iir_filter/butterworth prototype in SigProc becomes unstable for high-order filters

[jgaeddert]

Overflow occurs on output of the filter when the order is large and the pass-band width is small. This has been tested with 10-th order Butterworth low-pass filters with bandwidths of about 5% of the sample frequency.

Need to test if this is from

1. Overflow in the filter states
2. Unstable Butterworth filter coefficients (poles lie outside the unit circle)

[jgaeddert]

This is the nature of using floating-point precision in IIR filters and a great example of why this is bad practice.