deep filters with least parts

[Brian Marshall]

Ok, say i want a filter with a 24db roll off per octave from the cut off frequency,  I know i can cascade an rc network, but there has to be a way with less parts right?

[Mike Burgundy]

Not that *I* readily know off. Maybe with active parts, but that'll probably involve...more parts. Caps and resistors don't cost much, though.
If you want *good* 24dB/oct filtering, you might want tot consider a linkwitz-Riley circuit (google)

[Brian Marshall]

Not that *I* readily know off. Maybe with active parts, but that'll probably involve...more parts. Caps and resistors don't cost much, though.
If you want *good* 24dB/oct filtering, you might want tot consider a linkwitz-Riley circuit (google)

yeah, i'm toying with an idea that may or may not be all that great... just s direct box/cab sim.  It would run unpowered, so actives are out.  sorry i didnt speciy that.  I was trying to keep it super small, and cheap.  I guess i need to stop buying 5% metal film caps.  that would make it cheaper.

[Boofhead]

What is the filter input connecting to?
What is the filter output connecting to?

I should imagine a 24dB/oct RC filter isn't going to work too well, the roll-off will be very sloppy.  With no active devices for buffering the impedances are going to be quite high and you could end-up with loading and noise issues.

You could try an filter which uses inductors capacitors and resistors, but you need to assess loading effects.   Depending on the frequencies involves the inductors could be large and expensive.

[Ansil]

do you need multiple freq ranges...  only one freq.  which freq.

[Brian Marshall]

basically what i want to do is make something like the H&K red box.  I already have one, but I'm not too happy with it.  I just want to plug it in to the preamp out on my amp, and straight to my mixer. balanced or unbalanced is fine.  I suppose i am going to have to use an inductor/transformer on the input anyways.  I believe that is what the red box does to keep the circuit isolated.  I could be wrong.

basically what i want is this... -6db high pass roll off at about 80hz, -3db high pass (or maybe a notch)  at somewhere between 400, and 750hz, and a -24db roll off at 3.5 to 4khz.  something like that.  no knobs or anything.  I know how to do this with RC networks, but seems like a lot parts for something so simple, and i am worried about losing a lot of signal

I really want it to not use any batteries.

[Ansil]

basically what i want to do is make something like the H&K red box.  I already have one, but I'm not too happy with it.  I just want to plug it in to the preamp out on my amp, and straight to my mixer. balanced or unbalanced is fine.  I suppose i am going to have to use an inductor/transformer on the input anyways.  I believe that is what the red box does to keep the circuit isolated.  I could be wrong.

basically what i want is this... -6db high pass roll off at about 80hz, -3db high pass (or maybe a notch)  at somewhere between 400, and 750hz, and a -24db roll off at 3.5 to 4khz.  something like that.  no knobs or anything.  I know how to do this with RC networks, but seems like a lot parts for something so simple, and i am worried about losing a lot of signal

I really want it to not use any batteries.

ahh i got you..  well shoot the preamp is from an amplifier right???

len glasgo i htink that is his name.  anywya he had this idea for stealing some signal off of your amplifer speaker and making it line level. i have his pic somewhere on the HD. anyway you could do put a small load resistor  from the amp output and then use a rc network to do that. and it wouldn't eat as much signal that way.

[Brian Marshall]

my preamp out is buffered, and loses no signal when i plug it in.  I'm not using the fx send, i could, but dont see any advantage, it would only make it harder, because i would need a return, and would have to wory about the tone degrating.

I'm starting to think this may be easier using opamps, anyone have any idea how long a batery would last?

[Ansil]

ahh..  ok  a buffered preamp would be fine then.  but i wasn't talking about usin the fx loop i think you misunderstood me there.  anyway. you could steal the buffered preamp signal to acomplish this. quite easily.


personaly i would use opamps.  a nice quad opamp. would work nice.  or if you wanted to go with a little more natural sounding. use a 4049 since it overloads gracefully. anyway it also has six inverters.

[Boofhead]

Limiting the filter design to R's and C's will give something which is  just do able but the design will be very inflexible in terms of tone  - basically if it doesn't sound right that's it.

If you want 6dB/oct for the high -pass you can do that with an RC but the rest will require LC filter sections.  You would connect a number of filter sections in cascade to achieve the final result.

Designing such a filter is a difficult task, you would be best finding some LC filter design software and playing with that.  The software will design each section but it won't handle interractions and loading of each stage - ie when you connect the stages together things change. Coping with the interractions and loading is a not trivial matter and requires advanced techniques which are beyond most people.  If you wanted to play around with the loading your best bet is to plug the final filter into a circuit simulator and hand-tweak the values. How it sounds hasn't even entered the equation yet - that will require more adjustments, but at least an LC filet will get you there.  Chopping and changing the inductor values can be an expensive interative extercise.

When it boils down to it the filter for such a unit isn't simple.  An active filter design would let each section be designed in isolation because there is no interraction between the stages - they are also less costly to play around with.

[Mark Hammer]

I don't think there is any physical way around having an RC network for each pole, although there are physically larger and smaller ways to do it.  You *could* use op-amps, but you could also use transistors.  Look at the typical anti-aliasing filters on just about any commercial analog delay and you'll see plenty of examples.

Another method is to use one of the various switched cap filters like the MF4, MF6, MF10, made by National Semiconductor.  These are quite flexible, but they also need a HF clock section to drive the switching, and ultimately take up far more space for a 4-pole filter than any of the the 2-pole transistor circuit fragments.  On the other hand, if you want an 8-pole "brickwall" filter, they start to get economical and convenient.

[Paul Perry (Frostwave)]

Mark, are there any aliasing problems with those switched cap filters?
Because, the operating freq is 1/100 of the clock freq, which is bang in the audio range, for a low cutoff freq.

[puretube]

the appl.-notes want to see an RC hi-roll-off filter afterwards.
You won`t have problems with sw-caps if you use `em e.g. as
a 5kHz lo-pass for BBD-projects (500kHz clock), but surely
hear unwanted stuff e.g. when used for "wah"-purposes around 40 to 200Hz (4 to 20kHz clock bleedthrough and aliasing products)....

[Paul Perry (Frostwave)]

thanks puretube, that's what i suspected