frequency step network

[catalin gramada]

Hallo. How can be calculated a RC low frequency step network,please? I.m particulary interested to achieve a high order cut off low frequency pole in interstage aplication.Does it exist a formula to can be calculate for -18 or -24 db/oct cut off point,for instance? In absence of any inductor means need to be cascaded, right ?Can You help me how to design a very efective sub sonic filter using RC elements,please? Thank you.

[Gingertube]

Catalin,
I don't think a step network is what you want.
A step network gives you a step response but from reading above I think you want to attenuate low fequencies forever (as you go lower in frequency).

What you want is a high pass filter.
1 single pole high pass filter consisting of the coupling cap from the drive stage and the grid leak resistor(to 0V) of the following stage gives you:
-3dB point (corner frequency) when the Impedance of the coupling cap( 1 / [ 2 x pi x frequency x C]) is equal to the grid leak resistance.
It gives you a 6dB/octave slope.

If you want a higher slope you need additional poles, cascaded. Make 2 off RC networks set at HALF the cutoff frequency, one feeding the next, which gives you:
Same corner frequency as above but 12dB/octave slope

If cascading 3 off RC networks, then set the cut off frequency for each to 1/3rd the desired corner frequency to get:
Same cut off frequncy as single pole but 18dB/octave slope.

You often see 3 pole high pass filters in amps with tremlo applied in the preamp used to remove the tremlo oscillator frequencies.
Example: See the 4 pole filter in this schematic: https://schematicheaven.net/gibsonamps/ga40.pdf
Each of those 4 poles has a corner frequency of 31.8 Hz so its combined effective corner fequency is 4 x 31.7 = 127 Hz and has a 24dB/octave slope (ignoring the feed to the volume pot) so it will be down 24dB at 63.5 Hz and down 48dB at 32 Hz.

If you want a single pole high pass (6dB / octave slope) to give you -18dB at a particular frequency then you need to set the corner frequency 3 octaves above that particular frequency (that is at 8 times that frequency).

That is why 2 or 3 pole filter might be the go.

Have I interpreted what you want correctly?
Does this help?

Cheers,
Ian

[catalin gramada]

Hello. That is exactely what I need. Very sorry for confused terminology.. Thank You. Catalin

[catalin gramada]

So, two cascaded single pole filters will give the same frequency reponse as a single second order 12db/octave filter, taking in cosideration the same cut off point,please ?

[Gingertube]

So, two cascaded single pole filters will give the same frequency reponse as a single second order 12db/octave filter, taking in cosideration the same cut off point,please ?

YES but each of those single pole high pass filters needs to have a cut off frequency of HALF your desired combined cut off frequency, so that at your desired cut off (or corner frequency if you prefer that term) frequency each of them is contributing half of the attenuation.

I got this WRONG above initially but have edited it since to fix it. - for low pass each pole needs to be set at 1/2 frequency for 2 cascaded poles NOT twice as I originally said in my first response. SORRY.

Cheers,
Ian

[catalin gramada]

Thanks!

[nickb]

One little gotcha to be aware of is that when cascading networks like this, the second one loads the first one and will affect the result. Design the first to have a much lower impedance than the second. For example 20Hz first stage uses 47K & 0.15uf and for the second 470k & 15nF.

A variant would be to separate the sections so e.g. to have one on the way into the grid of and the other on leaving the plate.

[catalin gramada]

Thanks nickb. I used volume pot in this configuration as less intrusive volume control trying to keep 1:10 raport at least for reason You mention above. Unfortunately there are no so many spots to place this networks in ideal condition. a CF or low impedance plate followed by split inverter will be the best. 100n / 220k pot folowed by 10n / 2.2M (cathodyne grid leak) for ca. 15Hz cut off point for instance. a 12au7 should manage pretty well with no signifiant loses,think... I should have think at this aspect before . Thanks for You warn.

[catalin gramada]

[catalin gramada]

From the graph presented above even with 2.nd order seems it will not work as a effective "subsonic" filter. Maybe is better to cut one octave up ? It is a risk to get rubbish bass transients ?

[Enzo]

https://www.sweetwater.com/store/det...FdW6wAod29YClA


Read the specs on this $80 unit. 24db/octave, PLUS an independent subwoofer output with its own freq control.