Best 3-Band Active Tone Control?

[PRR]

You don't need these caps:



10uFd into 1Meg at input will carry sub-sub-sub-sonics. 0.01uFd is a reasonable value. And less leakage than an electrolytic.

Strongly suspect the three bands could be done in one opamp. Opamp magic.

[stonerbox]

I wanted to make the eq bass compatible but I guess the 10uF is a little overkill. Thanks for pointing that out PRR.

[ct_anthony]

Hi everybody,

I like to bring this thread back again I've used the EQ from STM (well, the mid-control part) quite a bit in the last time... superb control and extremely good noise ratio... Really enjoying this circuit!

What I´d like to ask (or request help) is how to exactly scale the caps for frequency selection... I´ve learned from my MXR 6-Band that I want to boost at 800hz and 1.6Hz for about 6db using a specific amp/guitar combination into the gain channel.. thats why I want to build these two bands using the EQ design from here (best noise performance I´ve seen so far). But the caps don't seem to scale linear: 440/56 != 630/33 != 1000/18.

So my question is, which cap values do I need for almost exactly 800hz and 1600hz ?

Regards, Anthony

Edit: Corrected frequencies

[Rob Strand]

From what I can see the actual frequencies of the MXR 6-band are 784Hz and 1.8kHz.
Using STM's circuit, to match these frequencies, I get  27n/2n7 and 12n/1n2.

However, the final response of the MXR circuit with two bands boosted is *quite different* to STM's circuit with two bands boosted.  This is normal as the two circuit add bands in a different manner.

The bandwidths of the MXR are more narrow and when two bands are added the shape of the bands skew such the frequencies skew upward.

STM's circuit is a cascade structure (series).  You can also use these "Bohn Equalizers" in a parallel structure like this:
[link remove until I find the ckt]

I did not see if the response of this structure matched the MXR closer.

The bottom line is different equalizers structures produce different results.
The MXR circuits are probably one of the worst for trying to understand how the bands interract.

[PRR]

> the caps don't seem to scale linear: 440/56 != 630/33 != 1000/18.

They should. But those are not the values on the "STM" on this page. In a 4-page thread, maybe you are pointing at another plan?

[ct_anthony]

Thanks for the answers!

I've taken the values from here:

Hi, sorry for bringing this old thread up, but I want to build the last EQ circuit displayed here, and I have some doubts about the schem. I drew the things I have doubts about in red. Could anyone confirm if my drawing is righ or wrong? Thanks.

Morocotopo

Well, I'll try the values you have provided... Currently I've breadboard the middle part with a TL072 and used the first half as a simply buffer (voltage follower). I only had a pair of 22n/2.2n caps at hand... but this covers the frequencies already pretty good... and it fits so well that it also nice always on into the clean channel... noise is really good, even on breadboard. I get some really low noise hum (which will disappears once it is shielded) und almost no hiss - perfect! - and way better noise performance than the MXR.

I guess Ill build two of these one-knob-units into one or two enclosures and be perfectly happy

Just some math for background:

100 / 220 ~ 0.45
700 / 33 ~ 21.21
1600 / 14.4 ~ 111.11
5000 / 4.7 ~ 1063.82

I was expecting the relation of freq / cap to be consistent... obviously it is not the case Is there an easy formula to calculate the cap(s) for  a given freq?

[Rob Strand]

I was expecting the relation of freq / cap to be consistent... obviously it is not the case Is there an easy formula to calculate the cap(s) for  a given freq?

The formula for that circuit is quite big.

See Equation 9 and figure 9:
http://leachlegacy.ece.gatech.edu/ece4435/f02/equalizer.pdf

The reason the caps don't scale like you expect is because the pots stay fixed at 100k but other resistors are in different proportions.  The Bass and Treble are have the same resistors but the Mid is different.  The values under the mid circuit should all scale up as you would expect:

(33n*630) = 20790
(56n*400) = 22400
(18n*1000) = 18000
Average 20400 with +/- 10% due to the fact the frequencies are nominal.

It's like Large Cap * Freq = 20400

The other circuit in this thread has different resistor proportions but they are consistent.

Not also the circuit you posted is +/-16dB whereas the other one in the thread is +/-14.5dB or so.   If you look closely and compare the resistors for the Mids in your circuit and circuit you can see they are close.  The difference is what gives one circuit a different boost/cut over the other.

[Edit: Added link to formula]

[PRR]

The Morocotopo drawing's freqs are close but not quite. According to SPICE the numerical idiot we get 16.5dB peak at:

56n, 5.6n = 393Hz  {"400"}
33n, 3.3n = 667Hz  {"630"}
18n, 1.8n = 1,220Hz  {"1K"}

> some math for background:
> 100 / 220 ~ 0.45

Without deriving I multiplied C times F.

56n * 393Hz = 22,008
33n * 667Hz = 22,011
18n * 1,220Hz = 21,978

For given filter ratios (width), C*F is a constant? (0.1% rounding error.)

So for the *narrow* band, use 22,000, divided by F, read that as nFd. Then also get a second cap 1/10th the size.

[ltnpd]

Sorry for reviving an old topic ...

... this is my first post

I mainly have experience with guitar electronics and I've been dabbling with passive filters... so active filters is new for me. I'd like to make STM's 3 band eq into a pedal. I only have experience building booster pedals at this point...

I have trouble understanding how to read the schematic. Please see image below.




I have several questions: (I've been googling and reading Electronics for Dummies ... and I haven't had any luck understanding more unfortunately).
1. I have trouble understanding how to wire the ICs. There are 8 pins but not all the ICs involved in the circuit seem to use all 8 pins?
2. V1 (orange arrow) is the input?
3. Where does the power supply come in (orange circle)?
4. The right part of the circuit is the volume knob?

I realize that the questions sound like basic circuit reading questions, so if you have any links that would help me understand that would be great also!

Thanks for the help in advance,
LTNPD

[roseblood11]

I'd prefer to control the bass before the distortion and the rest after it...

[GibsonGM]

Hi Ltnpd,  welcome to the forum!

Yes, some basic questions, but that is ok.  If you have more questions, feel free to make a new topic "Beginner's questions" or something like that, and we'll all jump in for you! 

1. I have trouble understanding how to wire the ICs. There are 8 pins but not all the ICs involved in the circuit seem to use all 8 pins?
If you look closely, you see "IC1A" and "IC1B", etc.  This is showing us that the circuit uses dual opamps...2 opamps in one IC package.  So the package only needs one power connect, and one ground, leaving 4 inputs and 2 output pins per dual opamp chip.   In other cases, maybe only one half of the dual opamp is used, and the other half isn't part of the circuit....or, in that case, you may sub a single opamp.

2. V1 (orange arrow) is the input?
Yes

3. Where does the power supply come in (orange circle)?
That is the bias supply...power would come from the "top", and points marked "Vcc" (in other cases, "Vref", "Vbias" etc) are connected between the 2 resistors.   This network is assembled somewhere convenient on the PCB when you wire it up.  You don't want power to pass over signal lines, if you can avoid it, but otherwise you place this where it makes sense.   Near where the power enters the PCB is of course best practice...and you take your power to the ICs from there.

4. The right part of the circuit is the volume knob?
Yes. The 1M pot in the feedback loop of the final opamp is setting the gain of the opamp...which is boosting the output signal. 

[ltnpd]

Ah... thanks GibsonGM! That clears a lot!

I was thinking of posting a new thread as a beginner question but since my questions are also specific to this circuit, I thought it would be best to post a reply to this thread. I'm sorry if I violated any rules.

Having this new knowledge and information... I have more questions also... (sorry!):
1. ... So what you are saying is that basically OPA2134 ICs have two opamps. Just to be clear... So IC1a and IC1b is the same single IC but one opamp (IC1a) is for the buffer section of the diagram and the other opamp (IC1b) is for the bass section of the diagram?
2. Also... (See photo attached below) Do all the "+9-18V" points (green arrows) connect in parallel to the battery (bias/power supply)? ... if so, wouldn't that mean that IC2a and "IC3 A and B" are technically connected twice to power (once via "+9-18V" and another through Vcc/2)?


Thanks again!
LTNPD

[PRR]

> another through Vcc/2)?

That's not really "power".

The audio can swing (almost) from zero to +18V. But audio swings both ways. To allow this, we "bias" it to half-way between zero and +18V. The two 10K lower left make a half-voltage reference from the zero and +18V.

[niftywiz]

Can you clarify the output stage opamp connections?

[DavidM]

Hi Paul and folks, sorry for reviving an old thread,

I am testing this circuit (just the mids section) and I would like to swap the 100K pot for 50K. I am doing so and I notice the response and probably the curve is not the same.

Any hint on what I should be targeting to get the same curves and response of the original circuit, but with a 50K potentiometer?

Thanks very much in advance!

David

The Morocotopo drawing's freqs are close but not quite. According to SPICE the numerical idiot we get 16.5dB peak at:

56n, 5.6n = 393Hz  {"400"}
33n, 3.3n = 667Hz  {"630"}
18n, 1.8n = 1,220Hz  {"1K"}

> some math for background:
> 100 / 220 ~ 0.45

Without deriving I multiplied C times F.

56n * 393Hz = 22,008
33n * 667Hz = 22,011
18n * 1,220Hz = 21,978

For given filter ratios (width), C*F is a constant? (0.1% rounding error.)

So for the *narrow* band, use 22,000, divided by F, read that as nFd. Then also get a second cap 1/10th the size.

[Rob Strand]

Any hint on what I should be targeting to get the same curves and response of the original circuit, but with a 50K potentiometer?

If you want to change the pots from 100k to 50k and preserve the response:
- half the resistor values
- double the cap values

[DavidM]

Hi Rob, thanks for your reply!

I know the workaround you proposed works perfect on passive tonestacks (like Fender) and other "simple" circuits, had no idea it would work as well in this application.

Will check and let you know in any case.

Best,

David

Any hint on what I should be targeting to get the same curves and response of the original circuit, but with a 50K potentiometer?

If you want to change the pots from 100k to 50k and preserve the response:
- half the resistor values
- double the cap values