Controlling the amplitude of the output signal from a pedal?

[niektb]

It's probably really hard to see what's going on. I haven't had time to change anything today.

hmm in most breadboards I know, the power rails are splitted halfway. Can you confirm that yours is different?

[GibsonGM]

It's possible that what niek said is the case.  But looks like you have gnd and 9V on the top left side...if you are getting 9V at the IC pin and gnd on pin 4, then that's not the case.   Quick meter check will show that.


OK - I see the bias divider and line going to the "2in+" pin (5), it looks like.     Is the blue wire there (left side) your guitar input, going thru the small ceramic cap?
That seems ok.     

What is the R and cap to ground just to the top right of the IC? I'd omit that and go for unity gain buffer at the moment...same with the output stuff, those 2 caps in series (?) and pot.    I'd send the output out thru 1 cap only, to the amp (start at volume ZERO!), to see it's working.     Appears the pot has wiper going to ground anyway, which would be wrong...have to eat dinner, mess around and I/we will be back in a bit!   

[YurkshireLad]

It's definitely not a split breadboard, I just checked. I haven't had time to remove the gain loop yet, so I put this picture together.







I don't remember where I bought the TL082s from, but there's always a chance they're fake. 

[GibsonGM]

Yurk?  If you just follow the schematic I posted before, you'll confirm if it works or not, and set up a foundation to start doing cool things like adding gain, volume pots and such    most of it is already built.  A buffer will test your opamp.   I don't like your pic, it over complicates things...series caps and all...nah!   

Please just rewire the opamp, even going to the other side and use my pin numbers (1, 2, 3 etc).   Then we can talk about any problems from a common viewpoint.    Seems you're making something that I'm not aware of, lol.   You are close (and what you've drawn SHOULD work, tho parts are not needed), but something ain't right.   I promise, if you wire it correctly, it WILL work, OR the opamp is junk (or battery if flat, or breadboard is shot or...he he).     

[YurkshireLad]

I'll try it tomorrow night. Thanks! Much appreciated.

[GibsonGM]

No worries. Hope to hear that it works   Stick with it and it will. 

[YurkshireLad]

I retire in about 12 years, will that be enough time to get this working? 

[GibsonGM]

If one just takes it step by step...check the voltage divider, check the power supply and ground at the IC...then hook up +,  - and output...one should be able to make it work in say, 10 minutes     But that's provided the opamp isn't damaged or something.  And one has to literally do EXACTLY what is on the schematic.

Once you get the buffer working, it shows that the IC is ok. Then you can add 2 resistors and get some gain going.   Then a pot on the output for volume.   Then a cap in the feedback loop to tailor frequency content.    Then you'll have gone thru about 90% of what you do with opamps in DIY!   That's why I suggest you scrap what you're doing and use my schematic to make a buffer.   If it works, I/we will show you how to add the 2 resistors to give it gain.      

[anotherjim]

Those series caps may be something seen elsewhere in the audio world - where they use bipolar power supplies. With common electrolytic capacitors that are polarized, they shouldn't get a reverse current flowing. With bipolar supply, the signal sits halfway at 0v so it swings above and below that. This means there isn't actually a right way around to fit an electrolytic cap. You either use 2 caps (twice the value because series caps divide and your 10uF are worth 5uF) with the negative ends joined in the middle, or else buy nonpolar electro caps that are already 2 in one can.

An advantage of the single unipolar power is the signal is biased to sit at a positive voltage (4.5v) so where one side of the cap is 4.5v (amp out or in), and the other side is going to 0v (jack in or out via what's plugged in or a pot or pulldown resistor) you can use one ordinary polar electro cap and you can work out which way the cap faces.

[YurkshireLad]

Ok, I rebuilt using GibsonGM's circuit and I think that works. I'm getting a clean sound of the guitar amp, with the same peak to peak as the input, (approx 300mV, depending on how I hit the strings). so the gain is definitely 1. Also, the amp isn't being overloaded in any particular frequency range, which is good.

[GibsonGM]

great!  Now I would connect the volume pot. (1 leg to output cap, wiper to output jack, other leg to ground) so you can control the level.

Then, you get to add your gain resistors!   Go ahead and do what you showed with the 20k and 1k resistors, but I would take them to "Vr", your voltage reference/bias point, instead of ground...that point is AC ground.  You don't go thru and R to get there, the cap will connect directly.

[YurkshireLad]

This is the bit I always have a problem with. Now I get no output from the op amp. This is what I wired up (ignore the pot values, I forgot to change them in the diagram).

Edit: I get output if I add a cap indicated by the red arrow. I get gain and distortion but it has the same problem as before, it sounds like the clean signal is amplified and the gain distortion is really quiet and hissy.

[YurkshireLad]

I may have to record a quick clip of me playing so you can hear it.

I don't have a logarithmic pot so I may have to reduce the gain.

[FiveseveN]

I get output if I add a cap indicated by the red arrow.

You also need an input cap and the biasing network. Like he said,

one has to literally do EXACTLY what is on the schematic.

and that's not what you're showing us.

[GibsonGM]

Remove the 1 Meg on the output.   Bring the 1k in the FB loop to VR, not ground.  The bias network is the center that the AC signal changes around; it's AC ground.  A linear pot will work for volume, just not as well. 


This is what you need to build.  I put a .1u cap on the feedback leg, that's quite trebly, you can use what you like for right now, such as the 10u you had there before; 1u is probably going to be a good value for cutting some bass but not too much.  That cap sets the bass cut when it's in the FB leg.           If you want less gain, make R2 smaller, say 10k (more appropriate for a single stage).

[YurkshireLad]

Success!

Thanks to a statement by GibsonGM that I read but didn't catch the magnitude of.

Bring the 1k in the FB loop to VR, not ground.  The bias network is the center that the AC signal changes around; it's AC ground.

This fixed it for me, and now I understand it, it seems really obvious. I guess the distorted signal was bleeding to ground, or words to that effect.

Anyway thanks!

I had to put a low pass filter (1k + 1nF for now) in before the op amp as I started getting horrendous RF interference. But I get much better distortion now, although I need to play around with tone next.

[GibsonGM]

Awesome! I'm glad you have success; now the rest of it will just be learning the different tricks to make them do what you want!

Yes - the signal has to ride around Vr, the bias point, at ~4.5v....by grounding that part of the opamp to 'gnd', we're actually disturbing the bias pretty badly (that's how I look at it...wants to be AC riding on 4.5VDC bias, and you're grounding it there) - that's the distortion!  And not the good kind     

Textbook opamps circuits mostly show everything based on a DUAL POLARITY SUPPLY (where such points WOULD be tied to zero volts)...since we don't often use them, and have to use that Vr bias network...we have to keep in mind where the ground as far as AC SIGNAL VOLTAGE is located...not DC...any time we see those circuits and want to adapt them for a single polarity supply.

* Another thing that may cause distortion here is simply too much gain. If you replace the '20k' (22k?) with something smaller, even 5k...you may find you get a boost without distortion.  Any time we ask the opamp to deliver more output than it is limited in doing by the power supply, it will cut off and distort. Mess around with values!  Even 10x is going to distort part of the input signal.

Have fun   

[YurkshireLad]

I haven't had time to work on the pedal for a while, but I plugged my guitar in last night. I remembered that the distortion didn't sound right, and it didn't, so I spent time checking and rebuilding the circuit. Later on in the evening, it finally hit me like a quadruple face palm.

My power circuit is using a simple voltage divider, instead of a more complicated schematic I posted earlier. So I don't need R5 shown in - https://i.imgur.com/EwKdbxv.png. There's no short circuit to ground with the voltage divider.

This extra resistor was biasing the op amp to 2.2V instead of 4.5V. I haven't tried removing the fix yet as the realisation hit me when I was staring a voltage divider schematic on a web page later on.

Doh! 

[GibsonGM]

Hey Yurk,

Sudden realizations are great, ha ha!  In the pic you linked, C2 is the path to ground for the signal if the bias R isn't there...at low frequencies, the audio (AC) signal will like that cap more than it likes the opamp input (infinite Z), so bye-bye audio.   

Not sure what you mean by a 'simple voltage divider' eliminating this (show a schematic if you can)...unless there's no cap there.  There should be; it's good practice to filter your power, even from a battery, which is very stable.  It is for 'decoupling' and has several purposes - don't leave home without it.    R5 isn't an extra resistor, it's a bias resistor, and is typically required.   Lecture over, I'm not the most 'theory oriented' person on here!   I mostly copy the conventions of other designers (since they already did all the math!), and they like that R to be there...something about what the opamp pin wants to see for a bias current and so on, out of a textbook I once read but forgot the 'why' about...

[YurkshireLad]

Simple voltage divider like this: