crank position sensor

[Tom Parker]

James Montebello <jamesm at talarian.com> wrote:

>The critical question is: how quickly (under normal operation) can
>the engine change speed?

In nutral mine can go from 2000 rpm to 6000 in probably less than a second. At
4000 rpm it's doing 60 odd revs per second, so in about 60 revolutions it's
speed tripples. Thats quite a big change per revolution, about 60 rpm per 
rev. In first gear it doesn't rev quite so fast, untill the tyres start
spinning.

If we assume that we get a reasonable idea of the engine speed at every pules
from the CPS, then we would need only 6 or so to get a resolution of 10 rpm
under a worst case condition.

The other issue is how good the clock is on the computer. At 2000 rpm you're
doing 30 odd revolutions per second. Thats 30ms per rev. One crank degree is
about 90us. If you've got 12 cps pulses, one every 30 degrees, then the
computer has to accuratly measure at most 2.7ms before firing the spark.

Looking at an 8000rpm limit, one crank degree is 20us.

If you want a 1/2 degree timing resolution then the clock has to stable enough
to measure 2700us with 30us accuracy and also be able to measure 600us with
10us accuracy.

The more CPS pulses you have, the shorter the maximum measured time will be.

How much of an issue this is, I do not know. I do know that you do use
different clocks to measure different lengths of time. A clock that has a
resolution of micro seconds is not often very accurate over several seconds.

--
Tom Parker - tparker at nznet.gen.nz
           - http://www.geocities.com/MotorCity/Track/8381/