Fuel injected lawn mower

[Andy Harrah]

Jeff,

> Yeah, I bet that alternator heats up pretty quick sinking 3500+ watts.

I've never run with much of a load, so it doesn't 'over' heat.  I did have it
connected to an old worn out 3 hp Briggs from my garden tiller.  At a fast idle,
if you crank the load up on the 'brake' it drags that motor to a stop almost
instantly.


Applying a load to a motor under test I find is a huge problem.  Having the
rest of a car hanging off the crankshaft provides a great load but its a 
pain to drive and fiddle with a keyboard at the same time.  Doing it on
the little engines seems like an easier place to learn.  I've imagined all
sorts of schemes to build a dyno for a lawn mower sized engine.  None is
as easily controled as the alternator.

All four-stroke cycle lawn mower engines and almost all four-stroke motorcycle
engine fire every revolution.  All the newer distributor-less microprocessor
controlled car ignitions do as well.  That extra spark just flaps in the breeze.
You'll notice these ignition systems use blocks of coils with twin outputs.
They connect the pairs to cylinders that appear one full revolution apart
in the firing order.  One is actually firing, and the other is wasted.

My system takes interrupts off the TDC reference and calculates the RPM.
Based on the RPM, it looks up the advance from a centrigual advance type table.
This value is then used to predict when to fire the next time, in advance of
actually receiving the TDC reference pulse.  The reference pulse actually is
set to about 6 degrees BTDC so that it can be used to fire the coil directly.
This is used before the micro is up to speed, or if it crashes.
Currently there is no load input to calculations.  The TTL level output of the
micro drives the ignition coil through a VW Rabbit amplifier module.

By the way, that scheme of one pulse per revolution isn't accurate on any
engine that can rev very quickly.  I read somewhere that Ford had trouble
with the controllers for the Taurus SHO because that motor can rev at
20000 rpm / second which was too fast for their systems to keep up with.
I guess that's why all the Motronic systems take pulses off the flywheel
ring gears as well as a TDC reference pulse.

A few years ago I played around with the same thing on a Chevy 350.  On that
one I used a manifold pressure sensor.  Before I had my gadget working, I
hooked it up to the stock system.  I drove around and collected the factory
timing curves, RPM and manifold pressure.  I copied that data into my system
as-is.  It was built around an NEC V40 which is kind of like an Intel 80188
with a different set of on-chip peripherals.  I had an 8K static RAM, an 8K
EPROM, an 8-input ADC, 8 bits of dip switch input and 8 LEDs for output.

My current work is with a 68HC11.  I'm using this table driven assembler called
TASM which is kind of shareware.  For $40 the guy sent me the source code.  I
converted it to work as a Windows app.  ( Windowless, just takes a command line)
I'm using Visual Basic to draw control panels with meters, etc.  The 'HC11
uploads data to the laptop over the serial line.  I'm hoping it'll prove to be
a nice setup.

I have a bunch of Chevy 305 and 350 injectors that I got for free, too.  I don't
think they'll work on the Briggs.  The per cylinder displacement of the V8
is about 625 cc.  The Tecumseh engines I've got range from about 125 to 200 cc.
Late model VW Bugs had L-Jetronic injection.  Those motors are 400 cc per cyl.
There must be some tiny motor in some imports that would match closer.
I'm very worried that the dynamic range of the injectors will be a problem.
Perhaps running a lower fuel pressure would help, but that might foul up
the spray pattern.

I believe the minimum pulse length for the injectors is about 1 to 2 ms.
There are also considerations on the electrical drive for the injector.  They
come in different forms.  Some are designed to be used with an external current
limiting resistor, and some are supposed to be connected to more intelligent
drive electronics that watches the current ramp until the inductance saturates
and then back the current down to a 'hold' level.  There are some decent
discussions of this in that blue Bosch injection book. Also, the Motorola
and National Semiconductor linear databooks describe chips to control
injectors.  Not that we can get them, but they talk about how they work.

I'm planning on using a VW CIS fuel pump, just because that's what I can get.
I'm not sure what I'll do about the regulator.  I've got a lathe, and I'm
reasonably handy with it.  I might try to make one.

I've got a VW Corrado VR6.  Someday I'd like to fiddle with the programming
in that beast.

.../Bill