[sdiy] SEM filter questions

[Harry Bissell]

A lot of the character of a filter is 'how it feels'... What happens as you increase the
resonance pot ? Is it a smooth 'musical' transition (if you like that sort of thing) or is it
a abrupt screaming transition to rail-to-rail oscillation (if you like that sort of thing).

I happen to LIKE the 12db state variable filter... mostly because as you increase resonance
(in the most common topologies) the pasband does not drop (like Moog ladder designs). In the ladder
filters I have designed, I incorporate a circuit to stop the passband drop from being so abrupt...so the 'hole' between
hearing (primarily) the passband and hearing the resonant peak is reduced.

Others probably like the circuit 'as-is'...

Its hard to say if the diodes in the resonant feedback of the SEM are a positive or negative, it depends mostly on taste.

(BTW I use zeners as a resonance limiter in my "Muffy" guitar processor, so that the filter won't overload at really high
Q values as it sweeps through the fundamental.  It makes the filter much more 'thin' sounding... but the alternative is to have
horrible distortion as its hard to simultaneously reduce the volume of six guitar strings. Hmmm... maybe I should have used the
ladder instead :^)

H^) harry 


----- Original Message -----
From: David G Dixon <dixon at mail.ubc.ca>
To: 'Neil Johnson' <neil.johnson97 at ntlworld.com>
Cc: synth-diy at dropmix.xs4all.nl
Sent: Thu, 19 Apr 2012 18:27:09 -0400 (EDT)
Subject: Re: [sdiy] SEM filter questions

> > A buddy has asked me to build an Oberheim SEM-inspired filter.
> 
> I guess if its just for personal use this would be OK, but 
> for commercial use I'd be a little careful now that Tom is back in the
> game:
> 
> http://www.tomoberheim.com/

I've no intention of replicating this as a product.  In any case, there is
already an SEM clone out in eurorack (the Bubblesound SEM/20).

> > Also, is there something special about this particular SVF 
> which makes 
> > it sound unique?
> 
> Possibly.  Structurally its a standard 2-pole state variable. 
>  Other than the CA3080 OTAs, you've got 741s as output 
> buffers, and an LM301
> with 5pF compensation capacitor as the input mixer.   All the BJT
> goodness is likely to add *some* colouration to the sound.  
> How much though?

It all probably adds more noise than colour.

> > I notice that the resonance circuit has a kind of 
> interesting looking 
> > resistor network and a couple of anti-parallel diodes.
> 
> That's nothing unusual.  The ASM-1 VCF has the same (might 
> even be inspired by it).  I have a similar scheme in my 
> 3-pole filter.  Many other SVF schematics I've seen have 
> something similar as well.  Its a good place to put 
> amplitude-limiting diodes, well away from the high-impedance 
> integrator nodes.

Hmm.  I put zeners in the feedback loop of the first (BP) integrator in my
2164 SVF schematic, along with a 2164-based VCQ circuit and small amount of
fixed feedback for oscillation stability, and the simulations worked very
well.  I'm curious to see how this method compares.

> > Do the JFET buffers impart any special qualities to the filter?
> 
> Unlikely, but possible.  What does your Multisim say?  I'd 
> think the bipolar op-amps would have more of an influence.

Yeah, I haven't done a simulation yet.  Tonight, perhaps.

_______________________________________________
Synth-diy mailing list
Synth-diy at dropmix.xs4all.nl
http://dropmix.xs4all.nl/mailman/listinfo/synth-diy

-- 
Harry Bissell & Nora Abdullah 4eva