FW: Processor selection for datalogger

[Roberto L. Landrau]

Hello everyone.  I am new to this list.  Please forgive me if I am
asking a FAQ (I didn't see indications of it on previous articles).  I
am not an FI expert trying to build my own controller from scratch,
yet, but I am trying to squeeze a few more horses out of my car right
now.  I would like to build a data logger to monitor a few engine
parameters.  I was wondering if someone on the list can save me some
time and ``trial and error'' iterations.  See below.

>>What I do is hook a ($20) LT1290 8 ch, 12 bit A/D converter to the
>>SPI.  4 wires (a data clock and an analog converter clock, data in
>>and data out).  A few lines of code and away you go!

>I've decided to stick with the HC11's A/D for the prototype version-
>I can live with the resolution limitation initially. Eventually, I
>will want better resolution, and this looks like a fairly good
>method.

>My only remaining concern is performance. I'll be monitoring:
>REF/EST (Timing), #1 Injector PW, VSS/DRAC (Vehicle speed), Coolant
>temp, Manifold Air Temp, 0xy Sensor, Barometer/MAP, Auxillary Boost,
>TPS voltage (position), Fuel pressure

Does anyone have a feel for what the injector PW signal looks like
(timing)?  I suspect it is a pulse between 1-10msec wide at idle.
The technical service manual for my car describes the sequence and the
crank angle at which the injection takes place, but it doesn't tell
you how wide the pulses are.

My specific question is: how wide is the injector pulse signal?  How
often do I need to sample to get a reasonable reading (i.e. is it
5msec +/- 0.000001msec, needing a 1e6 counter, or is it 1-100msec,
etc.)?  I realize this varies from car to car, but I am only looking
for an ``order of magnitude'' starting point.

Thanks.	

--------
Roberto L. Landrau	landrau at mitre.org
The MITRE Corporation	rll at linus.mitre.org
Bedford, MA 01730	rll at linus.UUCP