Question for Float Variable

kkatesuda · Joined: 09 Sep 2006 Posts: 8

I try to use Float to define my Variable

Ttelmah · Guest

Look at the file 'float.h'. This is a standard C include file (standard in any C offering a float library), defining the data sizes.
You will not get the sort of resolution you are talking about. The float, has a 'core' resolution 1 in (2^24)-1. It 'one normalises' the mantissa, and then stores 23bits, leaving out the leading '1'. It therefore has a basic accuracy of just over 6 digits.

Best Wishes

kkatesuda · Joined: 09 Sep 2006 Posts: 8

Please give me more. How i use float.h just include? plase give me some example.

Thank you very much.
Katesuda

SherpaDoug · Joined: 07 Sep 2003 Posts: 1640 Location: Cape Cod Mass USA

Calculating to one part in ten million is going to be tough. CCS 24 bit floats won't work. Neither will a scaled double integer. You are going to have to roll your own data type and math functions. I use 24 bit integers to measure ship hulls but all I do in the PIC is add and subtract. Anything more complex gets handed off to a PC. 0.15mm x 2^24 = 2.5km, good enough to measure the thickness of the steel or the length of anything afloat.
What is the application that requires this extreem precision? Maybe we can find a work-around for you.
_________________
The search for better is endless. Instead simply find very good and get the job done.

Douglas Kennedy · Joined: 07 Sep 2003 Posts: 755 Location: Florida

I posted some CORDIC routines in the code library it may give you the arctan precision you want without floating point. It only uses addition and subtraction and shifting so it could be modified to doubleprecision say 64 bits.
There are sqrt (x^2+y^2) polar to cartesion conversions etc. there as well.