Parametric eq noises scares me

[Elijah-Baley]

Hi, guys.
I really need to build an equalizer with centered frequencies. I built a 2-band Baxandall, and even a 3-band Baxandall. Both work fine, but I found the band too large.
Once I built a 5-band parametric eq from Sabrotone (GGG schematic), and it was really cool. But it was too noisy, not buzz, but something like a white noise, not so loud but annoying in the clean channel of the amp, and unusable in the gain channel. I was pretty sure the board was ok, I used TL07x IC, and I tried to put it in a box, and nothing solved the problem.
Now I need to try something like that, again. This time I want to try this: https://www.sabrotone.com/?p=578

Schematic:



Layout:



Probably, later I'll change some values to get a larger range of frequencies.

I'm wondering if the parametric eq are naturally noisy, and I afraid to get in the noise issue.
Probably, I'll put a high inpedance preamp in front and use it with an acoustic guitar.

Did someone build this eq? I need it will be absolutely clean and noiseless.

Thank you very much!

[ElectricDruid]

Anything with high gain is likely to suffer from noise. It's simply that there's always *some* amount of noise in a circuit, and the more you boost it, the louder it gets. A parametric EQ can use quite high Q values, so you're getting significant gain in a narrow band. That will emphasise noise in that band. That much is unavoidable.

However, the circuit you show uses a 1044 to generate a negative supply. That's not unavoidable, and I'd be surprised if a circuit based on one was as quiet as one *not* based on one. The switching frequency is well outside the audio band, so in theory it's not audible, but I've lost count of the number of times ultrasonics have shown up in my circuits as audio-band noise so I'm not convinced.

[antonis]

I've built many of single supply RG's parametric EQs variations with absolutely on issue..!!!
(just use the reasonably lowest resistor values for bias & gain and good op-amp use practice..)

[Elijah-Baley]

Anything with high gain is likely to suffer from noise. It's simply that there's always *some* amount of noise in a circuit, and the more you boost it, the louder it gets. A parametric EQ can use quite high Q values, so you're getting significant gain in a narrow band. That will emphasise noise in that band. That much is unavoidable.

However, the circuit you show uses a 1044 to generate a negative supply. That's not unavoidable, and I'd be surprised if a circuit based on one was as quiet as one *not* based on one. The switching frequency is well outside the audio band, so in theory it's not audible, but I've lost count of the number of times ultrasonics have shown up in my circuits as audio-band noise so I'm not convinced.

I built some years ago exactly this: https://www.sabrotone.com/?p=99
I open a thread about my noise issue, here: https://www.diystompboxes.com/smfforum/index.php?topic=116984.0 But I never solved the problem.
I think to remember the white noise was persistent at any setting.

I've built many of single supply RG's parametric EQs variations with absolutely on issue..!!!
(just use the reasonably lowest resistor values for bias & gain and good op-amp use practice..)

What schematic by RG you followed?
Power section 9v, or +9/-9 change a bit ground, Vref and -9 point. I don't know how much it influence the noise level.
The one I want to build have the +9/-9. If I understand this should be a good idea.

[thetragichero]

I'm a big fan of keep it simple stupid so without a specific need for it (i can't see one here) I'd leave off the charge pump
a resistive divider and voltage follower (if you have a spare op amp half) should be more than sufficient

[PRR]

> it was too noisy

At what signal level??

The basic EQ shown will have significant hiss when worked at Guitar Level. Brought-out by high gain after.

You can add say 4X gain in front and 4:1 divider after to get the signal over the noise yet still come out the same. An advanced technique is a high-lift in front and high-cut after; highs are weak and boosting won't clip before bass/mid, and the hi-cut after cuts hiss.

Didn't BOSS make a million of these? Steal their plan. I recall it was not perfect, but often useful.

[Elijah-Baley]

The old one, the 5-band, was too noisy at the same volume of the bypass signal, probably even in flat. It was better not have the eq in the signal. In the gain channel I really can't use it. Too much hiss.

PRR, thanks, but I really don't know how I can do (if I can) what you suggest to me. It seems complicated.
Anyway, I want to really hope that my failed old 5-band board had a mistake, though I didn't find any, or had something wrong or whatever. Because I can't believe there's around a project, verified, but pratically useless, and nobody admit or at least noticed it.

[idy]

That 3 band is a great tool. I use one as an acoustic preamp. No noise issues, but folks are right, a high gain + charge pump may whine. Ideal would be use an 18v tap and voltage divider. And try ig first with just 9v, maybe enough.

[Elijah-Baley]

Idy, did you build exactly the same 3 band tool in the schematic and veroboard?

I built a Klon Centaur using a ICL7660S-CPAZ, and I got no problem. I thought the MAX1044 in this project was ok.

Anyway, I'd like have a 9v version, I have to use it in a acoustic preamp like you, idy.
I could edit the veroboard taking off the MAX1044 and its two 10uF caps C9 and C13. And to connect pin 4 of the TL071 and pin 11 of the TL074 to the ground.
For voltage divider we mean the Vref, VB, 4.5v, etc? Because I don't see where I need it.

[idy]

Yes, I used the Sabro layout you posted, I then added a pot to the feedback loop of the last opamp stage (IC1D) to give a master gain. It has no issue with noise from the charge pump. I really like having a master gain on a preamp!

I think your new voltage divider with it's VB would serve R10 (to bias the first opamp) and maybe (just guessing, do gyrators go to ground or VB?) the frequency pots. You are changing a bipolar supply to single, some of the grounds will go to mid point.

Some one who knows will pitch in, will the gyrator stages need to be biased also? They are not DC coupled....but IC1D is.

[Kipper4]

is it strange that R11 and C11 don't go to -9V?

[Elijah-Baley]

Yes, I used the Sabro layout you posted, I then added a pot to the feedback loop of the last opamp stage (IC1D) to give a master gain. It has no issue with noise from the charge pump. I really like having a master gain on a preamp!

I think your new voltage divider with it's VB would serve R10 (to bias the first opamp) and maybe (just guessing, do gyrators go to ground or VB?) the frequency pots. You are changing a bipolar supply to single, some of the grounds will go to mid point.

Some one who knows will pitch in, will the gyrator stages need to be biased also? They are not DC coupled....but IC1D is.

Ok, thank you!
A gain stage, could be a good idea, maybe I could replace R7, (from the schematic, the veroboard labeling is different).
On the veroboard layout should be easy to rremove the bipolar supply parts, but a bit more hard add a voltage divider root and connect R10. I'd like to know if I should connect to the VB even the frequency pots, indeed. C11 and R10? Ground or VB? It seems ground is ok. I don't know.
I can tell i draw this schematic on my emulator software without the bipolar supply, and I use ground for GND and -9, and it seems work, or at least iget signal and frequency response.

is it strange that R11 and C11 don't go to -9V?

I'm wondering it, too.

[Kipper4]

How close of a tolerance do the resistors have to be?

Every time I tried breadboarding a gyrator eq it failed.

[idy]

I think you should leave some minimum resistor (1k? 5k?) and then add a pot in series for a gain control. Maybe not an issue but that's what I did.

[ElectricDruid]

is it strange that R11 and C11 don't go to -9V?

Not if the circuit uses a genuine bipolar supply, no. You're thinking like a pedalhead, not an op-amp audio designer!

How close of a tolerance do the resistors have to be?

Every time I tried breadboarding a gyrator eq it failed.

Depends how accurate you want the Q and frequency to be to your calculations. If you don't care too much, then wide tolerances are fine. It shouldn't stop it working. If you're trying to build a hi-fi 1/3oct graphic EQ though...maybe you want to make sure all the bands add up like they're supposed to and don't overlap too much in which case things need to be accurate. The problems will always be the capacitors really. 1% resistors are cheap, 1% capacitors not so much! Even 5% aren't that easy to find.

I think your new voltage divider with it's VB would serve R10 (to bias the first opamp) and maybe (just guessing, do gyrators go to ground or VB?) the frequency pots. You are changing a bipolar supply to single, some of the grounds will go to mid point.

Some one who knows will pitch in, will the gyrator stages need to be biased also? They are not DC coupled....but IC1D is.

I would think they need to be biased to Vb. The reason being that if they're grounded, the op-amp won't be able to drive the output negative - it'll already be biased as low as it can go. They might be AC-coupled, but that AC signal is going to want the op-amp to swing down as well as up, so unless the biasing makes that possible, we'll get clipping.

Tom

[Elijah-Baley]

This schematic is a monster http://byocelectronics.com/paraeqinstructions.pdf but really versatile!
I hardly think we could find a veroboard layout of that.

With the schematic of the Sabrotone layout I can't extend to much the range for each band even change some caps and resistors values.

[idy]

But the values on the layout provide very adequate shaping for a guitar. I'd be interested to know what the band ranges are in your simulation...

[Kipper4]

I note the lpf on the input.

[Eb7+9]

I can't believe there's around a project, verified, but pratically useless, and nobody admit or at least noticed it.

This a free forum intended for beginners ... don't expect pro behaviour anything here

[Elijah-Baley]

But the values on the layout provide very adequate shaping for a guitar. I'd be interested to know what the band ranges are in your simulation...

Probably yes, the original values are ok, but it weren't so large.
With Q at max (and full bosted) I got this:
Bass: 124 Hz - 560 Hz
Mid: 330 Hz - 1.7 kHz
Treble: 1 kHz - 4.5 kHz

I should say this I need this pedal to make some frequencies experiments.
I wanted something lower than 124 Hz, it seems it's not so low range. Some acoustic guitar preamp or guitar eq pedals have lower bass band, the simple MXR and the Boss GE-7 has the lower band at 100Hz. Some specs show 60-80 Hz.

Another thing I don't like too much is that the band "crossover" each other. I changed some value of the critical parts, so I moved a bit the bands. I got a larger range: 55 Hz - 5 kHz, but I even got a hole between 200 Hz - 350 Hz, unfortunately. The problem is that the band can reach just a limited range and If I try to enlarge the range more then the Q is too large at the higher side frequencies even with the narrowest set of the Q control. And reaching the lowest set 55 Hz and the higher set at 5 kHz I got this hole in the frequencies.
While where the higher mid band setting finishes the lower high band setting starts.
I hope I explained all this well.

I note the lpf on the input.

This is cool, it's a multi band eq. I got a white noise problem with the 5 band, I'm not sure why. Now, I'm searching something tweakable.

I can't believe there's around a project, verified, but pratically useless, and nobody admit or at least noticed it.

This a free forum intended for beginners ... don't expect pro behaviour anything here

Yeah, I know what you mean. I'm not a pro, indeed, I just know something more than one year ago, but it's true, I use to build pedal effect or circuits on stripboard. Totally amateur.
Anyway, a lot of us are guitarists or musicians, and if someone plays the device built himself strengths and weaknesses could hear. Often noise, oscillation, ticking... are reported.
Well, it's another topics.