Power Basic has a very powerful function called MTIME
 where it can measure the time it takes to do something
 in the 2 micro second resolution range.

 Back when I got my first copy of PB in about 1998 or so
 it took 210 microseconds to calculate the value of pi
 on the PC computers that were common in those days.


 pi# = ATN(1) * 4

 Now this is done in less than a microsecond on this PC board
 I am going to use for the EFI controller.
 In fact it can do this 100 times in about 51 micro seconds.
 In other word words it only takes about 1/2 microsecond.

 Lets say we wanted to know the exact RPM of the rotary
 running around 7500 RPM or 7 ms or 7,000 micro seconds per rev.
 We start the timer at rotor 1 TDC and stop it when
 TDC rotor 1 happens again. Our resolution on the exact RPM
 is one part in 3500 or plus or minus 2 RPM!!!!!!!!!!!!!!!!!!!!!!!

 This has far reaching implications for PC boards used
 as real time hardware controllers.

 BTW here is a picture of the computer in a $20 BTC key board
 we will use as a hard ware interrupter controller.

 Paul Lamar.


Paul,

I just discovered this message.  You did not tell the source of the
computer
board which is used to measure the short time interval.

Thank you for the additional information.

Gene

It was a VIA PC board but Intel makes something similar based on
the Atom processor for $90 including the processor. Available
from Fry's. There could be a better way of doing
it if you give me more info on what exactly you are trying to do.

We gave up on that project because PB refused to add USB support
to their DOS product and RS232 ports disappeared on mother boards.
The picture attached to that message was a keyboard controller. Not
the board that did the math. Here is the board that did the math.
The newer Intel version is probably faster.

Paul Lamar


Thank you for your answer to my inquiry.

I fly a gyrocoptor.  Rotor RPM is measured with a magnetic sensor counting
the teeth of  a starter gear used to prerotate the rotor.  The display unit
does not repeat the measurement often enough, about three readings per
second.  It would be preferable to udate the reading more often.  I also use
a tiny tach for engine rpm indication which has the same problem.  The
reading is not updated often enough.
My interest is to replace the indicator units for a more often and possibly
a more accurate indication.
Engine rpm is obtained by sensing the firing of one spark plug firing at
7000 rpm max.  The engine is a Rotax 618 (2 cycle, 2 cylinder)
Rotor rpm is typically around 350.  The starter gear is about 100 teeth.
Time between teeth is therefore 1/35000 minutes (about 1700 microseconds per
tooth).  It may be true that the sensor does not see every gear tooth due to
eccentricity of the gear tooth path.  This is a separate problem.
  In summary, I want a system which will measure the time between two
pulses then multiply the time by anumber which converts the time to rpm.  It
would be nice to do this for less than $50.

Thanks for your suggestions.

Gene

Get a a copy of the "CMOS Cook Book".
That has good tach circuits.

Paul Lamar