DIY_EFI Digest V4 #459

John Dammeyer johnd at
Tue Aug 10 04:58:58 GMT 1999

Phil wrote
>Hi All,

John wrote
>> Why would Honda put the MAP sensor just behind the throttle
>Good question. I guess from the tone of what you wrote that you
>suspected that this could be an issue.

Well,  I thought it a bit odd.  After all,  the Fuel Pressure regulator
is attached to the middle of the plenum.

>I agree with you it probably is. They could of course have a
>table of correction factors for air temp vs non linearity vs

True,  and perhaps at the high end it really doesn't matter.

>Next issue is why are they using an ambient sensor ? Doesn't the
>know what air pressure it is subject to ?

Guess you've never owned a SAAB 99 with mechanical fuel injection.  ;-)
Whenever we took it on holidays from home at 2200' ASL down to Sea Level
we always found the engine ran poorly.  Got it back up to home altitude
and that wasn't a problem.  The 900 series etc.  use MAF so they don't
care;  they already know how much air is being sucked into the engine.
Other vehicles with MAP use some sort of ambient air pressure sensor.  I
even disected an after market fuel injecton box and sure enough,  they
had two Motorola Pressure transducers inside: one for ambient and one
for MAP.

The 'Grippo' formula uses MAP/(MAP at WOT) to scale the PW.
Additionally there is a second part of the formula that uses MAP at Wide
Open Throttle (I had assumed Ambient Air Pressure due to the location of
the HONDA MAP sensor) minus Fuel Vapour pressure divided y the
temperature.  This number gets smaller as either ambient pressure goes
down or temperature go up thereby reducing the PulseWidth;  less air,
less fuel required.  Kind of like reaching 8000' ASL and adjusting the
Lycoming's mixture on on your Cessna till EGO readings are correct

Does anyone know what the Fuel Vapour pressure for Gasoline is or where
to look it up on the WEB?  I'm assuming (perhaps wrongly) that it's the
pressure generated by the fuel as it evaporates.  The number used in
Grippo's document suggests 3.1Kpa but that's at 85F.  What would it be
at -40F.

Interesting too,  if the MAP is taken somewhere further inside the
plenum then the WOT value will probably be lower than Ambient Air
Pressure.  If you used one sensor,  then you could save Ambient Air
pressure the moment the ECU powered up just before cranking.    Of
course if the vehicle runs for an hour straight up a mountain then the
fuel ration will just get richer and richer.  So the trick is to reduce
the PW as the air density decreases.

Al Grippo's equation reads:

Air Density = (0.0391568(Barometric Pressure - 31))
                     (Manifold Temp/10)    + 459.7

Barometric Pressure,  Manifold Temp and the 31 (3.1 Kpa) are all scaled
by 10.  Bruce Bowling does the same thing but in "Hg.  In either case,
this provides the number required to shorten the PW as altitude or
plenum temperature increase.

Next we get to drag our test frame with GCF (remember... Giant Computer
Fan)  up a small hill from Sea Level to 2500'.  Hopefully I should see
the O2 sensor keep the same values but I'd like to test at 10,000
without leaving the ground.

>> For a race car calculating a new injector value PW every stroke
>may be a
>> requirement but is that really needed for marine or aircraft
>> applications?
>Yes. and on second thoughts  Yes.
>If the calculations are done per cycle then the sudden throttle
>load change will encounter maximum of 2 bad combustions.
>Now that probably sounds excessively picky but if your engine
>goes from
>250 hp to 125 hp in 1 revolution and then back again the stress
>cycles on the crank are enormous. Even bailing wire has a cyclic
>strain limit.

Good point.  We'll call that the BWCSL.  ;-)

I understand what your saying though.   I'll check through my code and
see how many lean combustions I might have.  Obviously MAP will lag TPS
so a change in TPS is probably a better indication that it's time to
enrich the mixture.

>Question ? Who blows up all those F1 engines.

I'd like to try!!!!

>Try creating a 4 byte map.
>MAX. Vacuum  1000 rpm    Max Load  1000 rpm
>Max vacuum    6000 rpm    Max Load 6000  rpm
>if you interpolate b/w these 4 points and each on it's own is
>then you will be on average spot on. (This doesn't include
>camshaft timing alterations) this would require the 6 point map.
>By 65,000 points I didn't mean memory locations I meant possible
>calculated values from rpm 8 bit X vacuum load 8 bit.
>thus 255 x 255 = possible outcomes.

OK.  I see what you mean.

>You will at some stage have to get those 4 measurements as they
>are the basis for all further corrections.
>Full load 6000 rpm is easy just pull the stick all the way and
>the prop will do the rest.
>Full load 1000 rpm has more air  per cycle than 6000 rpm so give
>it 10% more

We can bog the engine down with the dyno to get that.

>Light load 1000 needs the rubber band to be removed.

Hmmmmm..  I think you forgot the page number of the Bosch Blue Book for
RB (rubber band) analysis. ;-)

>There ! that's  4 bytes of eeprom all used up.

You make it appear really simple.  I guess in many ways it is, after all
a simple SU carburator can do it.

>You have only to get rpm from the timer ticks vs a single A/D
>Vacuum Load and you are ready to go. Your last air, water and
>accel corrections are waiting from the last stored calc. apply
>these and squirt.
>Given a 1 mHz internal clock and 6000 rpm you have 10 ms or
>10,000 clock cycles to get this done. say 2000 instructions
>at worst.

Yes,  but I want to be able to play space invaders too.  8-).

> P.S. if this to and fro is boring the pants of others please
>chime in and tell me firmly.

Shussh everyone.  This beats the pants off the standard questions of "I
have a xxx ECU for my sports car and I want to modify the tables ...
what do I do?"



More information about the Diy_efi mailing list