Russell wrote...

>His design is
>a single stage 3 pole circuit (1 more RC input pair) which is the most you
>can reasonably stretch out of 1 transistor.

Yes, mainly because that emitter follower has somewhat less than
unity gain, as well as a non-zero output impedance.  Using an
opamp, or an NPN/PNP boosted follower, getting five poles is
pretty easy provided you can accept a slightly "non-classical"
response.

>I haven't checked
>Dave's component values but I assume this is a 3 pole Butterworth which has
>a reasonably smooth response curve (could be several other things instead).

I synthesized that filter the easy way: I let Microchip's
FilterLab program design it for me.  The response is more or less
a Butterworth.

>My main reason for posting (apart from commenting on Dave's commendable
>promptness :-) ) is to note that you can move the frequency within reason by
>scaling components. You don't touch C1, R1, R2, R6.
>
>By increasing C2, C3, C4 in the same proportion you decrease the cutoff
>frequency proportionately.
>By increasing R3, R4, R5 you do the same thing.
>Alter them the other way and the frequency increases.
>Do both together for greater effect. egg if increasing C's by 10 times and
>R's by 2 x the net frequency would drop by 2 x 10 = 20 times.

One thing to bear in mind when changing the resistors in that
filter circuit is their effect on the transistor bias; R1 or R2
might require tweaking to bring the output quiescent voltage back
to Vcc/2 (if that's important; it might not be).

DD

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu