[sdiy] Clock-controlled filters?

[cheater cheater]

Hi Tom,

On Fri, Oct 19, 2012 at 11:32 AM, Tom Wiltshire <tom at electricdruid.net> wrote:
> What sort of frequency range are we expecting to see for this pulse train input? Audio? LFO range? Higher? Lower? Give us a clue!

the range is audio range, needs to respond to frequency changes in the
LFO range, the pulse train could be a (large) integer multiple of the
desired frequency.

It's supposed to be an imaging (reconstruction) filter for a
variable-rate DAC. This is unrelated to David's recent post but that
post reminded me of an idea I had a few years back.

> You might be able to use a switched capacitor filter. They usually have a clock input to set the cutoff. The clock is at some multiple (x50, x100, etc) above the cutoff frequency.

Sounds good. I think that's more than likely exactly what I want.

> If you wanted to use one to filter a square to get a sine (like Dave Dixon did recently) you could set it up with a high frequency clock running the filter, and feeding a divider that brings the clock down to the filter's cutoff frequency. Then you have a square wave input with a tracking filter of pretty much arbitrary order. 8th order switched cap filters aren't uncommon.

That sounds pretty good actually. I'm guessing those parts don't
change much, and if they do, then they're mostly drop-ins (modulo
pinout), right?

> It is possible to build your own switched cap filters using CMOS switches and caps, though I don't know that I'd bother.
>
> If you've got a square wave which is derived from some other source (like feeding a guitar through a big muff, for example!)

That's an interesting square wave oscillator :)

> then you'd be looking at using a PLL with a divider to track the input and produce the required HF clock. This is going to be much trickier to make track properly, I'd have thought.

Not trying to do that. But indulge me - why would I need a divider?

Cheers,
D.

> On 19 Oct 2012, at 10:02, cheater cheater wrote:
>
>> Hi guys,
>> are there filter designs, not necessarily resonant, which tune to
>> pulse train input?
>> Chips are good too - although being able to make it discrete if/when
>> the part becomes obsolete would be a huge plus.
>>
>> Cheers,
>> D.
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>