> > The tradeoff between random noise attenuation and the step response of =
a
> > single pole filter is fixed. =A0No free lunch to be had. =A0However, yo=
u can
> > significantly decrease the step response at the same random noise
> > attenuation by cascading multiple filters. =A0This may sound like a fre=
e
> > lunch, but it's not since the computation requirements have gone up.
> > Fortunately single pole low pass filters are relatively easy to compute=
, so
> > using multiple poles is often a good option.

Lunch is essentially free for incremental guests if there is so much
food available that it would otherwise be thrown away and they wash
their own dishes and don't break any. ie if the processor bandwidth
and any extra memory required would otherwise not be used. But, I know
you know that :-).

> If you have enough CPU available, you can use a proper multi-pole
> butterworth filter, this web page is good for a simple design, as it
> gives C code for the filter:
> =A0http://www-users.cs.york.ac.uk/~fisher/mkfilter/trad.html

I've had excellent results in real world use from the usually ignored
Bessel filter. This is in analog implementations but the same applies
in digital form.

Dusting off dimming memory - may get something a bit wrong :-) -
Recollection is that two pole designs are identical for all 3 classic
filters (Butterworth, Chebychev, Bessel) but they diverge radically
from there. Also, if memory serves, both poles are identical for 2
pole which makes life even easier. (Doesn't sound right but that's
what memory offers. For something a bit better than simple RC like 1
pole response (which we've been discussing here) where the long
exponential delays are undesirable, 3 or 4 poles of Bessel is usually
excellent. The constant phase feature of this filter means that there
are less surprised than with the sharper Chebychev or the smoother
Butterworth.

For analog low pass implementations you can use emitter followers as
active elements so you can get 4 poles (5 if keen) with two
transistors and a small handful of passives. (You can realistically
get 3 poles in the first stage and two per stage thereafter.)

All that is less relevant to digital implementations - digital emitter
followers don't seem to be in much demand.


       Russell

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.