Op-amp booster circuit. Critique/advice sought

[suncrush]

I sketched up this simple op-amp-based boost circuit, and I'm trying to make sure it will work.

I realize I'm reinventing the wheel.  That was intentional; this is "homework" to make sure I understand how op-amps work.

I have a ton of 4558's sitting around, so I'll probably build it with that.



The first gain stage has a pot in it, so that you can reduce the total gain, in case your pickups are hot enough to produce distortion.  The max gain through that stage is 3.27.

The second gain stage is fixed at 2.8.  Max total gain is 9.15, which gives just over an 8-dB boost.  I might add a trimpot in series with R4, to bring the total gain to an even 10.

Does everything look kosher?

What should I use for the volume knob in the output section (R5)?  100KA?

[samhay]

This will not work as you have not biased either op-amp input.
If you are powering this wilth a typical (for stompbox) 9V supply (edit - as it appears you are), you need to created a reference voltage at half supply then bias both (+) input pins to this voltage via resistors. R1 will also have to be returned to this voltage and/or AC coupled to ground with a resistor.

I think you need to spend a little more time looking at actual guitar-centric circuits that use op-amps. I also suggest you don't build anything until you have it working on a breadboard.

Also, I don't see much point in using 2 op-amps here. Any op-amp you can buy can manage a gain of 10, so why not find somthing more interesting to do with the second one?

[suncrush]

Ok, I will read about op-amp biasing and post an update when I have the chance.


So, you would just use one half of the 4558 chip and do all the gain on that side?

I guess you could use a high-pass filter to separate the signal and run an optional treble boost through the other half of the 4558.

[Brisance]

the feedback pot will work as a resistor as the wiper is not connected

[JerS]

Also the audio out should be the one connected to ground - and the wiper to ground should be your audio output.

Good luck! Let us know how it goes.

[suncrush]

the feedback pot will work as a resistor as the wiper is not connected

Edit--Ah, I see what you meant.  Misclick. Easy fix.

[suncrush]

Also the audio out should be the one connected to ground - and the wiper to ground should be your audio output.

Good luck! Let us know how it goes.

  Indeed. oops.

[antonis]

the feedback pot will work as a resistor as the wiper is not connected

That is intended.

So you're of the kind of wealthy guys who make their pedals valuable depending only on the items price...

[suncrush]

I misunderstood what he meant.    I edited that post.  I missed with the mouse when clicking nodes.

[suncrush]

All right, take 2.

Using diagrams from here:

http://ocw.mit.edu/courses/media-arts-and-sciences/mas-836-sensor-technologies-for-interactive-environments-spring-2011/readings/MITMAS_836S11_read02_bias.pdf

I have set up a voltage-following biasing, using half of the dual op-amp, and putting all the gain in the other half.

[antonis]

You still have the non-inverting Input un-biased..
(actually, you have it miss-biased..) 

Delete the GND on bottom of C5 and connect it to U2's OUT..

P.S.
Use the AC coupled non inverting amplifier on second page of your attachement but exchange the R & C1 GNDs with Vbias..

[suncrush]

Oh, ok.  I think I understand what's going on now.  We're trying to get the inverting and non-inverting inputs to the same bias point, so that any difference between them is AC audio signal, and that's what get amplified.

Have I got that right?

[antonis]

Not actually but you're on the right way..

We want to set a "point" on which BOTH Inputs have to "be refered", so Amp will amplify any difference between them beyond of their actual values.
( i.e. - IN on 4.5V & + IN on 4.6V --> Amplification gap of 100mV )

This "point" should equally be laid between power supply rails (+9V & GND) for equal OUT voltage swing...

[ashcat_lt]

You still have the non-inverting Input un-biased..
(actually, you have it miss-biased..) 

How so?  The most recent pic I'm seeing looks pretty much fine.

Delete the GND on bottom of C5 and connect it to U2's OUT..

Why?  It works as is.  You could remove that cap and increase the bandwidth of the boost, but with a big enough cap it shouldn't matter either way.


Now:  Why the bloody hell do you think you need a volume control in this thing?  You going to boost it up (but not enough to distort!) then turn it back down?  Why?  Not enough noise in your signal?


Edit to add - Actually, you should probably put some smallish resistance (560-1K) in series with both opamp outputs because many opamps do stupid things when you try to drive a capacitive load.  Also, when you go to add switching, you'll want biggish (1M+) pulldown resistors on the outside ends of C2 and C1.  The V pot will do it for the output, but I still think that part is just silly.

[suncrush]

Yeah, fair enough.  Since the gain pot can give you anything from 1 to 11x, the volume is unnecessary.

Now, that being the case, would it make sense to use an log taper pot instead of a linear taper in the feedback loop?



And, for my education--what does the pull-down resistor do?

[ashcat_lt]

Yeah, fair enough.  Since the gain pot can give you anything from 1 to 11x, the volume is unnecessary.

Now, that being the case, would it make sense to use an log taper pot instead of a linear taper in the feedback loop?

Log usually.

And, for my education--what does the pull-down resistor do?

Gives the charge on the cap somewhere to bleed to/holds that end of the cap at 0V so it doesn't suddenly discharge when switched into circuit and create a pop.

[duck_arse]

..... because many opamps do stupid things when you try to drive a capacitive load. ....

this also applies to your U2 and its C6.

[Groovenut]

..... because many opamps do stupid things when you try to drive a capacitive load. ....

this also applies to your U2 and its C6.

Is C6 even needed? U2 is just a bias buffer so IMO, as long as there's a cap on the divider at the input you shouldn't need C6.

R5 will still need to be there but not as a pot, as a load resistor. 10k-100k works well with most opamps, then adjust C1 for the F3 knee you desire.

[suncrush]

TAKE 4



Left out a 0 on R6, which should be 10K.

R1 and R2's values shouldn't matter, as long as they're equal, and making them large will reduce power consumption, so 470K would be fine, correct?

How do I figure out how big to make the resistor between the U2's output and U1's non-inverting input?

[suncrush]

..... because many opamps do stupid things when you try to drive a capacitive load. ....

this also applies to your U2 and its C6.

Is C6 even needed? U2 is just a bias buffer so IMO, as long as there's a cap on the divider at the input you shouldn't need C6.

R5 will still need to be there but not as a pot, as a load resistor. 10k-100k works well with most opamps, then adjust C1 for the F3 knee you desire.

I googled F3 knee, and got videos of motorcycles......  What do you mean?