Best 3-Band Active Tone Control?

[Joe Hart]

Thanks everyone!!!

I'm off to Toronto armed with much to ponder on the drive up and back.

Thanks again!!!!!!!!!!!!!!
-Joe Hart

[stm]

Wow, this thread is getting interesting.  I see a good interchange of ideas flowing. These are the kind of threads where I learn...

-----
RDV:
-----
Your tone stack option is a valid alternative, however controls do interact and overall maximum cut/boost you can achieve is not as high/extreme as you can with active, separate band, Eqs.  A good thing though is you get the "voicing" from the particular tone stack you choose. *** Finally, the level boost section at the beginning is potentially risky, due to the fact that on single 9V power you can get clipping before reaching the tonestack.  Gain recovery stage should go AFTER to avoid this.  Then, a buffer stage would be needed BEFORE. ***  (the statements between asterisks are  now invalid since RDV changed the schematic according to these suggestions)

-----
Ben:
-----
It is true that a two-band baxandall allows you to cut/boost mids by raising or lowering both controls simultaneously. The catch is that your peak or valley is very wide in comparison to a dedicated notch/peak equalizer, so frequency control is rather diffuse, meaning that you raise the mids and some of the lows and highs as well.  Also, you don't have the chance of boosting Bass and Treble while at the same time cutting Mids, which produces a more severe effect in 3-band Eqs.  The other thing I dislike from baxandall design is that when you boost you get huge amounts of gain below 80 Hz and above 10 kHz (depending on Bass or Treble, respectively), where you don't want it or need it.  This increases potential for clipping and adding additional noise.

----
Hal:
----
Yes, using just the second stage gives you mid cut/boost.  Scale capacitors accordingly to adjust frequency. Keep in mind that values shown are for 550Hz.  Divide caps to increse frequency, and multiply them to lower it.

[stm]

OK, now I'm posting a circuit I did build, and that is very good and controllable.  No savings here on opamps, this is the no-compromise version, meaning that I am not trading functionality or performance for simplicity. Anyway, it is easier to understand and tweak in comparison to the gyrator based eqs., and eventually could be extended to more stages.



Be advised that a buffer stage is needed BEFORE in case you are going to feed it directly with your guitar, and eventually a gain control stage AFTER to match your setup level.

EDIT: Note the 10:1 ratio in the capacitors of each stage.  I optimized the resistor values (using MicroCap 7 optimizer) and the cap ratio to maximize the Q or peakedness of the bands, so this Eq has a much more definite mid band in comparison to a 2-band baxandall tonestack.

[Ben N]

Thanks for that great explanation, and for the cool 3-band eq (all bandpass--interesting).  Going back to your initial design, would that work something like the old Ampegs with Bax + inductor-based mid control?
Ben

[stm]

Ben, I'm not familiar with the Ampeg schems, so I cannot tell.

I have yet an even better and more flexible three-band eq in which I'm working now.  It is based on the concept that with three bands you actually have two degrees of freedom only, due to the fact there are redundant combinations which differ in volume only. For instance, raising the mids is similar to cutting bass and treble simultaneously.  Also, lifting all bands by the same amount is mainly a volume change, not new equalization curves.  This is why sometimes it is stated the 2 knob baxandall is comparable to the 3-band equalizer.  The problem with the baxandall is the implicit mid band is quite wide, which is not the best for guitar.

In my design, I have a MID knob, a three position switch to set the center frequency in three steps (400, 700 and 1100 Hz), a bass/treble balance circuit or TONE control, and a LEVEL control.

With this device I can reproduce the curves from Fender, Marshall, Vox and BMP style controls, so I intend to name it Universal Tone Stack.  Besides, it works also as a three band tone control with adjustable mid frequency.  Very flexible.  I want to arrange and curves showing its performance before posting the circuit.  Perhaps I will arrange a simple web site for this purpose.

[Brett Sinclair]

Pardon my ignorant question, but how do i calculate the frequency/cap value in stm's active three band schem?
It looks like a cool project to make a 5-band EQ based on this...

[Joe Hart]

Good question, Brett. I did some calculating (noticing that the only changes were the two caps and that they were in a 1:10 ratio), but the frequencies came out a little off. I would like to know they actual calculations.

Also, are the pots linear?

-Joe Hart

[stm]

Hi,

I don't have a formula for calculating the frequencies, but it is just a matter of scaling the caps.

For instance, 220n/22n produces a 100 Hz peak/notch. Divide the cap values to increase frequency accordingly.

Pots should be linear, however the preferred type would be this less common "S-shape taper", which I believe it's named "G-taper".

The reason of using caps in 10:1 ratio was to obtain higher Q or narrower bandwidth.

I've been working on an improved version of the three band Eq, and now I've settled for the following bands:

Bass: fixed at 100 Hz, wide
Mid: switch selectable for 400, 630 and 1000 Hz, narrow (as shown in the above schem)
Treble: fixed at 4000 Hz, wide

The above allows greater flexibility in obtaining the response curves of most tone stacks, including Fender (400 Hz mid), Marshall (630 Hz mid) and BMP (1000 Hz mid).

Regards,

STM

[Brett Sinclair]

Thanks for the reply!

So, to further understand how it works... for example to make the Q for the 100Hz band in your scheme wider, you would increase C9 to 220n?

[Ben N]

In my design, I have a MID knob, a three position switch to set the center frequency in three steps (400, 700 and 1100 Hz), a bass/treble balance circuit or TONE control, and a LEVEL control.

I imagine what you have planned here is an active control, with much deeper mid boost/cut available, but it sounds, interms of functionality, anyway, something like the enhanced Big Muff Pi stack (with "Body" control) proposed by Jack Orman.

Let us know if/when you post this stuff--sounds cool, very useful.

Ben

[stm]

This is my actual 3-band Eq.  I call it semi-parametric due to the switchable mid frequency selection in three steps.  If you look at the previous circuits I posted on this thread you will see I've been tumbling around eqs for a while.  This is the design I like the most so far, very flexble indeed.

Notice the Bass and Treble bands are wide.  In particular, low frequency band is suitable even for 4-string bass guitars tuned in E (41 Hz), and particularly for drop-tuned guitars.

Also, a good characteristic is that it doesn't end up at 20 Hz or 20 kHz with tons of gain as opposed to conventional Baxandall tone stacks.

In addition, boosting both Bass and Treble will produce a sort of "mid cut" effect which is DIFFERENT from just cutting the Mids (much narrower).  In this aspect, there are different possibilities to play with while adjusting the controls, and hence more curves to dial in in comparison to a two-band equalizer.

This is the schematic.  Notice an input buffer and gain leveling stage are needed.



These are the performance curves for each of the bands set at maximum boost.  Mid band shown at 630 Hz only.



Ben: I abandoned the TONE+MID circuit because it was less flexible than the three-band eq.  Anyway its simplicity is interesting and it has much less interaction between the tone and mid frequency controls than the AMZ Body control because of the use of an active design with separate op-amps.  Unfortunately I don't have it here at the office to post it.

Best regards,

STM.

[Space Jm]

Hi STM,

I am about to build a preamp with EQ active control and I studied your 3 band schematic and  found it very very interesting.

However, I am just wondering something about +/-15dB. In my understanding, with your schematic you can only boost the bass+ 15dB by raising the 100k pot ? Is it true ?
IN fact your can not decrease -15...a  band ? Am I right or not ?

Thanks for your feedback.
Space Jm

[stm]

Hi Space Jm,

The circuit has BOTH boost AND cut capability.  The curves in the above picture show only the boost capability so as to simplify the graph.  Cut curves are mirror images with respect to the 0dB line.  Don't forget adding a switch for the midrange frequency greatly expands the flexibility.

If you want a simpler design you may try the Tonemender also, which is allows getting most of the Fender, Marshall and Vox tonestack curves: http://www.runoffgroove.com/tonemender.html

Best regards.

[Space Jm]

Hi STM,

Thanks very much for this very clear explanation. I will try this week end.
See you later.

[didier]

what program did you use for those simulations?

[Space Jm]

STM,

I just check the Tonemender, is it an active EQ of not ? Because I can not see 1Aop for each band like your design...
Thanks

[Mark Hammer]

G'day gents, I'm a little late to this one.  Apologies.

The original request was for a dedicated 3-band thing to tack at the end of a distortion circuit.  I'm going to be a bit of a heathen and suggest that much of what has been proposed so far is more or less inappropriate to this particular task.

Why, you ask?  Well, let's start by asking the question "What are equalizers normally used for?".  The average EQ unit, regardless of how many bands, is intended to address a signal whose properties you have absolutely no idea about.  Could be a symphony coming through large home speakers.  Could be a kick drum you're trying to keep from generating wolf-tones in a small performance space.  Could be a singer with a reedy voice you're trying to make sound a little huskier.  The point is that the bands, bandwidth, and number of controls, are really a function of not knowing what the unit will have to do in future.  The epitome of this is the multi-band parametric EQ where you have adjustable selectivity and resonant frequency "just in case".

But what about distortion pedals?  Well, by definition, they add harmonic content, and by popular demand they should generate less fizz and better fizz-to-fuzz ratios, in which case any sort of Baxandall type treble boost function is a sort of irritant at worst, and moot at best, since half of what it does is to introduce more of what we apparently don't want.  Additionally, where non-fuzz EQ systems are intended to eliminate undesirable characteristics (and unintelligibility in a mix can be thought of as one of those), EQ on a pedal is really intended to change character, both in terms of emotional properties and in terms of personage. 

So what is it about the "character" of a fuzz that strikes us as important to pay attention to?  Well, I don't think this is anywhere near an exhaustive list, but certainly one of the things that is important to us is how "angry" the pedal sounds.  Some overdrives sound in earnest, while some fuzzes sound irritable and still others sound positively angry and overwrought.  The emotional tone of the output is partly a function of the distortion control, but also a function of the extent of upper harmonic content that comes out.  Showing "restraint" involves cutting back or filtering out the raspiness.  And that is likely best done with some sort of lowpass filtering whose properties would depend on how much nonadjustable filtering occurs elsewhere in the pedal, plus how much additional harmonic content is introduced by the clipping circuit in the first place.  It is easy for me to imagine contexts where there is a need for more aggressive lowpass filtering, and other contexts where something as flaccid as what you get on a TS-9 is perfectly adequate.  Does anyone need treble boost?  Depends where it goes.  Upper-mid/lower-treble boosting capabilities are great for pre-clip applications since they change the quality of the distortion produced, and indeed you'll find a number of pedals that use fixed resonant pre-boosts to do just that (A lot of DOD's pedals really differ in terms of the pre-boosts used).  On the output, though, I generally find that a Rat-type lowpass filter has tremendous power in changing the critical high end that determines perceived character.

Another aspect of distortion character might be called personage.  Some units have a distinctive personage much like Dustin Hoffman's Ratzo Rizzo character in "Midnight Cowboy", or some roles that Ben Stiller has had; small, puny, bereft of physical power, but so irritable that you just get out of the way.  Here, we are often discussing units that produce a more nasal, strident, and bass-weak (or at best bass-neutral) sound.  Jordan Bosstone could be the poster child for this.  At the other end of the spectrum are those units whose personage is the sort of character that picks up your 4 x 12 stack, mutters "Where do you want it?" in a surly tone, and tosses it in the van without blinking or sweating. here, we are often talking about a sturdy big bottom, some mid scoop to accentuate the bottom, and just enough upper mids and treble to indicate determination and force of will.

What I get out of this ridiculous digression is the sense that the "ideal" 3-control for a post-clip EQ would be a variable lowpass filter (with maybe a bit of resonance), a nice shelving bass control, and a midrange resonant cut with a very shallow and broad scoop.

Is this "enough"? Well, to be fair and honest, and to echo the cornucopia of distortion tones that people have exhibited in this forum, a great many distortion circuits come with their own intrinsic character, and it is folly to think that we can turn any single one into any single other simply by dint of the EQ controls on board.  The difference between a Big Muff, a Fuzz Face, and a Scrambler is a whole lot more than the EQ settings.  So maybe the thing to consider is what the range of personalities a given circuit can have, and design an EQ circuit to make them more visible/audible/accessible.  This is clearly an argument for the idea that there can be no single "generic" 3-band system, regardless of how well it is designed (and Sebastian's stuff is always top-notch). 

Instead, the ideal 3-control system should be planned out based on:

• the inherent character of the clipping circuit itself, including any fixed filtering it might use (e.g., small input capacitors)
• whether the EQ is going before or after the clipping circuitry, or even divided up between those locations
• what sorts of personality changes can be extracted from the unit
• how huge a contrast from the bypass tone the units "default" personality is

[WGTP]

To add my 2 cents, reviewing frequency response curves of various 12" speakers such as Celestion and Emenance, it appears that most have various amounts of dip (from 0 - 10db) at around 1.5Khz and boost at around an octive higher at 3 to 4Khz.  Most are fairly smooth from the 1.5Khz dip down to their bass limit with slight differences in the mid-range. 

To simulate different speakers (which I think is an often overlooked contribution to the tone) a pretty narrow dip at 1.5Khz and boost at 3Khz is needed, along with a boost and cut at 750Hz and 200Hz.  The low 200Hz should cover the difference between a 4-12" stack and an open back single 12" cabinet.   

[Joe Kramer]

Hey Joe H!

Here's a suggestion no one's tabled yet (I think): If you want the ultimate in CONTROL, you could try a state-variable filter (a la Anderton's Super Tone Control).  It only requires a quad op amp, very few parts, and it'll give you just about every response-curve in the book.  May take up a bit more knob real estate than you want though.  FWIW.

Joe K

[stm]

The Tonemender is a passive tonestack wrapped around buffers to allow having unity gain or even boost capabilities without loading your guitar.  The tonestack topology comes from a Fender tonestack and its values have been optimized for:  1) clean guitar equalization,  2) smooth and progressive control throughout the rotation of each pot,  3) reproducing as much as possible of the traditional fender, marshall and vox responses in a single unit (thus the switchable mid frequency).

On the other hand, I also use the proposed 3-band active equalizer to buffer and adjust my guitar eq prior to other effects.

You can still use the 3-band or tonestack tonecontrol approach to adjust the bass and mids at the output of a distortion pedal, however the high control might not be adequate if you don't provide extra high frequency removal.