low pass filter , high pass filter , rc filter DIY. > 36db per octave

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
That's a terrible way of going about it. You should look into active filters (using op-amps) and cascading them instead of going passive all the way. I don't know of any active filter design that gives >36dB/oct slopes but cascading a couple of 18dB/oct filters will give you the same result.

Good luck,

Mike
 
6dB/oct is easy. 36dB is six inter-connected 6dB/oct stages which makes it 36 times harder.

Why do you NEED a 36dB/oct filter? There are very few musical uses for such a sharp knife.

When I really have trash that close to treasure, these days I use a digital filter.

The Falstad L-C-L-C.... filter appears to be 84dB/oct and 50 ohms impedance. Not easy to drive with common audio amps. If you scale to 10K impedance you need 2.9H and 5.8H coils, which are typically huge.

The cascade R-C-R-C... is easily built but has a VERY soft corner. It runs 1dB over the first octave, 2dB over the second octave, etc etc. The deep bass is not cut much while the midbass is weak.

On all simple linear filters (active or passive) the phase response is linked to the sharpness of the amplitude response. If you want a steep filter, you must accept huge phase shift.

That plot looks zig-zag because you get 90 deg shift for every reactance, you have 14 reactances, you have over 1,000 deg phase shift, the plot only covers a 360 degree (circle) range, so it auto-ranges.

(as Mike just pointed out) An 18dB/oct "Active Filter" is one opamp and really sharp enough for most musical purposes. Use a filter-table (Walt Jung, Active Filter Cookbook) to get good 12 and 18dB alignments, or cascade two stages for optimum 24 or 36dB alignment.
 
That's not a singularity in the phase diagram. Think of phase shift as time delay. at 360 degrees phase shift signal is delayed one period. You will have total phase shift of 540 degrees, Now set the scale on the phase axis to 720 degrees. What looks like a singularity magcally disappears-
ry to plot group delay instead together with phase and it will become clear.
 
Gabriel Sousa said:
but using active filter i have the same problem in phase.
In the analogue domain phase shift is directly related to the steepness of the slope in amplitude frequency response. The steeper the slope, the more phase shift. It's those pesky laws of physics, which unfortunately can't be broken. The only alternative is the linear phase (FIR) filter, which only exists in the digital domain (OK, it might be possible to implement in analogue with a bucket brigade delay, but performance would be terrible). That will also produce a phase shift, but it will be proportional to frequency instead of slope.

Why is phase shift a problem? Is it for a crossover?
 
johnR said:
Gabriel Sousa said:
but using active filter i have the same problem in phase.
In the analogue domain phase shift is directly related to the steepness of the slope in amplitude frequency response. The steeper the slope, the more phase shift. It's those pesky laws of physics, which unfortunately can't be broken. The only alternative is the linear phase (FIR) filter, which only exists in the digital domain (OK, it might be possible to implement in analogue with a bucket brigade delay, but performance would be terrible). That will also produce a phase shift, but it will be proportional to frequency instead of slope.

Why is phase shift a problem? Is it for a crossover?

no is for a lofi efect

a pedal efect to use on vocals.

is a problem not to have a linear phase ?
 

Latest posts

Back
Top