synthacon weirdy filter mutates!

[danial stocks]

I just made a mutated version of the synthacn weirdy filter in an envelope 
filter.. this is the mods I did.. some of these may appear odd, but was the 
best way to get it to fit into the 82x53x30mm box I wanted to use without 
having to order anything really obscure.. I cant scan etc, so you need to 
look at the diagram Jim Patchell had while reading this description..

first, it has to run on a single 9V supp.. so synthetic gnd is used..
-the transistors I used was 2SC1815Y, roughly gain matched.. No real reason 
for these TR's , I just like jap tr's ..
-caps were changed to 10nf in the filter network.. 18nf is even nicer 
sounding, but 10nf was selected as a good compromise for range as the filter 
does not sweep a really broad range with so low supply vtg
-the output trans stages were ditched. the 100n cap from the diode chain 
feeds a non inv opamp stage with AV=+2.. level pot is drawn from this opamp. 
also res pot is from here.. the res gets really out of hand high amplitude 
oscillations, so the o/p of the op amp feeds thru 2k2 res, into 100K pot. 
this pot has 1n4148 diodes for clippimg across it, and a 1 nf cap . the o/p 
from the pot goes to another AV=+2 amp which drives the feedback point on 
the filter network.
-input network on Q1 [CV in] is now 8K2 res i/p, 2k2 res to gnd.. no attempt 
was made to scale this filter to anything in particular, these vals sounded 
best in the env filt system and so were used.
CV in range is nominally gnd - v+, but about - 0.5V is needed to make the 
filter fully cutoff. I think this may well be the case if the filter is 
built as per the normal design too.. no amount of fiddling altered the fact 
that filter was not cutoff at gnd.
-from o/p cap on diode, a 1m5 res is conn'd to gnd on op amp side.. [non inv 
amp likes to have a gnd ref if no DC bias on it.. ] also this res could be 
craftily used to form a hpf with the dc blocking cap.. a rolloff of gain @ 
low freq's, allowing res settings such that self oscillations cease at lower 
freq's.. so filter could be made to osc & stopped by sweeping the freq..
-res to V- from the emitter junction of q1&q2 is now 2k2..
-Audio input res has been reduced to 2k2.. this filter can be run at fairly 
high amplitude with good results, the 47k res spec'd in the original seemed 
highly unneccessary [maybe they ran it from a 10Vp-p VCo?] infact vals of 
res this high made it seem a bit noisy.. OTOH the val of res here too low 
adversely affects the bandpass resonance performance. It does not however 
affect HP or LP res.

Results.. sounds nice..  nice diodey sound, but best to run it from much 
higher voltage..
Cheers,
Dan
_____________________________________________________________________________________
Get more from the Web.  FREE MSN Explorer download : http://explorer.msn.com