Personal web pages of

Tim Stinchcombe

Ladder SVF

This module formed the engineering project of my electronics HNC, which I completed in late 2005, but the idea for it had been in my head for several years. The basic idea was this: the fabled Moog transistor ladder filter is believed to derive much of the character of its sound from the unusual transistor/capacitor combination of each filter stage, and the way these are configured in a 'ladder' structure - thus would it be possible to take two of these stages, and arrange them to form a state-variable filter structure, and if so would it sound 'good'?

Well, the answer is partially 'yes', in that I have built a working filter: as to whether it sounds 'good' is a matter of subjective judgement. From my point of view I think it sounds better than I was expecting, which is satisfying. The filter is not strictly 'state-variable': such filters use integrators, whereas the ladder stages are standard first-order lowpass sections. Thus I had to manipulate the topology slightly to ensure I got what I wanted, and that it would self-resonate. The name I have given it, 'Ladder SVF', is thus something of a misnomer, but it is easier than 'Two ladder filter stages arranged in a near state-variable filter set-up'!

I designed-in voltage control of the resonance, and the main output is a 12dB lowpass; the output equivalent to the bandpass of the SVF is a 6dB lowpass, which I have also brought out (but which probably isn't really worth it); the equivalent to the SVF highpass turns out to be an 'allpass', but isn't accessible without a lot more circuitry, so I didn't bother with it. The input attenuation is scaled such that it can be operated 'normally' or if cranked up, can be overdriven into distortion.

Much of the design of the filter, particularly the feedback circuitry, was done 'just to see if it could be done' (it can), but it means the design is not really that practical: for example there are 9 trimpots necessary to make it work, thus requiring quite a complicated set-up procedure. Consequently I haven't decided yet whether I'm going to publish the schematics or not. However I am working toward a second version, with a more sensible feedback mechanism, so I may publish the details if and when I get that working (and on the infinitesimally miniscule off-chance that anyone is interested in building it under licence, I'm receptive to any reasonable offer, so get in touch!).

Sound samples

Here are a couple of samples of some very simple filter sweeps, to illustrate the basic capability of the filter. In each: the input is the output from an A-110 VCO; it's the 12dB output recorded; the cut-off frequency is modulated by an LFO, and between each sweep I turn the resonance setting about a fifth of a turn (so 6 sweeps, from min to max), and this pattern is repeated for three input level settings (about a third, two-thirds, and max), so the overall volume level increases too, and the increasing distortion can be heard. (Both are mp3 files, approx 1Mbyte big):

Low frequency sawtooth input

Low frequency square wave input

(I'd like to get a more sonically interesting example up too, but my musical skills are not great, and time seems to be getting away from me at the moment!)

Here's a shot of the PCB: it was laid out using Easy PC, and etched by PCB Pool (Ireland). The faceplate was made by Schaeffer AG (Germany).

[Page last updated: 14 Dec 2009]