Has anyone had any success with the bandpass filter example in section 12.9? I implemented the coefficients as shown and when trying to filter my data the output diverges within 10-15 data points. I mean it oscillates at every step between ever-increasing positive and negative values until the magnitude gets too large.

I can see how equation 12.8.11 comes about, but the jump to 12.9.12 is beyond my abilities.

Is there anything about the coefficient's values in a IIR filter that can be used as a sanity check? :eek:



Thanks
Bill

I had the signs reversed in the d1 and d2 coefficients. I was subtracting the d1 and d2 terms when I should have been adding them. The sign reversal was already taken care of in the sign of the coefficients, and I didn't know that.