Two band Parametric EQ

[jatalahd]

You will need the input buffer as shown in the BYOC schematic, but leave R4 out as it most likely will attenuate the signal. You will need to increase the value of C1 of course. The added 3.9k resistor causes the input impedance to be low in the circuit, but R11 and VR2 are taking the input impedance down as well. The buffer will fix both of these.

I would also leave C4 and C9 out, but you need to test if they are really needed. C2 and C3 are there most likely to isolate DC from the cut/boost pot, which will scratch with DC offset. This leads to think that maybe caps are needed close to VR3a/b as shown in Rod Elliott's page.

For output C18 and R49 are enough, but buffer would be preferred here also. Without buffer, the second cut/boost stage is vulnerable to be loaded by following stage and will reduce gain.

Before making any soldering work I would suggest building the circuit on a solderless breadboard:
https://en.wikipedia.org/wiki/Breadboard#/media/File:Electronics-White-Breadboard.jpg

There you can test how the circuit actually behaves with/without capacitors and try out different components.

Also if you notice in testing phase that the circuit is noisy, as commented in the BYOC comment section, you most likely want to read this:
http://www.byocelectronics.com/board/viewtopic.php?f=33&t=53506

[Guitar_Duderino]

Thank you again, jatalahd, for your great help.

You will need the input buffer as shown in the BYOC schematic, but leave R4 out as it most likely will attenuate the signal. You will need to increase the value of C1 of course. The added 3.9k resistor causes the input impedance to be low in the circuit, but R11 and VR2 are taking the input impedance down as well. The buffer will fix both of these.

Of how much should be my C1?
And R2,R47 = 5.1kΩ, R45,R46 = 1.6kΩ, R48 = 330Ω ?

I would also leave C4 and C9 out, but you need to test if they are really needed. C2 and C3 are there most likely to isolate DC from the cut/boost pot, which will scratch with DC offset. This leads to think that maybe caps are needed close to VR3a/b as shown in Rod Elliott's page.

And there they are of 100μF instead of 10μF.

For output C18 and R49 are enough, but buffer would be preferred here also. Without buffer, the second cut/boost stage is vulnerable to be loaded by following stage and will reduce gain.

From simulation (above) input buffer cuts all output in dB and output buffer boosts/recovers it.

Before making any soldering work I would suggest building the circuit on a solderless breadboard:
https://en.wikipedia.org/wiki/Breadboard#/media/File:Electronics-White-Breadboard.jpg
There you can test how the circuit actually behaves with/without capacitors and try out different components.

Yes. With so many options for capacitors I already wanted to do that.

Also if you notice in testing phase that the circuit is noisy, as commented in the BYOC comment section, you most likely want to read this:
http://www.byocelectronics.com/board/viewtopic.php?f=33&t=53506

Had seen it, yes, and considered it. Since I want my EQ on 9V battery, some of his changes I cannot do. He also seemed to have problems with low quality parts from BYOC...let us see whether these happen to my circuit.

I intend to put 125mW resistors, as BYOC indicated. May this still be? Would 250mW be apter?

[jatalahd]

Of how much should be my C1?
And R2,R47 = 5.1kΩ, R45,R46 = 1.6kΩ, R48 = 330Ω ?

C1 should have its value so that the input buffer has its high-pass cut-off lower than the lowest frequency of your filter frequency range. Since you have the SPICE simulator up and running, use it to find suitable value. If 0.1uF is not enough, try 0.47uF or 1uF. Generally for the buffers you can select freely the component values. If the values shown in BYOC are working for your circuit, then use those.

I intend to put 125mW resistors, as BYOC indicated. May this still be? Would 250mW be apter?

My opinion is that there should not be any difference between 125/250 mW resistors, maybe using 250mW is more common. Better to use 1% tolerance metal film resistors in the filter sections at least. For capacitors in the filter sections I would recommend using 1-2% tolerance film (polystyrene) capacitors, as with relatively high accuracy in value, they will not limit the high-Q behavior of the filter that much and seem to have lower noise.

[Guitar_Duderino]

Just this



works well. There is some attenuation in dB (compared with guitar directly connected to amp). But at some time I heard some click and sound went up. Might be some poor connection on breadboard?...

With that input buffer there is no attenuation (perhaps even augment), but clean sound becomes slightly distorted. Tried output buffer and did not notice any difference. Why this distortion? What would solve it?

[jatalahd]

With the circuit shown, the attenuation might be due to the small input impedance, as we noted earlier. By placing a 470k resistor in parallel with small input resistance will not change anything. With buffer in front, the input impedance problem is fixed.

With the input buffer added, try to isolate the part in the circuit which distorts. I would set both VR1 and VR4 to midway, so there should be nothing going into the filters, then you basically have two unity gain op-amps in the signal path and there should be no boost nor cut. Is there still distortion? Is there still attenuation or gain in the signal? Show us the schematic with the input buffer attached to the circuit. If there for some reason is gain at the input buffer, that might overdrive your amp, especially with the boost activated.

[Guitar_Duderino]

Thank you. I will try that.

Besides that there is truly some rain noise (not solved by buffers) independent from pickup volume...obviously stronger with distortion effect. Is it from these opamps? From this voltage inverter? I read that thread (linked above)...but its solution was to use stronger opamps, which drained 9V battery very fast...and I need this to run on 9V battery.

On Reviews section of BYOC (https://buildyourownclone.com/products/paraeq) someone says that it is very noisy, someone says that it sounds amazing. How can this be? And here (https://youtu.be/KHSdJiRUaE8) at 2:54 some noise is heard...

Above all I am trying to do something, which was already done in the 70s... What opamps and voltage inverters would they have used and not get noise?

[antonis]

there is truly some rain noise

Falling on tin or tile roof..?? 



P.S. Just for info, it doesn't help a lot.. 

[Guitar_Duderino]

hehe Being no rain noise expert I here put this sound file.

https://www.dropbox.com/s/ydpnywmkfmjbtyi/EQ%20noise.m4a?dl=0

From clean to strong distortion...here and there moving frequency potentiometers.

With the input buffer added, try to isolate the part in the circuit which distorts. I would set both VR1 and VR4 to midway, so there should be nothing going into the filters, then you basically have two unity gain op-amps in the signal path and there should be no boost nor cut. Is there still distortion? Is there still attenuation or gain in the signal? Show us the schematic with the input buffer attached to the circuit. If there for some reason is gain at the input buffer, that might overdrive your amp, especially with the boost activated.

With input buffer (just opamp added to circuit above after R1) and VR1 and VR4 at middle postition there is no distortion and I think sound (dB) is equal to guitar directly to amplifier. As VR1 and VR2 are turned to maximum, so particularly some frequencies distort sound.

[jatalahd]

Sounds bad.

Since we have gone so far with this project, I decided to follow you on the build and wired the first stage (50 Hz to 2 kHz) on my breadboard. Since I had only one quad op-amp chip left I used that with basic emitter-follower BJT buffer infront. The op-amp used was LM224N and BJT was BC549C. In addition I used two 9 volt batteries with 1000uF caps connected parallel to both batteries, then also mostly basic carbon resistors and polyester film caps (basic caps, no fancy HiFi caps as I proposed).

The result was surprisingly good. Most importantly, the circuit had actually less noise than my commercial DS-1 Distortion pedal, even with full boost. And it gave a spot on steady +20 dB gain for the frequency band that we calculated (well, I tested only from 50 Hz to 1000 Hz...)   I made a video recording when I have full boost on and I am feeding the circuit with 380 mV sine wave from my laptop. The circuit output is 3.60 volts (basically 20 dB gain). The output waveform does not show any sign of distortion. Video here:

In the end you will hear some additional noise, but it is due to the fact that I had my window open and the smartphone audio recording seems to make a mess on the noise from cars driving by. So not a good proof, you need to take my word for it ...

However, when I plug on my guitar, there is still no "rain noise" as I am not playing, but when plucking the strings I get a buzzing distortion mostly on thick strings (low freqs). This happened even with full cut activated. At this point I am not sure where this distortion is coming from, maybe the caps are microphonic or ... well just don't know yet.

Maybe we find the solution after some more testing.

Your noise problem is most likely due to the power source or making wrong ground connection at some point of the circuit. In the beginning I used to mess up all op-amp builds by connecting the input to -Vcc instead of ground, but then I did not use two supplies, only one split supply, so the ground can be misleading concept in that case.

Again try to build your circuit piece by piece while keeping the audio path connected. Then you will most likely find out where the rain noise is coming from. That should be fixable, but then we still need to find a fix to the buzzing distortion problem.

[Guitar_Duderino]

Again most grateful, jatalahd, for your help in this since its inception.

Tomorrow (Monday) I will immediately avoid that Max1044 charge pump by connecting two 9V batteries. Let us hear what that does to noise. Someone suggested in its stead ICL7660S. Any opinion on this?

On distortion someone told me to connect to "aux in" on my amplifier or some direct channel to avoid some gain, which clean or other amplifier channels may have, adapted to passive guitar circuits...

Keep me informed on exact elements of your circuit. Did you put C1 R2 R3? But this filter is not necessary, right? R4? I did not feel much difference between input buffer with R4 and without R4. What of C2 C3 C4? Do you understand their effect through your oscilloscope? And what effect C17 and C18 have? C18 is connected reversely compared with C7 of ESP. And ESP has that C1...to what effect? Putting ESP C4 and C5 on my circuit annulled any EQ effect...either that or I heard no guitar sound...cannot rightly remember.

On my breadboard I simply have that circuit from my last image (and charge pump). I have Vishay 1% capacitors, Multicomp 1% 125mW film resistors... What do you say about those C19 C20 C21 (see BYOC circuit)? Across websites I saw various values for these capacitors... BYOC image has all of 10μF, but its kit then has C20 = C21 = 220μF. Elsewhere someone had C19 = 47μF and C20 = C21 = 100μF. I tried all these and heard slightly different rain noises... From these options that with all of 10μF sounded less bad...
Just low band connected produces some rain noise, just high band connected produces slightly more noise (or more audible), both bands connected produce most noise.

[jatalahd]

I got rid of the distortion by changing the quad op-amp to LM348N. Don't know why that fixes the distortion. Experts have told me in this forum to not use LM348 for audio (it is very old uA741 based design), but for this it seems to work. Now there is very little background noise and no distortion while playing, so it is good to go

The oscilloscope still gives +20 dB gain at max boost from 50 Hz  to 2 kHz. Now I made a video playing with the guitar though it. The samples are "full boost 100 Hz", "full boost 2 kHz", "full cut 2 kHz" and "full cut 100 Hz". Then some random stuff at medium state followed by full boost randomness:

Unfortunately I could not find my schematic anymore so I could have added the input buffer I used. But essentially my circuit is the first stage of your latest schematic (with mods explained after...)

Just this

I added BJT emitter follower buffer with 1uF input capacitor in front, which is connected through 10uF cap to your R1 (actually I did not have R1 in my circuit). VR1 was 50k since I did not have enough 10k pots. Then I had the filter without the Q controls VR2, R11, R12. Then I just had 1uF cap at the output of U4, from where I took the signal to my amplifier. Otherwise same components as in your drawing. According to the oscilloscope trace, there might be some DC offset around VR1 when turning into full boost mode so maybe it is good to have caps at both sides (as in the BYOC circuit).

Hopefully you have had some progress to suppress the noise and distortion. It seems now that the circuit can be made to work properly (Although I noticed some tendency to feedback when turning the amp volume high  )

[Guitar_Duderino]

Excellent, jatalahd! And any noise with strong distortion effects? (My noise is barely heard with clean sounds.)

If those capacitors are polarized, how did you put them?

Unfortunately here noise continues with two batteries instead of Max1044 charge pump...

Noise is heard on breadboard and on PCB (I made one)...

[jatalahd]

I tested now by placing my DS-1 distortion pedal between this EQ and amplifier. I put full distortion on the DS-1 and full boost on EQ. Noise is there, just like in the youtube video you linked previously, but I don't think it is as loud as in the sound clip you shared. Here is the video of my testing:

If the DS-1 pedal is placed before the EQ circuit, there is no noise:

Noise will definitely arise if you are placing an amplifier circuit after an amplifier circuit, The EQ filters a large portion of the noise, but the DS-1 amplifies it with large bandwidth. Maybe some guru has some idea how to avoid it. I don't know how.

The 1uF caps are not polarized, the 10uF is polarized and when using the transistor buffer, the + side is obviously connected to the emitter (I have wired the tranny to use only VCC +9V, so the emitter voltage is about +5V).

[Guitar_Duderino]

Thank you again.

Aah... I recognize your EQ noise...... I recorded my file with loud sound. And your EQ noise might not be so loud having one band... My EQ noise becomes louder with two bands...

So whence this noise? Opamps? I want this on my guitar...onboard...so no other option there... How can this noise be diminished? Can those three capacitors you put (input + between input buffer and first band + output) do something to it? I do not have 1μF capacitors now...but have 0.1, 10, 47, 100, 220μF. What should be my criterion for choosing these capacitors at these three points?

What simple circuit may be done to hear opamp noise?

[jatalahd]

What simple circuit may be done to hear opamp noise?

Well, EQ gives +20 dB, DS-1 might give another +20 dB on top (it is an op-amp based distortion). Just take two op-amp circuits with +20 dB gain each, connect them in series and I am sure that you'll hear noise (The gain is then 100 x input). If no noise yet, then add gain until you do hear noise. Noise is everywhere in every circuit, you cannot avoid it. The more gain you want, the more noise you get.

Your second stage surely adds more noise since it boosts at so wide frequency range and I would say all frequencies above 10 kHz are just creating more noise than audio.

Those capacitors that I added are just there to block DC, they don't affect to the noise in any way. You need to choose the capacitors to allow your lowest frequencies to go through without attenuation. If you don't have 1uF at the moment, use the 10uF instead (for testing purposes).

As I noted previously, I don't have any solution to the noise problem. In my opinion the circuit works well, since as a standalone unit it passes audio cleanly through. As an idea though, if you are using the circuit always with distortion, you could attenuate the signal at the EQ output. Maybe something like a 12 dB attenuation could make the noise level so low that the distortion stage boosts it to "normal" noise level. You can test this attenuation simply by using a 500k pot at the EQ output as a voltage control. Now you are injecting an already boosted signal to a distortion stage, which in my opinion makes the distortion sound bad.

Other option is to try the BYOC trick, attenuate at the input and restore at the output.

But I think I cannot help any further here. Hopefully others in this forum come and share their knowledge.

Or you can start a new discussion topic concerning the noise problem.

[ElectricDruid]

It's not only a question of where the noise is coming from in the circuit - what makes you think your input signal is so quiet?! There'll be some noise going in, so if you boost it x100, you've got a significant noise level, even if the rest of the circuit is perfect (which it's not and never will be).

All you can do is make sure your input has as little noise as possible, and make sure your gain stages *only* amplify the audio you're interested in (e.g. highpass and lowpass the signal to just the audio range).

It's audio. We're always battling the noise floor. Sigh.

[Rob Strand]

Falling on tin or tile roof..?? 

P.S. Just for info, it doesn't help a lot.

You could get the wrong message with the rain analogy.  On the whole for electronics Shot Noise is white (but there are cases when it's not).

Shot noise is like rain because it come is "drops" or packets.   When you have enough drops, as you do with any reasonable current ,it becomes indistinguishable from thermal noise.

The sound of rain as we know it has filtering (ie. EQ) from the acoustics of the environment.

https://en.wikipedia.org/wiki/Shot_noise

Shot noise is distinct from voltage and current fluctuations expected in thermal equilibrium; this occurs without any applied DC voltage or current flowing. These fluctuations are known as Johnson–Nyquist noise or thermal noise and increase in proportion to the Kelvin temperature of any resistive component. However both are instances of white noise and thus cannot be distinguished simply by observing them even though their origins are quite dissimilar.

[Guitar_Duderino]

Hehe Thank you, ElectricDruid and Rob Strand.

Playing through this EQ with clean sounds...no problem...with distortion effects comes that noise, but it does not bother playing...only when not playing it is heard... I will suppose it to be normal for now... I did not thought it normal, since in all these years of playing with many effects I had never heard such noise...

About input and noise... My guitar wiring now is:

Pickups
>Potentiometer 500kΩ linear
>(optional) Equalizer (in: 470kΩ, out: 100kΩ)
>(optional) effect circuit (in: 1.5MΩ, out: 47kΩ)
>(optional) Preamplifier PA-2 (in: 500kΩ, out: 2kΩ) "It is meant to be used at output of guitar just before output jack" (EMG page)
>Output jack
>Other pedal effects etc.

(optional = may be turned off)

Any errors in this order and in match of input and output impedance? What do you say? Can you shed some light on choosing these impedances?

And on connecting electrolytic capacitors parallel to battery... Equalizer has one of 10μF...effect circuit has one of 100μF...Preamplifier is EMG PA-2...do not know what capacitor it has... Remembering that Equalizer and effect circuit may be turned off through external bypass (battery and signal cut off) but Preamplifier through internal bypass (thus it is always fed) do these three circuits need to have these capacitors or could just ONE capacitor in ONE circuit (hopefully Preamplifier) cover this need? What do we want with this capacitor? How should we choose it?

I now wonder...since this PA-2 on off position makes input signal go through voltage follower (so I interpret from its instructions pdf), could it come before Equalizer and avoid need for Equalizer input buffer? It would be good for power consumption and space...

[Rob Strand]

If you have this EQ -> Distortion then the distortion will boost the noise.

You should have a buffer before the EQ.  You can greatly decrease the noise using the technique known as pre-emphasis/de-emphasis.   You boost the treble before a noisy circuit then cut at the output.  The treble cut at the output cuts the noise of the circuit.   The overall response is flat.

You can see the idea here:
http://guitar-fx-layouts.42897.x6.nabble.com/file/n39376/m114.gif

The first opamp is a buffer at low frequencies but the 51k+12k+2n2 boosts the high frequencies.   Then at the output the 5.1k + 1.2K +22n cuts the treble so the response is flat.    For best results you should use small resistors around the buffer like 10k or less , the 51k is a bit large and will add noise.  For example you could just use 5.1k + 1.2k +22n at the first opamp.

[Guitar_Duderino]

The first opamp is a buffer at low frequencies but the 51k+12k+2n2 boosts the high frequencies.   Then at the output the 5.1k + 1.2K +22n cuts the treble so the response is flat.    For best results you should use small resistors around the buffer like 10k or less , the 51k is a bit large and will add noise.  For example you could just use 5.1k + 1.2k +22n at the first opamp.

With 5.1kΩ 1.2kΩ 22nF at input and 5.1kΩ 1.2kΩ 22nF at output:




Is this it?

Not having here those parts I tried with 5.1kΩ 1.6kΩ 100nF(non polarized)/10μF and 220μF (polarized). This is its noise at those positions from above graphic with Guitar (turned off) > Equalizer > Amplifier (with strong distortion):

(no emphasis till 5s, emphasis with 100nF till 10s, 10μF till 15s, 220μF till 20s)
https://www.dropbox.com/s/i722osmgqk25h5k/nul-0.1-10-220.m4a?dl=0

Did you expect this? What values will diminish most noise? Is effect different between polarized and non polarized capacitors? Which should be preferred? From these three I incline more to 10...220 almost introduced some wavering high noise compared with 10...

And, ElectricDruid, you said

what makes you think your input signal is so quiet?! There'll be some noise going in

But here with pickups cut off what noise input is there?