Bench racing ecms

[Ludis Langens]

Sorry for not responding earlier - I've been spending too much time in
technology recycling facilities.

"Bruce Plecan" <nacelp at bright.net> wrote:
> A 50% dwell frequency generator for VSS,  hopefully this will work,
>     and not have to bother with Buffer/amp, or DRAC..

There is no need for a 50% duty cycle.  The C3 hardware measures one
cycle of the VSS (from one ___ edge to next ___ (same) edge.)  Further,
all the ECM software I've seen averages two of these cycles before
calculating vehicle speed.  The custom VLSI chip does provide a way to
directly read the raw VSS (and MAF and knock) input, but I haven't yet
found any software that makes use of this feature.

> Some 100 ohm resistors for like charcoal canister/ EGR solenoid
>     loads.

Dummy loads aren't needed with the '7747, except in the EST module
interface.

> Generating a MAF (hot-wire, hot-foil) type of signal for the ecm.  So far
> got a shop vac. for generating flow across an actual sensor.

Frequency generator.

> 12v lawn tractor battery, and circuit breakers.

For my ECM hacking I'm using an adjustable dual output RatShack supply. 
One output provides VBat, the other provides VIgn.  This makes it easy
to make the ECM do its key off power down operation - I just twist the
VIgn adjustable output knob to zero.  I don't need to unclip any wires
or fully power down everything.

There are no external loads the way I'm using the ECM.  The ECM's power
draw of less than 500 mA is well within the 1 to 1.5 amp limit of the
supply.  I don't know how the RatShack supply would react to running a
full set of real injectors/solenoids/motors.

> With as much talk as the O2 being handles as a switch crossing the
>     .44 as a O2 sensor crossing count was thinking of a divide by
>     rpm signal source for generating a 0-.8v for O2 input????....

Here's an idea similar to one that was already posted here.  Do a duty
cycle to voltage conversion of the injector output (i.e. an RC filter.) 
Filter this voltage through another RC filter with a longer time
constant.  Use an opamp (open loop like a comparator) to compare the
voltages at the two filters.  (Use an opamp here because of its
push-pull output.)  Resistor divide and bias the rail-to-rail opamp
output to about a 0.2 to 1.0 volt range.  Throw a capacitor into the
output node of this resistor network.  Add voltage followers to your
personal taste.

The idea is to create a long term and short term measurement of the
pulse width.  If the short term is bigger than the long term, tell the
ECM it is running rich.  Otherwise say it is lean.  This will tell the
ECM that its pulse widths are bouncing about the correct point.  The cap
in the final output network simulates a slow reacting O2 sensor.  It
also filters out the high frequency (injector pulse rate) erroneous
switching of the opamp/comparator.  You may need to use pots in place of
the R's in the filters to tune this circuit for different engine rpms.

> A rpm signal generator, but have to work out an interface to module
>      is a transformer necessary?.  Ya think just a 1.2v DC pulse would
>     fire the module?.  Use a monstable 555 for the 1.2v, and then
>     an adjustable astable for rpm?.

For my ECM hacking, I've been running my own software in the ECM.  This
software plugs away at the control registers of the I/O chips to try to
figure out the register definitions.  For many ECM functions I've needed
to simulate an ignition module reference input.

Generally, what I do is this:  One of the three PWM output channels gets
programmed to generate an output waveform.  This output is then fed back
into the ECM via the EST interface.  Because the same internal clock is
used for both the PWMs and fuel/ignition timing, I get nice stable
measured values.  Anyway, the EST inputs are very high impedance.  I
originally looped the (PWM) quad driver output back in (with a pullup to
5 volts.)  But it also works to take the NMOS VLSI chips PWM output and
directly loop this to the EST input.  Just remember, there are two EST
reference signal inputs, one of which needs to be grounded.

-- 
Ludis Langens                               ludis (at) cruzers (dot) com
Mac, Fiero, & engine controller goodies:  http://www.cruzers.com/~ludis/

    The time to notice that one corner of a (junk yard)
    car is supported by a stand under just a piece of
    plastic trim is not while you're under the car.