Purpose of opamp follower / buffer in power supply

[erikor]

Total noob here working on my first guitar pedal (basically the Box Of Hall reverb). For the power supply circuit, I have copied the circuit found here (my version is attached) which provides both regulated 5V supply and the split power supply required by the Accutronics BTDR-2 reverb brick.

If I understand correctly, the opamp follower (aka buffer) is there to avoid impedance mismatch between the power supply and the reverb circuit. The problem is that I don't really understand what those words mean .

What would happen if the buffer was not there? What is impedance mismatch in simple, small words?

Secondly, what are the three capacitors doing? Just filtering noise? And why is C3 a different value?

Thank you for any shoves towards understanding you can provide!

[EBK]

The voltage divider that gives you your reference voltage will really only give you V/2 if no current flows out of the middle of it.  Otherwise, the voltage will be less than V/2 or fluctuate as the load varies.  Sometimes, the fluctuation of this node can cause signals to undesirably bleed through to other parts of the circuit.  The buffer essentially stabilizes the V/2 voltage under a varying load, or keeps it from sagging under a heavier load. 

Someone will soon tell you either that it is not as simple as I described, or that it is simpler than I described.  They will also be correct. 

Edit: I forgot to say welcome to the forum! Welcome!

[erikor]

Well I hope no one corrects you because this makes perfect sense to me! 

Thanks!

[highwater]

Secondly, what are the three capacitors doing? Just filtering noise? And why is C3 a different value?

Yes, filtering noise. When used that way, the capacitors only need to be "big enough"... too big, and the worst that'll happen is that they cost more, take-up more space.

Someone will soon tell you either that it is not as simple as I described, or that it is simpler than I described.  They will also be correct. 

Sometimes those capacitors are enough by themselves, and you can leave-out the opamp -- that's how it can be simpler than EBK describes. Sometimes, you want to know *exactly* how simple it can be -- that's when it can be MUCH less simple than EBK describes.

Anyhoo, welcome to the forum! And don't forget to let us know how your build works-out... don't leave us hanging just because it works perfectly straight-away (I can only speak for myself here, but mine almost never do).

[Kipper4]

And why is C3 a different value?

R1 C3 form an RC low pass filter and the values would be designed to filter out undesirable freqauncies.

[EBK]

And why is C3 a different value?

R1 C3 form an RC low pass filter and the values would be designed to filter out undesirable freqauncies.

Another way to look at it is that C3 is arbitrarily large.  Without the buffer, C3 would help stabilize the reference voltage some.
With the buffer, I'm not even sure C3 contributes much of anything (I'd probably still include it though).

[duck_arse]

hello a welcome also. not to muddy the waters, but 100uF on the output of a regulator is hardly ever the correct value. by a factor of about 100.

[EBK]

hello a welcome also. not to muddy the waters, but 100uF on the output of a regulator is hardly ever the correct value. by a factor of about 100.

I typically use 10μF.  Am I off by a factor of 10 or 1000?

Edit: I just looked at a datasheet. I'm off by a factor of 100.  100nF ceramic is recommended.  Guess I will save my electrolytic caps for other things. 

[Ben N]

Guess I will save my electrolytic caps for other things.

And your space, too.

[EBK]

What is impedance mismatch in simple, small words?

I'm going to attempt to answer this in a practical way that may bother some people (apologies in advance).
Impedance "matching" is something we don't generally need to be concerned with in pedal circuits.  Instead, think of impedances as input/output, high/low.  A circuit block acts as a good neighbor if it has a high input impedance and a low output impedance.  What is a good neighbor?  It doesn't ask too much of the circuit block that feeds its input, and it is generous and flexible with its output. A buffer can turn a bad or mediocre neighbor into a good one. 

[erikor]

Duck_arse, I found that circuit on the internet. It must be right. 

I will join EBK and drop it into nanofarad range.

[PRR]

> What would happen if the buffer was not there?

Then you have a 4-pack opamp chip with only 3 opamps used. You have to do "something" with the leftover opamp or it may run wild, squeal, disturb the audio. Using it as Vref buffer is a zero-cost answer.

Whether a Vref needs buffering, or just a big cap, or nothing at all, is a system design question. If heavy loads hang on the Vref, Vref will shift or wander. In Box Of Hall, the 25k Reverb pot full-down end can kick reverb signal into Vref. If not swamped-out somehow, that leaks into IC1c and IC1d and contaminates those signals. One is unintended regeneration, the other is reverb in the output when no reverb is wanted.

A big cap on Vref could swamp these effects for midrange. If you want very-very low leakage to deep bass, it may have to be a really large cap. As big as 50uFd. Which is what this guy already has just to swamp the 5k divider. The "advantage" of the buffer is even less cross-leak under 100Hz.

Oh, you want "some" cap on the divider. Otherwise it feeds half the crap on the "9V" into the heart of the circuit. 10u will greatly reduce 50-120hz crap. 47u is generous but may not matter to size or cost; maybe the Designer buys 47u like candy.

[erikor]

> What would happen if the buffer was not there?

Then you have a 4-pack opamp chip with only 3 opamps used. You have to do "something" with the leftover opamp or it may run wild, squeal, disturb the audio. Using it as Vref buffer is a zero-cost answer.

Oof didn't even think about that. As it happens, the rest of the reverb circuit leaves one opamp unused. I better do something about that.

Thanks!

[erikor]

Thank you SO much to all for these very informative responses. I will post back with results of the build!

[Ben N]

maybe the Designer buys 47u like candy.

If you salvage CFLs, you will have plenty of these, which, in my experience, test fine, but they are portly, being rated for higher voltages.

[erikor]

So I have gathered all my parts and am about to build. But while reading about opamps and studying my circuit above, I notice I have drawn the voltage input connected to the inverting input of the opamp instead of the non-inverting input (which is actually not the way this power supply circuit is shown on The Internet). But after reading a bit more and even simulating the circuit (on circuitlab.com), it seems it does not matter.

Does it matter?

Thanks again!

[EBK]

Yes, it matters.  Flip your circuit around to match the others you've seen.  Positive feedback makes an amp oscillate or saturate (possibly both), not act as a buffer.

[erikor]

Well after swapping around the power supply op amp pins and swapping around the LED pins which I also had backwards in my schematic (not shown above), it works! Sounds pretty good even.

A couple issues to work on for the future:

1. There is a switch pop (obvious if nothing is being played, imperceptible mid-riff). Looks like Mr. Black has some good information about that here: https://www.mrblackpedals.com/blogs/straight-jive/6629778-what-really-causes-switch-pop

2. There is a bit of a hum when nothing is playing.

A BIG thank you to all of you for your help on this first build!

Baby picture attached!

[antonis]

Looks like Mr. Black has some good information about that here: https://www.mrblackpedals.com/blogs/straight-jive/6629778-what-really-causes-switch-pop

IMHO, Mr. Black expounds a very "personal" stance of popping..

[bluelagoon]

What is the correct placement for the Filter cap on a Vref Voltage Follower?
I see they are not all placed same in every circuit. Take for instance the Angry Charles circuit from the PedalPCB.com pdf,
it has the cap placed after the op amp. See attached image.

This is contrary to the circuit from the opening post in this thread.

And then I read from a contributor on the electronics stackexchange that
"Op-amps generally can't handle excessive capacitance on their outputs"

That doesnt say much then for the Angry Charles Vref voltage follower if that case holds true.


https://electronics.stackexchange.com/questions/487659/voltage-reference-with-op-amp-buffer-and-additional-filtering

So which actually is it, or does it even matter?

I had in mind to do a vref follower similar to the angry charles , but using only a 10uF cap rather than the 100uF on the A. Charles schematic.