Help with passive EQ design

[Vyvyan]

I'm in the process of designing a passive eq with some inductors I've been collecting. What's the best online tool for calculating capacitor values with a given inductor and frequency values?

Also I'm lost on how to choose resistor values in RLC networks.

I have a large amount of parts form old consoles solid state and tube. Im planning on using 50k pots with center detents, but I have plenty of other pot values at my disposal.

Planning on a high shelf, hi-mid, low mid, and low shelf each with a few selectable frequencies. Using iron age audioworks m2 melcor style opamp as makeup balancing between the -/+ inputs like an API 553, but I'm open to suggestions for alternative makeup configurationd, also making another with a G-pultec style makeup stage with 600:600 input transformers on the input to the passive filter section.

Any help with resistor values and/or a good online tool for calculating capacitor values give inductor and frequency values would be greatly appreciated!

 

[gyraf]

Impedance Nomograph - print this out, use a handful of color pencils to estimate every aspect you'll need: http://www.testecvw.com/carl/images/ImpedanceNomograph.pdf

LC: Resonant Frequency Calculator
RC: Guitar Pedals: R-C Filter Calculator

Pro tip: Even if you measured and quantified the inductance of the iron you have around, this will not necessarily reflect with any precision on how it will behave in a real-world filter - both because inductance is often heavily influenced by working/measuring frequency, and because loading/saturation characteristics are quite hard to quantify, and even harder to really predict..

/Jakob E.

 

[ruffrecords]

Pro tip: Even if you measured and quantified the inductance of the iron you have around, this will not necessarily reflect with any precision on how it will behave in a real-world filter - both because inductance is often heavily influenced by working/measuring frequency, and because loading/saturation characteristics are quite hard to quantify, and even harder to really predict..

/Jakob E.

Which is why most pros do not use iron. Instead they use ferrite which is much more predictable.

And the R in RLC circuits defines the Q (sharpness) of the resonance)

Cheers

Ian

 

[5v333]

A resonant LC circuit has a resonant freq of
f_0 = 1/ [ 2*pi *sqrt(LC) ]

Q-value depends on the R, L and C

 

[ruffrecords]

A resonant LC circuit has a resonant freq of
f_0 = 1/ [ 2*pi *sqrt(LC) ]

Q-value depends on the R, L and C

Strangely, Q depends only on f, R and L (but of course f depends on C so you are correct). But from the point of view of designing EQ you would start with the required resonant frequency f and sharpness/bandwidth Q and use those to work out L given Q = 2*PI*f*L/R. Once L is calculated you can work out the required C for that frequency. R is determined by the circuit topology (mainly) and by the inductor's dc resistance (slightly).

Cheers

Ian

 

[Vyvyan]

Which is why most pros do not use iron. Instead they use ferrite which is much more predictable.

And the R in RLC circuits defines the Q (sharpness) of the resonance)

Cheers

Ian

I have been salvaging inductors from old Yamaha boards, pm-1000/430/700 and old radio shack graphic eq's (realistic 31-1988/31-1987/2018) and ADC sound shapers (basically rebranded versions of the realistics) so I have a handful of inductors of these values:
8.2H
5.6H
3.2H
2.5H
1.4H
1.2H
1H
.6H
.68H
.37H
.4H
.16H
.1H
.18H
.03H
.06H
.041H
.047H
.021H
.012H
.01H

Also have an inductor from an old wah pedal, 10H I believe.

I understand that the resistor in each frequency network determines the Q but still have no idea how to go about chosing that value. It has an effect on the over impedance of the circuit right? So does that give me a kind of limited range to choose from given over effect on the impedance?

I would love any help pointing me to specific circuits to study and possibly base a design on given the range of inductor values I already have. I'm still a novice with all this stuff, but wanting to learn as much as possible, so I'm not necessarily just looking for anyone to do the work for me or anything. With each project I do there are so many "ah-ha!" Moments where things click and ideas and theory I've read countless times finally falls into place and makes sense lol, but impedance is still quite a mystery to me. And I've read up on many EQ designs and been through countless forums on Passive EQs but unless I'm missing something the one thing none of them ever seem to address is resistor values. They might say something like "R determines the Q" but never seem to expand on how exactly or how to choose and calculate values.

So any help or leads in the right direction would be lovely

 

[NOON]

Try putting a potentiometer in the R position then listening to the effect as you vary it, then look at what it's doing to the bandwidth using a software tool like REQW. If you're using salvaged parts then there's no accurate way to calculate R for a desired result without going down some serious rabbit holes, so select to taste by experimentation.

 

[ruffrecords]

Check out my old universal EQ project at:

Custom Tube Consoles - DIY

Scroll down and click on the universalEQ folder.

Cheers

Ian

 

[Script]

Visual Analyzer, a free software for Windows, has a bidirectional sweep function allowing you to 'see' what the filter does. Just for visual feedback.

 

[5v333]

This is kind of how far i have got on the subject.
For just a RLC circuit the damping factor is 4L/(R^2*C)
and i think Q = 1/(2*damping factor) in this case.
This can be derived when solving a diff equation of the circuit.
But with extra resistors in parallell somewhere, the diff equation
becomes N diff equations of different order(dB/oct) where N is
the amount of loops the circuit contains.
The solution gives us the transient response of the circuit and is
the sum of N solutions.
So for many solutions i dont know how to work out the total Q-value.
Does a 6db/oct filter even have Q-value?
The damping factor tells us if the solution is complex or real.

 

[5v333]

I think i need to come back later on and correct my self about what
damping factor or damping ratio is.
But most of the description should be appropriate when defining
a RLC system.
wiki says Q = 1/(2*damping ratio)

 

[5v333]

yes i was wrong about the damping ratio.
for a series RLC circuit it should be (R/2)*sqrt(C/L).
does not really apply to parallell configurations etc...

 

[ruffrecords]

yes i was wrong about the damping ratio.
for a series RLC circuit it should be (R/2)*sqrt(C/L).
does not really apply to parallell configurations etc...

You will find that most passive EQs use series resonant circuits rather than parallel.

Cheers

ian

 

[5v333]

Yes but you need extra resistors in parallel etc in order to make shelfs and bells.
And then the damping ratio becoms something else.

 

[ruffrecords]

Yes but you need extra resistors in parallel etc in order to make shelfs and bells.
And then the damping ratio becoms something else.

This the case in all LC based passive EQs whether you use series or parallel networks. Calculation of what I call the characteristic resistance of the entire circuit (which determines Q) is very difficult which is why I use LTspice to do it.

Cheers

Ian

 

[5v333]

cool

 

[gyraf]

..actually, the nomograph paper above will also output/show the characteristic circuit resistance in a quite intuitive way, letting you estimate Q..