Need help identifying a tone control topology based off schematic

[bushidov]

Hi all,

I have traced a pedals' tone control and then created a PCB for it. I know that it actually works, but I am not totally familiar with the topology of the tone controls in the schematic I traced out.

Coming from the input of the signal, where VR4 and VR5 are at, I recognize this as a standard active Baxandall tone control.

However, after that, the VR6 and VR7 pots appear to be doing some other type of tone control that goes into separate, unity gained, inverting op amps (inverted twice, it is back to normal, I suppose). It looks like "half-a-baxandall" but I am just guessing at that one. It doesn't look like a gyrating band EQ.

I do know that it cuts/boosts about 12dB at 1kHz for the VR6 and cuts/boosts about 12dB at 2.8kHz, but I only know this because of the pedal I traced the PCB states that on their page.

So, my question is, what kind of tone topology is this? Any examples on how to calculate frequency boosts/cuts?

[merlinb]

Yes they are Baxandall mid controlls. It's a classic topology that is normally done with just one opamp and three bands (see image below), but they've split them up because they want more than one mid band (which you can't do around one opamp with this topology). You can imagine the 220k resistors in the last two opamps to be acting like dummy treble/bass controls permanently set to the centre position, with only the mid control actually implemented.
http://sound.whsites.net/articles/eq.htm


Normally the boost or cut would be equal to the [resistance from output back to wiper], divided by the [resistance from input to wiper]. But for a mid control this figure is reduced and the maths gets hard. Better to sim. Looks like your circuit is using 22k pots, so the 'ideal' max boost/cut would be [22k+2.2k]/[2.2k] = 11 (20.8dB). But this is reduced by virtue of it being a mid control, so you only get 12dB in reality. A normal graphic EQ gyrator topology would have made a lot more sense.

[bool]

Splitting mid-bands like that is a workaround to avoid tone controls interaction. And the output will get back "in phase".

[Rob Strand]

So, my question is, what kind of tone topology is this? Any examples on how to calculate frequency boosts/cuts?

First notice the caps in each stage are in a ratio of 10:1.  If you change the circuit keep the cap ratios 10:1.

Next, larger capacitors  produce lower center frequencies.

So if you take the 1kHz circuit and reduce both caps by a factor of 2.8 you will end up with 2.8kHz.   When you do that you will notice it doesn't quite match the 2.8kHz bands.  That's not because the method I mentioned is wrong, it's because either the 1kHz or 2.8KHz bands (or both) are a little off.

If you search the archives I've posted quite a few times on this ckt.

[FWIW, the 1kHz band is about 1050Hz.  So that means the other band is higher than 2.8kHz.]

[Rob Strand]

Plug these formulas into a spreadsheet:

[antonis]

"Peaking" EQ with only 1.12 Q..?? 

[bool]

What is it then? A "dipping" eq?

... "hey look mom, here comes the dipper"!

[Rob Strand]

"Peaking" EQ with only 1.12 Q..??

Q's of around 1 sound pretty natural for instruments.  Obviously not the choice for a 1/3 octave equalizer.  Sometimes you might want higher Q' when cutting.   Some bass preamps use lower, say 0.7.

The mids on those 3-band "baxandall" style equalizers are very low, like 0.5 sometimes less!  You have to work really hard to get higher and after all that the controls end-up interacting quite a bit.

You have to be careful about a Q spec at full boost.  If you have a Q of 1.4 on a +/-15dB eq, then you set it to only 12dB the Q drops to about 1.   Many equalizers are proportional Q ie. the Q increases as the amount of boost (or cut) increases.

[antonis]

What is it then? A "dipping" eq?

Using general filter terms, peak (or notch) should correspond to a Q of 10 or higher, shouldn't it..??
(of course, useless for audio purpose due to very narrow bandwidth..)

A filter of Q=1, at 1kHz say, exhibits -3db bandwith from 500Hz to 1,5kHz - should you name it "peaking" ..??

[merlinb]

A filter of Q=1, at 1kHz say, exhibits -3db bandwith from 500Hz to 1,5kHz - should you name it "peaking" ..??

What other options are there? Hump and scoop? Trough and barrow?

[Ben N]

A filter of Q=1, at 1kHz say, exhibits -3db bandwith from 500Hz to 1,5kHz - should you name it "peaking" ..??

What other options are there? Hump and scoop? Trough and barrow?

Manic depression?

[Rob Strand]

A filter of Q=1, at 1kHz say, exhibits -3db bandwith from 500Hz to 1,5kHz - should you name it "peaking" ..??

Actually that's for a *bandpass* filter.   A peaking filter with a boost of A0 at the peak is 1 + (A0-1)*bandpass.  Think of 1dB peaking boost, there's no -3dB point below the peak.  1dB means a gain of 1.12 so A0 = 1.12 and that means you have 1 + 0.12*bandpass.   

For a 1dB peaking boost, when the band-pass hits its -3dB points the output can be though of as 1 + 0.12 * (1/sqrt(2)) or 0.7dB.   This isn't actually 100% correct because the band-pass filter also as a phase shift at the -3dB points so we can just add like I have here.  The point is at the band-pass -3dB points the peaking filter produces an output somewhere between 0dB and 1dB.

If you have 15dB peaking boost then the band-pass -3dB points are going to be quite close to the peaking filter's 12dB points either side of the 15db peak.