FI for Mini

John Dammeyer johnd at autoartisans.com
Wed Oct 6 17:58:55 GMT 1999


Hi All,

My two cents here after finishing off a FI system for a Honda.  I did this
incrementally with a minimum of sensors to start with.  What you are really
interested in is calculating at any RPM the amount of air that will flow into
the cylinders at any RPM.  To determine the flow and therefore the air mass you
need to know the pressure outside the throttle plate and the pressure on the
other side of the throttle plate inside the manifold (MAP).  If you assume
standard pressure and temperature you end up with only needing MAP.  An O2
sensor helps here for setting up a volumetric efficiency table.

Calculate the amount of fuel the engine will need for 14.7:1 air fuel ratio
given cylinder size and injector size.  This is your starting point and is
usually a number in milliseconds that you convert to timer ticks so your micro
can program a counter to trip on and off the injector.  This value assumes 100%
volumetric efficiency.  i.e. when the intake valve closes and the piston is a
BDC the airpressure inside the cylinder is atomospheric. (almost never happens
except at cranking with throttle wide open and cranking is a special case)

Next determine the ratio of MAP/MAPWOT.  Assume MAPWOT is the same as
atmospheric pressure (emperically it's usually 1/2" Hg less than Atmospheric
(30.50"Hg)).  You can also determine MAPWOT when processor first starts up
before engine cranking begins; MAP == MAPWOT.   This ratio,  always less than 1,
is a quick way of calculating the mass of the air being pushed into the intake
manifold by the external air pressure.  Remember,  the engine doesn't really
suck in the air.  It creates a low pressure zone and the higher external air
pressure pushes the air in to fill the low pressure zone.  Just like our weather
systems.  The difference in air pressure determines how fast the air moves and
air always flows from a high pressure zone to a low pressure zone. (Can anyone
spell hurricane?).

Now,  because air has mass,  it doesn't start moving right away and doesn't stop
immediately either.  At idle RPM the piston is moving slowly so the change in
MAP is also slow.  This means with a closed throttle and low MAP (10"Hg) not a
lot of air gets into the cylinder.  The air movement has a chance to slow down
and stop moving before the next cylinder intake valve opens.  At high RPM the
piston creates this low pressure area much faster.  Although the throttle may be
totally open, and there is a small difference in MAP verses MAPWOT, the air is
moving quickly and doesn't have a chance to stop moving before the next cylinder
needs air so each cylinder is filled with more air molecules.

Therefore the amount of air that gets into the cylinder at any RPM is called
Volumetric Efficiency and is somewhat independent of Throttle position.  You can
use 200RPM steps and a single dimensional table that needs to be calculated for
VE.  Values range from about 0.3 at idle to 0.8 at WOT for street engines and
can exceed 1.0 for race engines with wild CAMs.

So the formula is PulseWidth = FuelAirRatio * (Map/MAPWOT) * VE[RPM];

This works but doesn't handle acceleration very well because the opening/closing
of the throttle can lean/richen the mixture before the math catches up.  So to
make the engine drivable,  it's a good idea to use a Throttle Position Sensor
(TPS) to add an enrichment parameter to the formula based on both absolute
throttle position and change in the TPS (dTPS).

You'll also find you need to use a different FuelAirRatio for idle verses
cranking verses running verses WOT.

A couple of extra points here.  Air temperature changes the density of the air
so an air temperature sensor will help set the mixture more precisely.  An
external air pressure sensor is also important because the vehicle may change
altitude without the engine stopping and restarting.  i.e. When you start, you
know atmospheric pressure because the MAP sensor reads static air pressure
before the engine starts turning.  But once the engine is running you may change
altitude by 5000' during a 1 hour drive.  At that point the Atmospheric Pressure
is considerably less than it was when the engine was started and so there is
less external pressure available to push the air into the intake manifold and so
for a given MAP value,  the mixture will be too rich.  This is why pilots always
have to lean out the carburator mixture as the aircraft climbs in order to avoid
fouling the spark plugs.

It's really that simple but implimentation and tuning etc. is still a lot of
work.  Don't take it all lightly.  The O2 sensor will give you a guide as to
where you are but the stock auto ones are notoriously inacurate on either side
of 14.7:1.  A second method of determining optimum mixture at high RPM is to use
Exhaust Gas Temperature (EGT).  A K type thermocouple inserted into the exhaust
manifold as close as possible to the valves results in about 1400F at optimum
power.  A temperature of 1600F suggests a too lean mixture and could potentially
damage the engine and a 1200F temperature is too rich.

Whew!  There's four months work in explained in a few paragraphs.

Regards,

John



>>Subject: How completcated does Efi HAVE to be?
>>
>>I am into minis, something a lot of americans probably won't have even
>>heard of never mind seen! Up until recently minis had single point throttle
>>body injection. You can pick up the inlet manifold, throttle body with the
>>following sensors, The system is a MAP system. I was just wondering how few
>>sensors I need to make it work. Make my own ECU and keep it simple.
>>Calibration would be done on a dyno using a lap top. No need for a Heated
>>o2 sensor, don't need it to learn really.
>
>There is a EFI Control System called the Alpha-N which calculates air flow by
>measuring only engine rpm (via coil) and throttle position via a Throttle
>Position sensor.  Pros are simple installation and compatible with radically-
>cammed engines.  Cons are no closed-looped operation, inferior part-throttle
>cruise metering and entire fuel map must be reprogrammed.  A MAP can be
>optionally
>used but mainly as a input for barometric pressure for different altitudes.
>This system is used for racing or heavily-modified street engines with big
>cams
>and low vacuum.
>
>Bob, Burlington, Ontario
>'97 FR CC Sport, 5.2L, 3.55 SG, auto., 15.234 @ 87.26
>
[snip]
>------------------------------
>
>I vaguely think that you only need a throttle position sensor and a map sensor
>to do a reasonable job of EFI. However, you WILL need the crankshaft position
>sensor, which means you need the flywheel, transfer case and the actual
>sensor.
>- --
>Tom Parker - parkert at ihug.co.nz
>           - http://www.geocities.com/MotorCity/Track/8381/
>





More information about the Diy_efi mailing list