<snip
I'll assume for the moment that the MAF output is the square root of
the
flow.
This might sound strange but perhaps the best way to make a MAF linear
is use an 14 pin 16F684 pic chip to PWM drive a injector switching
FET and do away with the 555. The program could be just be a look up
table loop setting the PWM. The price in quantity is under a 3 bucks.
http://www.microchip.com/wwwproducts/Devices.aspx?dDocName=en010214
Another possibility is use a manifold pressure sensor that has a 0
to 5 V linear out put to control a 555 instead of a MAF. We have a
lot of experience with those as that is what Tracy uses to feed his
computer.
Paul Lamar
HOLY COW PAUL!
*NOW* you're going to a microcontroller!?!
Does this signal the end of the MAF/555 experiment?
IMHO - YES YES YES!!!
1) use a microcontroller to A/D the MAF voltage
2) use a lookup table to determine duty cycle
3) alter pulse width with the INVERSE of RPM (divide by RPM)
4) Inject-Per-Revolution (IPR) the resulting pulse width
5) go to sleep until next injection trigger
Damn that was easy. Right?
If you want a Speed/Density (MAP sensor) system, there are a few more
steps, but still "simple". Start with MegaSquirt FAQs to avoid
reinventing the wheel. They have done A LOT of the hard work and have
proven systems out there.
*****
Since you keep plugging PIC chips, I'll return the favor by plugging
TI chips (I have no relationship with TI). The Texas Instrument
MSP430 chips cost under $1 in quantity. BUT ... you can get a
COMPLETE dev kit for $20 - try and beat THAT for getting started!
http://focus.ti.com/docs/toolsw/folders/print/ez430-f2013.html
In fact, I think the MSP430 chips beat all others point-by-point,
across the board. CERTAINLY worth consideration for new projects with
no historical design baggage.
I know you don't want to learn a new instruction set and
microcontroller family (reasonable), but the TI MSP430's are really,
REALLY nice low-power high-performance RISC chips that are very
elegantly laid out - from the instruction set and registers to the
onboard peripherals. It's a real pleasure to work with these as
opposed to the Intel and Motorola microcontrollers. The best day in
my recent EE life was when an engineer friend twisted my arm into
buying the $20 USB dev kit. I haven't looked back since.
I recommend these for anyone starting out and learning to play with
micro/embedded projects. After all, $20 isn't much to risk when
learning something brand new. (Contact me offline if interested:
vince at pacstates dot com - I'll pay it forward and twist YOUR arm.)
-Vince Orton
Weird Yes? The straw that broke the camels back was the non linear nature
of the MAF. I can program the chips as I already have a Motorola
/Freescale
programmer and the experience to do it. I'll sell them for less $5 plus
shipping.
You will need four as I would still do the 555 modular concept of one
controller per injector.
Or you can do it Vince if you wish at whatever you feel like charging.
Does the TI chip have A to D and PWM out?
Paul Lamar
Doesn't that USB iDAQ have a PWM output capability? I would think that
thing has everything you need right there. Just need to make your own
program for it.
Sad to see the 555 sytem dropped so suddenly.
C. Smith
Yes I think that might work. I see programming it as a problem.
It has an immediate application, as is, to record the fuel map as a
function of engine RPM and intake manifold pressure. I need to get
mine back
from Mark Suspinski. Hint :)
BTW if you are having trouble tunning Tracy's system this device would be
handy to determine what the fuel map is at any RPM and intake manifold
pressure level.
Back to the 555.
Think of the microchip as a sophisticated 14 pin $5 555.
We will call it the "Five dollar 555 system" :)
It does the same thing as a 555 with about the same number, more or less,
resistors and capacitors. Instead of responding to an ideal liner voltage
with air flow function it responds to the square root of the voltage
out of
the actual MAF with a linear increase in pulse width.
We will need a new PC board design but everything else about the 555
system
will be the same.
If we find that the engines run consistently rich or lean in some
RPM/power
range we can adjust the chip and swap it out if it is plugged into a
socket.
No big deal. $20 bucks for a new set of chips. In fact since they are
flash
memory we might be able to merely re-flash them.
It only takes a few minutes re-flash them. You can also store an infinite
number of maps on your hard disk for use or re-use and people can share
maps. I can save them and offer ready to go maps burned on the chips. Or
Vince can do it all.
Paul Lamar
Paul, I for one am not ready to toss the 555 chips.I agree that
the early design stages need to evaluate possible alternatives
and points of view. I don.t consider writing c++ programs and
mapping look up tables for one engine configuration and applying
it to one with different intake an exhaust systems to be the
best solution. The 555 can direstly drive the mosfet which runs
the injector and I downloaded a circuit which allows it to drive
four mosfets. In the interest of simplicity and reliability I
propose one trigger magnet and two hall effect sensors, one per
rotor.I would put the complexity of operation on the circuit
board rather than trying to mechanically align four to eight
hall effect sensors. As for the multiple oscellator ideas. NO
I say NO!!! each firing must have the same mixture. By the
time you start leaning some charges will be too lean to fire or
worst yet detonate. This would require an overall richer charge
and higher fuel consumption.Pulse width modulation is the
obvious plan. Wider pulse for more fuel. If the time available
for fuel flow at 7000 RPM at 70% duty cycle is too short you
need larger or multiple injectors per rotor. I see this as a
sysem of timers.And a timer is what the 555 is. I would set the
hall effect sensor to fire the maximum spark advance point, then
thru the Schmidt trigger count down two timers, one for each
plug to fire the plug with the count down driven by a resister
on the drain pin of the chip.I would like to see the injector
chip also time delay the start of the injector pulse to center
the 70% maximum duration of injector pulse to coincide with the
rotor intake.Once turned on the injector would flow until the
reset pin on the 555 chip is signalled to turn it off. This
would be done by a voltage comparitor. The output of the drain
on the 555 falling voltage is compared to the 0-5 volt mass air
flow output voltage and at convergence trigger the reset on the
injector driving 555.
The manual leaning would be a panel mounted variable resistor in series
with the mass air flow output feeding the comparitor. I realize
that going from two to five 555 chips per rotor sounds like a
lot of complexity and a physically larger circuit board which it
is. But all the complexity is on the board with minimal
hardware on the engine to adjust, and a means to richen the
mixture for take off and climb and lean for cruise. The basic
thought is to provide a consistant mixture on each rotor filling
and fire the spark plugs at the propper time. I view the Ford
mass air flow sensor as a perfect part for this application. It
by its very nature adjusts for density altitude and factors air
temperature and humidity error on circuit board on the sensor
and outputs a single variable voltage. The rest is mostly
timing the events of ignition and fuel pulse width which is
exactly what the 555 chip does. There is an online calculator
which allows you to enter resistor and capacitor values for the
555 and gives the time interval the chip would produce. I am
currently collecting circuit schematics off the internet which
would achieve these goals.
I am currently looking for one which flip flops the voltage
comparitor in the event the 555 drain is
at a lower voltage than the MAF and would not allow the voltage
signals to converge at the comparitor. My first search on
google came up with 1,200,000 results so it will be out there
somewhere. The problem with being older than dirt is so many
ideas so little time. I still remember when they invented dirt
and thought it would be a good idea. I should have bought some
stock in it at the initial public offering.
george grimes
You won't have an competition :)
I am going with the $5 555.
Paul Lamar
--
The Rotary Engine NewsLetter. Powered by Linux.
ACRE NL web site.
http://www.rotaryeng.net
Youtube key word UTUBPLEASE
Copyright 1998-2009 All world wide rights reserved.