Hello again everyone,
 
After a lot of testing I'm still having the same basic problem:
 
1) I need to implement an IIR filter with the following data:
    processor BF537
    fixed point format 1.31 (my samples are 32 bits wide)
    sampling freq: 2000Hz
    type of filter: low-pass, 2nd order, 40Hz cut freq.
2) I found an application note (EE186) which implement an IIR filter for the 1.31 arithmetics
3) when I compute the coeff using matlab and quantize the filter the amplitude response is far
from the wanted but if I change the quantized parameters to 2.30 instead of 1.31 the response
goes back to normal.
 
so, now I need one of these two solutions:
 
1) A way to fit my filter coeff to the 1.31 arithmetics.. OR
2) An implementation that can accept 32 bit samples and coeff and also that assumes A0 to be
0.5 instead of 1
 
Please, can anyone lend a hand here?
 
Thanks a lot in advance..
 
Juan Carlos Gálvez
       
____________________________________________________________________________________
Building a website is a piece of cake. Yahoo! Small Business gives you all the tools to get
online.
http://smallbusiness.yahoo.com/webhosting