Hi Michael,

I can point out a couple of thoughts for you, concerning the need to use double precision variables. At the University of Texas at Dallas, a couple of professors in the geology department got a fancy new GPS system several years ago. They celebrated by putting up a brass plaque proudly proclaiming the location of the plaque in degrees/minutes/and seconds, carrying out 6 decimal places for the seconds. It was like a first year engineering student with a new calculator.

Since the U.S. military ended the SA (selective availability) limitation of the GPS system, all users can enjoy the same precision that the military enjoys, that is approximately a 10 meter radius. If you use certain land based beacons you can improve your relative (to the beacons) precision to something on the order of centimeters, but your absolute precision is still only about 10m.

Now for a quick back-of-the-envelope calculation. The meter was once defined as 1/1,000,000 of an arc from the equator to the north pole. Therefore, 90°/1,000,000 m = 0.00009 degrees/meter. multiply by 3600 seconds/degree and you find that there are approximately 0.324 seconds per meter. With a 10 meter precision, it is pointless to talk about anything less that 3.24 seconds. (Carry this exercise further and you will find that the geology department mentioned above had declared the location of a decameter scale plaque to a precision of µms).

I guess this was just a long winded way of saying that GPS units, like calculators, can give you more precision in a result than exists in the measurement. This is the exercise I would do before deciding to use double variables. And as for your principle question, I would be quite surprised to learn that an 8052 couldn't handle what you described. How hard is it to keep up with a serial port?