Subject: EFI block I/O diagram
From: Rotary Engine
Date: 2/21/2007, 8:52 PM
To: AARotary Engine



    >>      >     Attached is the Input-Output block diagram for the
   simple EFI
    >>     system.
    >>      >     This diagram is intended to discover any missing
   sensors or
    >>     engine
    >>      >     outputs (i.e., other than for pilot information).
    >>      >
    >>      >     Please comment, especially omissions. I would rather
   provide
    >>     for a
    >>      >     sensor up front rather than have to patch or edit the
   code
    >>     later on. I
    >>      >     realize that not everyone will want all the sensors.
    >>      >
    >>      >     Fuel flow can be calculated from the fuel pressure
   given a
    >>     calibrated
    >>      >     hole in the slide throttle. The two pumps on each side
    >>     (tank?) can be
    >>      >     operated in parallel, i.e., both on all the time; in the
    >>     event one
    >>      >     fails, the uC can notify the pilot and increase the duty
    >>     cycle of the
    >>      >     remaining good pump in order to maintain proper fuel
    >>     pressure. Or, a
    >>      >     switch could be provided for the pilot to switch to
   the other
    >>     pump.
    >>      >     Builders choice.
    >>      >
    >>      >     The uC system will have total redundancy a la Tracy,
   i.e., an
    >>     A uC and
    >>      > a
    >>      >     B uC. If one fails, switch to the other and land.
    >>      >
    >>      >     Doc Custer
    >>      >     Retired Software Engineer and sometime Audiologist
    >>      >
    >>      >     -----------------------------------------------------
    >>      >     Looks OK to me Doc. I am glad you did that diagram.
    >>      >
    >>      >     I am thinking out loud now.
    >>      >     You can get fuel flow from the pulse width you are
   feeding
    >>     the pumps.
    >>      >     It must be calibrated because all pumps do not in general
    >>      >     pump an amount of fuel directly proportional (in a
   linear
    >>     manner)
    >>      >     as a function of the the pulse width. Even tho the
   average
    >>     voltage
    >>      >     applied to the pump is directly proportional to the pulse
    >> width.
    >>      >
    >>      >     In other words applying a 50% wide pulse to a 12 volt
    >>      >     pump is the same thing as applying 6 volts DC. In any
   event
    >>      >     you may need a voltage regulator on the 12V supply.
   If the
    >>     battery
    >>      >     is 14 volts the pump gets 7 volts.
    >>      >
    >>      >     Another way of doing it is feeding the battery
   voltage to a
    >> 1/3
    >>      > divider
    >>      >     and feed the output of the divider to one of the A to
   Ds in
    >> the
    >>      >     chip so the computer knows the battery voltage at all
   times.
    >>      >     It then uses a look up table to adjust the fuel pump
   pulse
    >> width
    >>      >     according to the available battery voltage.
    >>      >
    >>      >       This is the cheap
    >>      >     way to do it as you don't need a high power voltage
   regulator
    >>      >     on the 12 volt supply. If the battery voltage is low
   you widen
    >>      >     the pulse to compensate. Pumps can take up to 7 or 10
   amps
    >>     and as far
    >>      >     as I know there are no  single chips regulators that
   will
    >> handle
    >>      >     that. Well at least the last time I looked. In any
   event you
    >> can
    >>      >     display the battery voltage on the instrument panel as a
    >> trouble
    >>      >     shooting aid.
    >>      >
    >>      >     The other possibility is to ignore the voltage level
   as most
    >>      >     alternators have a regulator that is pretty much set
   to 13.5
    >>     or 14
    >>      >     volts or so. Assume that will always put you in the
   mixture
    >> ball
    >>      >     park and the pilot in the loop will fine tune with the
    >>     leaning knob.
    >>      >
    >>      >     The EGT is a bit hard to use as a feed back sensor as the
    >>     response
    >>      >     time is very slow and
    >>      >     it changes direction as you pass through peak EGT.
   You are
    >>     better off
    >>      >     with the O2 sensor IMHO if you are looking for a
   closed loop
    >>     system.
    >>      >     I would forget the closed loop and keep  the pilot in
   the
    >>     leaning
    >>      >     loop. The engine will run over a wide mixture range.
   10:1 to
    >>     20:1.
    >>      >     Air/fuel.
    >>      >
    >>      >     EGT rises as you lean the mixture and then drops as the
    >> mixture
    >>      >     gets leaner and leaner past peak. A tough nut to
   crack as far
    >> as
    >>      >     programming is concerned. You can lean the mixture
   and see if
    >> it
    >>      >     rises. If it rises you are on the rich side of peak
   EGT. If
    >>     it lowers
    >>      >     you are on the lean side of peak EGT. This takes a
   lot of time
    >>      >     in addition to the slow response of the TC. .
    >>      >     the only way to speed up the response of a thermal
   couple is
    >>      >     use a smaller one. That has life limit implications.
    >>      >     --
    >>      >     Paul Lamar ...No rotor no motor.
    >>      >
    >>      >
    >>      > Doc,
    >>      >
    >>      > Fuel pressure alone won't give you fuel flow. You will
   also need
    >>     to use
    >>      > the MAP value. You need the pressure differential across
   the hole
    >>     to get
    >>      > your flow rate.
    >>      >
    >>      > Do you plan to use any barometric compensation?
    >>      >
    >>      > Ken Bauman
    >>      >
    >>
    >>     Good point Ken Bauman about the pressure differential,
   especially in
    >>     a low
    >>     to moderate pressure system.
    >>
    >>     I hadn't considered barometric compensation as I assumed
   manifold
    >>     pressure
    >>     would suffice. Is this wrong?
    >>
    >>     Doc Custer
    >>     Retired Software Engineer and sometime Audiologist
    >>
    >>     We can simulate it by pulling a vacuum on a glass jar and see
    >>     how long it takes to inject X amount of fuel at Y pulse width.
    >>     I have an electric vacuum pump.
    >>
    >>     Paul Lamar
    >>
    >>
    >> Took this from an old MegaSquirt posting.
    >>
    >>
    >>  > > I am totally confused here. My experience with reciprocating
    >> aircraft
    >>  > > engines climbing to altitude is that it gets leaned out, both
    >> carb and FI.
    >>  >
    >>  > (Yeah, I know. A bunch of you are thinking "here we go
   again," but
    >> I've done
    >>  > this often enough that I think I can answer the question
    >> satisfactorily in a
    >>  > single e-mail.)
    >>  >
    >>  > Imagine an engine with constant VE, it's 100% across the tables.
    >> We only
    >>  > run this engine at 70f, so the only variable is manifold
   absolute
    >> pressure.
    >>  >
    >>  > Throttle the engine to 50 kPa at sea level with an ambient
   pressure
    >> of 100 kPa,
    >>  > and pulse with is, for example, 5.0 ms. Now fly up to 5000 feet,
    >> ambient
    >>  > pressure drops to 84 kPa and make sure we are still throttled
   to 50
    >> kPa in
    >>  > the manifold. What is our pulsewidth now? Correct, it is _still_
    >> 5.0 ms.
    >>  > Change to any altitude you want, but as long as you have a MAP of
    >> 50 kPa,
    >>  > the PW is 5.0.
    >>  >
    >>  > Well, this is only taking into account the intake side of the
    >> engine! In
    >>  > a real engine, the volumetric efficiency is not dictated
   solely by
    >> the intake
    >>  > tract, but also by the exhaust system (headers, cam timing,
   exhaust
    >> valve
    >>  > size and all that can change the power output/VE, right?).
    >>  >
    >>  > Imagine we are running the engine in a dyno cell ambient 100 kPa,
    >> 50 kPa of
    >>  > MAP, resulting VE = 100% and we get that 5.0 ms PW. Now put a
    >> vacuum cleaner
    >>  > on the end of the exhaust pipe, and pull it down so that the
    >> exhaust system
    >>  > is under a perfect vacuum (extreme case, but gets the point
    >> across). The
    >>  > VE magically goes way up, because we've got a super extractor
    >> exhaust now.
    >>  > Suddenly the engine goes lean, because the VE went up, but we
    >> didn't know
    >>  > it inside MS because we only measure MAP, not the ambient
   pressure
    >> at the
    >>  > exhaust outlet!
    >>  >
    >>  > This is _exactly_ what happens as you gain altitude. The MAP
    >> sensor corrects
    >>  > the fuel pulse based on MAP, so altitude changes are of no
    >> consequence here,
    >>  > but MS has no way of knowing what is going on in the exhaust
   system.
    >>  >
    >>  > Here's the bottom line:
    >>  >
    >>  > Barometric correction ENRICHS the mixture as you gain altitude to
    >> correct
    >>  > for the increasing efficiency of the exhaust system.
    >>
    >> Ken
    >>
    >> Interesting. Looks like we need another pressure sensor.
    >>
    >> Paul Lamar ...No rotor no motor.
    >
    > I agree. Thanks Ken.
    >
    > Attached is the latest diagram.
    >
    > Doc Custer
    > Retired Software Engineer and sometime Audiologist


   why 2 manifold temps?  would there really be enough difference to
   need 2?
   how could fast moving intake air change a significant amount during the
   brief period it passes thru the intake?   kevin


Redundancy.

Doc, that will be a very nice setup. Are you talking 4 fuel pumps? Or 2
pumps with redundant PWM? How will the torque sensor work?

Ken

The Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site. http://www.rotaryeng.net
Copyright 1998-2006 All world wide rights reserved.