Datalogger's

[Matthew B. Watts]

-----Original Message-----
From: luna3!angel <angel at luna3.lc.lucent.com>
To: diy_efi at efi332.eng.ohio-state.edu <diy_efi at efi332.eng.ohio-state.edu>
Date: Monday, March 02, 1998 9:18 AM
Subject: Re: Datalogger's


>RPM is typically a pulse, use a counter to measure it. Any counter
>will do since its a fairly low rate.
>
>Angel


Almost.  Your counter clock (sampling speed) has to run fast
enough to get decent resolution at high RPM and the counter
needs to be wide enough so that it doesn't overrun at idle or
startup.

For example, a 16-bit counter running at 1MHz will give you 4614
counts at 13000 RPM but will overrun (200000 counts) at a startup
RPM of 300.  In this case, we need to slow the clock down by
a third to eliminate overrun (only have 65536 counts possible),
which reduces our high RPM reading to 1538 counts.  This still
gives us very good resolution of:

1 count = about 8.4 RPM on the high end
1 count = about 0.0045 RPM on the low end

(which should be plenty good even for an Indy Car)

Keep in mind that resolution gets better at lower RPM,
so decide what MAX RPM is and go from there.

It's useful to think about the engine in reverse.  Instead of
revolutions per minute, think about counts (or unit of time)
per revolution.  This method makes it much easier to look at
fuel injection systems because one knows that the injection
cycle must be some multiple or fraction of the total time
per revolution.  It's also easier to figure out where the
engine is in its rotation.  From the example above, if we're
running at 13000 RPM and have seen 384 counts (1/4 of
a total revolution, 1538 counts), then we are at 90 degrees
past the timing mark.

Hope this helps.

Matt Watts
University of Utah
Formula SAE Racing
_________________