valve but the amount of overlap a mass air system can tolerate
is finite.  Since the airflow is directly measured, mass air systems can
tolerate larger variations (in cam specs, cylinder head flow, etc.) than
speed density.  However, there are still tables in the ECU that may need to
be reprogrammed.  Most MAF-based systems still have speed-density modes for
when you're off the range of the MAF or need to go off stoich for accel/decel,
etc.  The sensors will need to be recalibrated for different injector sizes.
This can be done in the sensor itself or the calibration curve in the ECU.
Mass air is particularly good for idle and emissions which is why many
OEM systems are mass air.

Speed density systems can be cheaper and more reliable than mass air systems
because they don't require mass air flow sensors.  They can also make a bit
more power than mass air systems since a mass air flow sensor is a restriction
in the intake flow path.  Speed density is somewhat less sensitive to cam
overlap than mass air but still has limitations as overlap can cause unsteady
vacuum readings.  The size and location of the plenum can influence this. The
GM guys have used speed density with fairly hot cams, though.  A vacuum
chamber to smooth out the signal helps.  Speed density systems are less
tolerant to changes than mass-air systems in the engine before requiring
re-programming of the volumetric efficiency tables.  The commonly available
EEC tuning tools (twEECer, PMS, etc.) don't typically support speed density
systems since the majority of Ford performance applications are mass airflow.
However, Ford's Cosworth F1 V8's ran speed density EEC-IV's for many years but
I wonder if they were running in an alpha-n mode.  The situation is reversed
on the GM side and many GM guys will change from mass air to speed density.

Alpha-N is best for big lumpy cams.  Most of the aftermarket EFI systems
like ACCEL/DFI, Electromotive, Haltech, BigStuff3, etc. are either speed
density or Alpha-N.  Electromotive for one will allow a blend between the
two.

The type of manifolding can influence your selection of control scheme.
Speed-density does not work well on highly modified engines that lose
manifold vacuum upon any throttle opening.   With independent runner
throttle bodies, throttle position is a better indication of load than a
MAP sensor once off idle.  This is because a small opening in the throttle
body will cause manifold vacuum to go to atmospheric.  Beyond say 10-15%
throttle opening, there is little response to a MAP.  Alpha-N is the way to
go with naturally-aspirated independent runner but it doesn't do much for
part load/part throttle fuel economy.  That's where it's benficial to blend
in the MAP.  The engine I'm building for the Pantera has a lumpy cam (238/242
@ 0.050", 300/304 advertised, 0.621"/0.595" lift with 1.7/1.6 rockers, 110
LSA).  Mike Trusty has been running IR EFI for many years.  It's been his
experience that MAP dependent ECU's don't handle IR systems (with or without
big cams) well.  The MAP changes so rapidly, even with a mild cam, that at
just off throttle the MAP goes almost to atmosphere which makes the ECU think
the throttle is wide open even though the throttle position doesn't verify
that.  Add in the hard to get rid of pulsations even with manifolded vacuum.
In his experience what works best is a system that can look at both throttle
position and MAP and allow you to give one or the other more authority in
different power ranges.  I know some guys give throttle position all of the
authority.  That works but doesn't do anything for part load/throttle economy.
Electromotive has a "Blend" mode in their software which is a function that
allows you to blend the authority between the throttle position and MAP.
John Meaney (the guy behind most of the aftermarket EFI systems like FAST
and Big Stuff3) says his latest and greatest Big Stuff3 unit allows the user
to tailor the cell width for both rpm and load (MAP or TPS%) and that by
increasing the resolution I can get the sensitivity I'm after.  Still not
sure if I buy that yet.

Supercharged applications tend to have cams with less overlap so speed density
or mass air may be better.  Supercharged applications using Alpha-N require a
MAP upstream of the throttle to act as a barometric compensator.

Independent runner manifolding will tame down a big cam.  Since the runners
are isolated from each other, reversion from adjacent cylinders does not
foul the intake stroke, allowing a longer cam duration with a streetable
idle.  Kirby Schraeder runs a PPC-sourced IR EFI system on his 377 cubic inch
Cleveland stroker (iron 4V heads with Weber lower and 48mm TWM throttle
bodies).  He runs a fairly large overlap 288FDP Crower solid flat tappet oval
track cam on the street.  Specs on his cam are 254/258 degrees at 0.050"
(288/294 degrees advertised), 0.569"/0.580" lift (0.022"/0.024" clearance hot)
with 105 lobe centers.  That's a lot of duration and tight lobe centers for a
street car.  According to Kirby, with a 700DP Holley on a Ford aluminum dual
plane intake manifold, it had a wild idle and wouldn't start pulling well
until 3000 RPM (Crower rates the cam range as 3500 to 7000 RPM).  When he
installed the independent runner EFI, the first thing he said was "Where'd
my idle go?".  He noted it now pulls 5th gear from 1500 RPM.  Kirby also
noted it's tough staying off the 7200 RPM rev-limiter in lower gears.

Independent runner also allows tuning of the inlet tract length generally not
possibly with single 4 barrel plenum type intakes.  I know a lot of people
like the looks of the cross runner manifold but I suspect for many applications
the runners may be too long.  I ran a series of Dynomation simulations to pick
the runner length for my Fontana stroker.  On that engine, the short stacks
that fit under the stock Pantera engine cover and decklid is close to ideal
length.  Longer runners increased low and mid-range torque but that came with
a big loss of HP in higher RPM ranges.  With a converted carb intake, you're
stuck with the runner lengths cast into the intake.

A big benefit of EFI is being able to adjust the spark curve easily in ways
a mechanical system won't permit.  If you want to lean out the mixture at
cruise for best fuel economy, you'll also need to adjust timing.  Combustion
gets much slower under lean conditions and if you don't adjust spark timing,
the combustion occurs much later and exhaust temperature climbs.  That's bad
for the seats and valves.  However, if you adjust for MBT spark at each A/F
ratio, exhaust temperature will actually decrease relative to stoichimetric.

Narrow band O2 sensors have a very steep voltage output versus O2 content
and basically just toggle between rich or lean trying to hold the air-fuel
mixture at stoichiometric (around 14.7:1 for typical pump gas).  Narrow
band sensors are employed because catalytic converters are happiest at
stoich.  For power you want to go slightly rich and for economy you want
to go slightly lean.  For tuning purposes (a human using the output to tune
a carb or EFI), a narrow band is all but useless.  In a computer, you may be
able to datalog and get usable info.  In general, wide band O2 sensors will
yield a much more usable output.  Narrow band 02 sensors are generally only
closed loop during cruise.  At wide open throttle, the control schemes
usually revert to tables.  Some of the latest aftermarket controllers (e.g.
BigStuff3) are using wide-band O2 sensors for data logging and also in "learn"
modes. They also have a simple model built in to get you in the ballpark for
start-up operation.  You simply enter your bore, stroke, compression ratio,
etc and the computer defines a start-up map.  A downside to using a wide band
O2 sensor closed loop all the time is when the sensor fails it could lead to
a dangerous lean condition that could melt a piston.

BTW, most of the controllers are designed around serial data interfaces but
most of the laptops have gone to USB ports.  There are supposedly
USB-to-serial adpaters but you may need driver software upgrades.

One big thing in favor of Ford or GM-based systems is BIT.  I'd wager
a majority of the code in those ECM's is for built-in test and diagnostics.
That helps a huge amount when trying to debug a system.

and that's just some of the basics...

Dan Jones

> *When asked about the effects of reversion their response was that the
> "throttle body/ adapter/ meter combination work together to be virtually
> unaffected by reversion".

Simply not true.  Due to the mass air sensor location, their system is worse
than the stock Ford mass air set-up which is itself susceptible to reversion.
It's physics that applies everywhere except in the marketing universe.  For
many engine combinations, this set-up will work fine.  For others, it will
not.  If you have a lot of compression and overlap, I'd look elsewhere.

> *The system uses a pair of "conventional O2 sensors not wideband".

Their system uses a stock Ford EEC-IV computer which uses narrow band O2's.

> They seem to agree that is designed to be plug and play "requiring no tune up
> to get up and running". I just wonder if it is possible to tune this system
> properly once installed??

You can tune it with one of the aftermarket tuners (Eectuner, TwEECer, PMS)
if you buy the right processor.  Ford makes many different processors and not
all supported by the aftermarket tuners.  Understand that tuning will not
overcome the mass air sensor limitations.  There is no magic here.  Mass-Flo's
EFI systems use a factory EEC-IV mass air processors with no modifications
other than a device to convert the GM frequency output to the analog voltage
the Ford EEC expects.  The Mass-Flo systems are based on the A9L Mustang EEC
which is supported by all of the above mentioned tuners.

You can duplicate the Mass-Flo system with a single plane intake (set up for
injection with injector bungs), a 90 degree elbow with throttle body, a Ford
EEC-IV processor, and a mass air meter matched to your injector size.  That
set up would have the added benefit of allowing you to move the mass air
sensor, if necessary, and would not need the extra electronics that conditions
the GM frequency output to work with the Ford analog voltage input.  Also, the
large single disc style throttle bodies usually make a bit more power than
the 4 hole carb style.  A friend runs that set up on his drag race Mustang
with a PMS piggyback controller.  Alternatively, you could do the same thing
with an aftermarket speed density or Alpha-N ECU (without the mass air meter).
That approach will be more flexible but will require additional tuning.

> {short back-story - my thoughts are this system may have the potential to
> match the performance of properly tuned race carburetor with the added
> benefits of fuel injection.....drivability, easier starting, fuel economy
> ;o), etc....  I still believe an IR injection system would be ideal but I
> am exploring my options with a single carb style manifold right now.}

The manifolding is a separate issue.  The 4 barrel carb intake approach is
fine if you choose the right control approach.  Previously, I gave a general
overview of the three basic engine control schemes and what their limitations
are.  If you choose to use the mass air approach, pick your cam accordingly
and read the Probst book on the Ford EEC.

> Also, the "pie in the sky" method that they describe their system is somewhat
> annoying.

That's a big red flag to me.  When a company makes claims like that they are
either incompetent (don't understand the limitations of their system) or
dishonest.

> This is not a good analogy but here is goes anyway.  If this system was the
> answer to all prayers, then engines in Formula 1 and all other forms of
> performance vehicles would be using it.

Ford ran the speed density EEC in Formula 1 as the mass air approach was
less reliable (a mass air meter in each stack to fail) and made less power.

Dan Jones