Op-amp booster circuit. Critique/advice sought

[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.

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

Sorry, the frequency at which the amplitude of the signal is down 3dB in a filter

[suncrush]

If C1 is large, then that frequency should be outside of guitar range, right?

[Groovenut]

If C1 is large, then that frequency should be outside of guitar range, right?

Yes, you can use this formula to find the F3 knee

F= 1 / (2PiRC)

[suncrush]

If C1 is large, then that frequency should be outside of guitar range, right?

Yes, you can use this formula to find the F3 knee

F= 1 / (2PiRC)

Yeah, ok, that one's easy.  With a 10K resistor, it works out to something in the ballpark of a 0.5uF cap, but going to a 100K resistor lets you use a 47 nF.

[Groovenut]

If C1 is large, then that frequency should be outside of guitar range, right?

Yes, you can use this formula to find the F3 knee

F= 1 / (2PiRC)

Yeah, ok, that one's easy.  With a 10K resistor, it works out to something in the ballpark of a 0.5uF cap, but going to a 100K resistor lets you use a 47 nF.

You might also wants to make the worst case scenario assumption that the circuit might be plugged into something that would bring the load down even further than your circuit load resistor and compensate so the cap value is large enough to pass the desired signal frequencies.

[suncrush]

Roger.


Ok, the only thing I think I need to figure out before I go put this on vero is how big a resistor to use to connect the output of U2 and the + input of U1.

[antonis]

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.

Typo...

[Groovenut]

Roger.


Ok, the only thing I think I need to figure out before I go put this on vero is how big a resistor to use to connect the output of U2 and the + input of U1.

For standard electric guitar and bass , 1M should be fine.

What is the purpose of the R and C to ground from the output of U2?

[suncrush]

Roger.


Ok, the only thing I think I need to figure out before I go put this on vero is how big a resistor to use to connect the output of U2 and the + input of U1.

For standard electric guitar and bass , 1M should be fine.

What is the purpose of the R and C to ground from the output of U2?

The MIT notes include the cap to ground in the voltage follower schematic.  duck_arse suggested the R to avoid issues of the op-amp driving a capacitive load.

[ashcat_lt]

The MIT notes include the cap to ground in the voltage follower schematic.  duck_arse suggested the R to avoid issues of the op-amp driving a capacitive load.

That cap is meant to filter any noise that might be riding that bias signal before it gets injected into the opamp input.  I suggested the resistor, but I wanted it to be directly after the opamp.  Before the junction off the cap and the bias resistor.

Groovenut suggested it might work just as well with that cap-to-ground on the non-inverting input of U2, and it probably would, but this way we're also shorting any (probably negligible) noise that the opamp itself might add.

[Groovenut]

The MIT notes include the cap to ground in the voltage follower schematic.  duck_arse suggested the R to avoid issues of the op-amp driving a capacitive load.

That cap is meant to filter any noise that might be riding that bias signal before it gets injected into the opamp input.  I suggested the resistor, but I wanted it to be directly after the opamp.  Before the junction off the cap and the bias resistor.

Groovenut suggested it might work just as well with that cap-to-ground on the non-inverting input of U2, and it probably would, but this way we're also shorting any (probably negligible) noise that the opamp itself might add.

Ah I see, but the way it's currently drawn it's not a filter for the bias rail. To function as intended the resistor would need to be the first thing at the U2 output followed by the cap to ground. The bias rail would then need to be taken from the junction of the two.

The problem I see is unless you dont mind using a huge cap value, you may be injecting shot noise with the series resistor. Say we wanted to use a 100uF shunt cap, the resistor would need to be about 3k or so to get filtering down to .5Hz. I would think there may be more noise from the resistor than the low impedance bias converter opamp. Granted the current flow is so minuscule this may all be academic 

Thoughts?

[Gus]

Do a search here for booster, opamp booster etc.  There have been a number of threads over the years.

[ashcat_lt]

Ah I see, but the way it's currently drawn it's not a filter for the bias rail. To function as intended the resistor would need to be the first thing at the U2 output followed by the cap to ground. The bias rail would then need to be taken from the junction of the two.

I mentioned moving the resistor, too.  As drawn, it's a shelving high-cut, so it kind of works, but not as well as it could.  I'm not convinced we need to go all the way down to 0.5Hz with this filter, but I'm also not afraid of a 100uf cap.   

...course then that raises the point that both this cap and the actual rail filter at C3 should probably have smaller film or ceramics parallel to them to make up for how electrolitics fail at higher frequencies.

..course C3 is only necessary if the 9V is not actually a battery.  Because it's there, I've been assuming that there will at least be a DC jack to allow external power which would want to be filtered.

...bnut then we should probably put some smallish resistance in series with the DC input (before C3) to limit in-rush current and help set the cutoff of that rail filter. 

...and then I'd want to put a diode in series with that for reverse polarity protection...

[suncrush]

Pedal will NOT be battery powered. I run everything off a wall wart.

[suncrush]

Ok, I can't figure out C6.  It's there because the MIT dic shows it there, but it also shows it going to ground. If I'm hooking the inverting input up to the output of U2, it seems that cap would be counterproductive. Since the cap would block DC signal, wouldn't I want to take it out so that I pass the same voltage to both inputs?

[ashcat_lt]

Ok, I can't figure out C6.  It's there because the MIT dic shows it there, but it also shows it going to ground. If I'm hooking the inverting input up to the output of U2, it seems that cap would be counterproductive. Since the cap would block DC signal, wouldn't I want to take it out so that I pass the same voltage to both inputs?

Honestly doesn't matter either way.  You don't want DC gain there, so you need either no DC path or no DC difference or both.  In this case you've got both.  C6 is kind of redundant, but as long as it's big enough won't hurt anything.

...but you still need to move R7, and I think that C6 wants to be on the other side of that along with everything else.  So, what you want is for the output of U2 to connect directly to its own inverting input and one end of R7.  The other end of R7 connects to C6, and the R going back to U1 (can't tell if it's a 6 or 8 or what, it kind of looks like there are 3 resistors labeled R8 on my screen), and one end of C8.  The other end of C8 goes to ground.

[antonis]

Pedal will NOT be battery powered. I run everything off a wall wart.

Then - if your project hasn't educational purpose - delete IC2, place a resistor of 100R in series between +9V and C3, oversize C3 to 470μF, make R1 & R2 10k and make C4 47μF...
(perhaps add also a 100nF ceramic disk for PS decoupling..)

[suncrush]

Good call, antonis.  Let's simplify!




A couple things.

Is the reverse polarity diode in the right place? Can I use a germanium diode for that to minimize potential loss?

The schematic I based this on has C4 going to ground.  Earlier people suggested hooking it to the bias voltage.  Other people suggested it is irrelevant.  I'm going with the irrelevant crowd and keeping it simple.  Let me know if I've done something horribly wrong.

[Cozybuilder]

Suncrush-
You might read up on low pass filters & how to calculate the frequency:

http://www.learningaboutelectronics.com/Articles/Low-pass-filter-calculator.php

[antonis]

The 100R resistorr should be is series with the rest of the circuit (between +9V and C3's positive plate..) but you've also have to consider for a voltage drop depenting on circuit's current consumption..

D1 should be the first item after +9V..
You may use a Schottcky diode for less voltage drop or a reversed 1N40XX + 100R(*) between +9V & GND for no voltage loss..
(*) optional..

You've deleted the resistor connecting Vref with + Input so the Input impedance is low enough (R3//R4)..
Bypass the connection between C2 and R3/R4...

Newahhhhhhhhh.... 
Give me a minute to fix your scheme..