Post by thread:[PIC] tan function on a 8 bit PIC

Alan, About a year ago I wrote a fixed point transcendental library. It is not a problem to calculate tan on a micro including the 18F but I would avoid using it if it is possible to use a separate algorithm in applications. w.. alan smith wrote: > I did some searching, and saw some routines on arctan, but wondering is doing a tan calculation out the the reach of a 18F processor. What I am eventually implementing is audio eq calculations.

On Wed, 27 Aug 2008, alan smith wrote: > I did some searching, and saw some routines on arctan, but wondering is doing a tan calculation out the the reach of a 18F processor. What I am eventually implementing is audio eq calculations. Hi-Tech PICC-18 using 24-bit floats: #include <math.h> void main(void) { float x, y; x = tan(y); } Memory Summary: Program space used 947h ( 2375) of 20000h bytes ( 1.8%) Data space used 43h ( 67) of F00h bytes ( 1.7%) -- John W. Temples, III

alan smith wrote: > I did some searching, and saw some routines on arctan, but wondering is doing a tan calculation out the the reach of a 18F processor. What I am eventually implementing is audio eq calculations. Could you look into C18 SRC folder? -- Ciao, Dario

On Wed, 27 Aug 2008, alan smith wrote: > I did some searching, and saw some routines on arctan, but wondering is > doing a tan calculation out the the reach of a 18F processor. What I am > eventually implementing is audio eq calculations. Yes you can calculate tan on an 18F. I guess the real question is how fast do you need to calculate it. I've written a floating point tan library function for the XCSB compiler and that generates code for the dumber 16F. Regards Sergio Masci

You can approximate via interpolation and a discrete TAN lookup table rather quickly. sergio masci wrote: {Quote hidden}> > On Wed, 27 Aug 2008, alan smith wrote: > >> I did some searching, and saw some routines on arctan, but wondering is >> doing a tan calculation out the the reach of a 18F processor. What I am >> eventually implementing is audio eq calculations. > > Yes you can calculate tan on an 18F. I guess the real question is how fast > do you need to calculate it. > > I've written a floating point tan library function for the XCSB compiler > and that generates code for the dumber 16F. > > Regards > Sergio Masci

alan smith wrote: > I did some searching, and saw some routines on arctan, but wondering is doing a tan calculation out the the reach of a 18F processor. It will certainly be able to do it. Whether it will be able to do it at the speed and accuracy that your application requires is another matter.

> I did some searching, and saw some routines on arctan, but wondering is > doing a tan calculation out the the reach of a 18F processor. What I am > eventually implementing is audio eq calculations. Hi Alan, Why do you specifically need to calculate tan()? You might get away with the approximation: tan(x) ~ x for small x But if you need to calculate all the way out to tan(pi/2), you're in for a surprise! Scott

Yep...come to the conclusion that the lil ol 18F won't cut it. True I can use a look up table, but I found an easier method....find a part that has things built in, in as much as the function I need for the audio eq. Wolfson has a part, and yes it limits the choice of centre frequencies I can choose, for this particular app it will fit just fine. Thanks for the replies --- On Wed, 8/27/08, Scott Dattalo <spam_OUTscottTakeThisOuTdattalo.com> wrote: {Quote hidden}> From: Scott Dattalo <.....scottKILLspam@spam@dattalo.com> > Subject: Re: [PIC] tan function on a 8 bit PIC > To: "Microcontroller discussion list - Public." <piclistKILLspammit.edu> > Date: Wednesday, August 27, 2008, 11:11 AM > > I did some searching, and saw some routines on arctan, > but wondering is > > doing a tan calculation out the the reach of a 18F > processor. What I am > > eventually implementing is audio eq calculations. > > Hi Alan, > > Why do you specifically need to calculate tan()? > > You might get away with the approximation: > > tan(x) ~ x for small x > > But if you need to calculate all the way out to tan(pi/2), > you're in for a > surprise! > > Scott > --