DIY_EFI Digest V4 #453

John Dammeyer johnd at autoartisans.com
Fri Aug 6 22:35:01 GMT 1999


-------------------------
>
>Date: Sat, 07 Aug 1999 03:03:04 +1000
>From: Phil Lamovie <injec at ains.net.au>
>Subject: Acceleration enrichment Re: DIY_EFI Digest V4 #450
>
>Hi All,
>
>For a simple solution to acc enrichment start by creating a FIFO
>stack of say
>ten counts. Read  TPS every 10 ms. Algorithm should respond in a
>linear fashion to the diff b/w first and last. Say 100% of
>additional fuel for 200% change in A/D value. A single integer
>value of X ms will give you an adjustable ramp rate. This is a
>simple feed forward system

Hmmm.  I'll see if that technique will work for my system.

>
>The harder part is to arrange a proportional decay vs rate of
>change of rpm


>
>Of course for an engine  that has a very poor chance of acc
>quickly relative
>to say a blipped throttle during a down change this is really of
>little import.


Yes,  a propeller at 2500RPM (6000 Engine RPM)  has a fair amount of
momentum (or lack of it at low RPM).  8-)


But that brings up an interesing point that we discovered today.  The
MAP sensor vent for the Honda engine is located about 5mm behind the
throttle plate.  Once the throttle is opened wide my MAP goes to
atmospheric.  At 5200RPM and wide open throttle the MAP stays at
ambient.  I see very little change or no change as the RPM increases.
Al Grippo in his fuel injection document discusses acceleration
enrichnment and derivative of MAP and TPS as the signal to do
acceleration enrichment.

That's all find and good but when should the enrichment stop?  For a car
engine in first gear a safety could trip it at max RPM.  A car going up
the hill in fourth with the pedal to the metal would either a) slow down
and stall if the hill is too steep or b) drop to a given speed and go up
the hill with perhaps a 12:1 A/F ratio. (I'm guessing)

In the case of a propellor load for water/air (assuming no cavitation)
the prop is usually sized so that the engine turns just under maximum
RPM at WOT (static).  The prop can turn faster if the vehicle is moving.
Carburated engines just smoothly increase in RPM to their limit given
the engine/prop combination.

So I think I can assume that if the delta TPS is large or for that
matter if the TPS is at maximum value,  that the mixture can be
increased to WOT at Max RPM.  (as a guide).   Then,  if either MAP or
TPS changes in the negative direction,  switch over to normal A/F curve.
ie: If at WOT and max RPM pulse width was 12.5ms then run injectors at
12.5ms until TPS or MAP changes.  At that point switch over to the
normal closed loop control algo. and the engine shouldn't miss a beat
but mixture should even out and fuel economy will return to expected
value.  (fuel economy on aircraft is _really_ important).

My SAAB Turbo has a switch at 70% TPS that tells the engine computer to
put the boots to it.  I would assume that a smaller change in TPS will
just increase the mixture based on a number of engine parameters with a
corresponding slow increase in acceleration.  That I already have since
I can slowly reach 5200RPM (loaded) with the gradual opening of the
throttle;  it just doesn't have a rich enough mixture when the throttle
is opened wide.


Suggestions,  comments?

>
>On the matter of duty cycle it's more to do with lousy linearity
>of an injector
>at either end of the scale 0-2 ms is considers to be unusable due
>to lack of repeatability. With 0.5 - 1.1 ms just to open this
>gives rise to variations cycle to cycle of more than 8 %. Makes
>it hard to pass US pollution laws.
>
>At the other end of the scale the rate of collapse of the
>magnetic field and the fuel pressure together with pintle bounce
>mean that 90 % duty cycle could in fact mean that the injector
>never closes.


That's what I've read too.  Also too long an on period causes heating
with a subsequent possibility of a stuck pintle.


>
>F1 engine at 16,000 rpm has 3.75 ms per rev losing 0.5 to open
>and 0.65 to close would give a max. of 2.6 ms. Thus one injection
>per 2 revs.
>And you were worried about timing the injection to coincide with
>valve train events.


sigh... I worry way too much.


>
>Be wary of any O2 sensor voltage vs  A/F ratio inferences. If you
>don't correct for temp there is a drift of 1.5 A/F over 400 deg
>c. And that's for the
>expensive BROAD BAND device. A single wire sensor can't be
>trusted
>at all. For any peace of mind use a 4 Gas analyser and EGT.



Yes.  I've found that it seems to work well under steady engine loads
but doesn't respond well as the engine load is changing.  Doesn't matter
much.  Loop opens up for accel, and decel and engine warmup and then
mixture defaults to pure table values.  This works well enough and just
on the rich side.  Today we ran the Hovercraft Fan (Giant Computer Fan
really),  up to 5380RPM and ran it for a few minutes.  Smooth and
mixture sat nicely at slightly rich.  Be nice to have EGT for this part.
Best with carburator has been 5200RPM.  Man is it loud.

>

>Be sure to refer to the properties of steel/wire/bailing  in the
>BLUE BOOK
>to aid with exhaust mounting redesign


Do I look in the index under SWB or steel/wire/bailing?  ;-)


>
>For the key to all things in auto tech The Bosch Automotive
>Handbook
>
>SAE  order no is BOSCH4 (4th Edition) It's blue !!  ISBN
>1-56091-918-3
>

I'll order one tomorrow.  Sounds good.

>DISCLAIMER .......I am a hopeless addict of the SAE book store
>and would buy everything if I could.


SAE book Store?  Try any book store for me.  We don't have paintings on
our wall,  we have bookshelves.

Regards,

John Dammeyer






More information about the Diy_efi mailing list