High/Low Pass Filter Question

[JustinSelf]

I'm starting to get a decent idea of how filters work in a tone stack, and how to calculate the cutoff frequencies.  I've been doing some tweaking with a big muff tone stack and a couple things occurred to me.  So a 6.8n cap and a 33k resistor in a low pass filter will result in a cutoff frequency of about 710Hz.  A 3.3k resistor and a 68n cap will provide the same result.  Or to go even more extreme, say a 220n cap and a 1k, or even a 1uF cap and a 220 ohm resistor, will also give just about the same cutoff frequency.  Could I use any of these, or are there other considerations that prevent going this route?  Does the large change in resistor values affect the current too much or something else entirely?  I'm quite curious.  It seems every time I start to have an "aha" moment with learning about pedal circuits, I end up with more questions than answers, which I love.  Just more things to learn. 

[Elijah-Baley]

Do you know Tone Stack Calculator software? http://www.duncanamps.com/tsc/
It could be useful to you.
Then try some different values and hear with your ears how it sounds.

[phasetrans]

I'm starting to get a decent idea of how filters work in a tone stack, and how to calculate the cutoff frequencies.  I've been doing some tweaking with a big muff tone stack and a couple things occurred to me.  So a 6.8n cap and a 33k resistor in a low pass filter will result in a cutoff frequency of about 710Hz.  A 3.3k resistor and a 68n cap will provide the same result.  Or to go even more extreme, say a 220n cap and a 1k, or even a 1uF cap and a 220 ohm resistor, will also give just about the same cutoff frequency.  Could I use any of these, or are there other considerations that prevent going this route...

Yes, you can use any pairing that meet the formula, in theory. Here are some cliffs notes notes for you about real world parts:

Cost. Large capacitors can be expensive. Silicon is cheap now. So depending on the application, something like a capacitance multiplier https://en.wikipedia.org/wiki/Capacitance_multiplier, or using an opamp in a bias circuit (see opamp U2b / D  in the Engineer's thumb compressor) can be a real cost savings.

Larger value resistors are noisier https://en.wikipedia.org/wiki/Johnson%E2%80%93Nyquist_noise so try to use smaller ones in the signal path.

As the size of the capacitor changes, the type of capacitor can change. And the internal "parasitic" behaviors of different capacitor types are all different. Capacitors have their own resistance and inductance, for instance. Rarely does that matter at low frequencies (like in audio). But it is very important in other circumstances.

Capacitors have more ways they can wear out / drift out of spec than resistors.

Ceramic capacitors now. Some of the more common capacitor dielectrics (the thing that helps store more charge) are ferroelectric. Common ferroelectric dielectrics are X5R and X7R. Ferroelectric materials are also piezo electric, and therefore can be microphonic in some circumstances. Ferroelectric dielectrics also have less capacitance when subject to a DC bias voltage. Depending on the bias voltage, the effective capacitance can reduce by half.

Paraelectric ceramic dielectrics (e.g. C0G) aren't piezoelectric, and don't have the DC bias effect, but are more expensive.

Capacitors can be physically the largest components in a design.

There are certain "magic" values for price point. 0.1uF MLCC are very common as bypass capacitors, so that particular size can be cheaper than similar values.

I'm sure there's plenty more I am forgetting to add.

[antonis]

I'm sure there's plenty more I am forgetting to add.

Loading.. 

Resistors exhibit a "constant" load (despite frequency) where capacitors dont..

[PRR]

What antonis said. You pick R-C filter resistors in context of the impedances of the circuit it is in.

0.01u and 1Meg is a 17Hz high-pass. But if you put it in line with an 8 Ohm speaker it becomes a 2 MEGA Hz high-pass (radio frequencies).

In passive filters the available impedances limit what you can do.

That's why op-amps are terrific. You can see <1k on one side and >1Meg on the other side, giving wide choice of impedance (and thus cap values).

[phasetrans]

I'm sure there's plenty more I am forgetting to add.

Loading.. 

Ugh. So obvious I forgot to include it 

To the original poster, the input and output impedance of the circuit has substantial influence on the filter behavior.

This is especially true in tube circuits that can have 100k+ output impedance of some stages.

[edvard]

I would add to this that the simplest way to figure out values for your filters would be to learn how to use a circuit simulator like LTSpice or Tina-TI to eliminate a lot of guesswork. You'll be able to quickly see the effect of surrounding components on filter curve and efficiency.

[ElectricDruid]

I would add to this that the simplest way to figure out values for your filters would be to learn how to use a circuit simulator like LTSpice or Tina-TI to eliminate a lot of guesswork. You'll be able to quickly see the effect of surrounding components on filter curve and efficiency.

+1 agree. Even something simpler like:

https://www.falstad.com/circuit/

..will do for sorting out basic filter responses. There are some basic RC filters in the "Circuits"->"Passive filters" menu to get you started.

[merlinb]

Another point:
In the big muff tone control you want the pot resistance to be larger than the impedances in either filter in order to keep them 'separate', so you get a decent variation as you turn the pot. But you also don't want to use an enormous pot resistance since it will add noise. With a 100k pot, you therefore want to keep the resistances used in the filter below 50k, ideally, as indeed they are in the big muff. (But if they're too small then they will load down the previous transistor!)

[Joncaster]

Another point:
In the big muff tone control you want the pot resistance to be larger than the impedances in either filter in order to keep them 'separate', so you get a decent variation as you turn the pot. But you also don't want to use an enormous pot resistance since it will add noise. With a 100k pot, you therefore want to keep the resistances used in the filter below 50k, ideally, as indeed they are in the big muff. (But if they're too small then they will load down the previous transistor!)

That BMP Tonestack was many nights of fun on the breadboard. I settled on a 50k pot for mine (wider sweet spot on the knob), 22k both HP/LP values, ended in slightly less scoop at something like 1.2khz. Creamy.
A nice way to do a mid hump, just switch the 4n7 to a 22n, makes the mid hump in the 500-800hz range depending on R. Woolly.

[phasetrans]

I would add to this that the simplest way to figure out values for your filters would be to learn how to use a circuit simulator like LTSpice or Tina-TI to eliminate a lot of guesswork. You'll be able to quickly see the effect of surrounding components on filter curve and efficiency.

To Justin the OP, this is good advice. you can build filters with just a voltage source and resistive load, and change the resistance of both the source and the load to see how the loading effects the filter behavior.

If you are mathematically inclined, it is worth learning about the Laplace / z transform at some point, because it is darn clever. But given that we have modern software to simulate filters, I would suggest starting with the intuition about how the filters behave and visit the math later.