Redundant parts in this circuit?

[John Lyons]

Is the .1 cap and 10K redundant after the diodes?



It seems that I could use a 20K / .047 cap before the
diodes and achieve the same thing.
(caps in series divide, resistors in series add. )

Am I overlooking something?

[joegagan]

i don't know what the caps are doing there, but i would think that the diodes would react differently with a 20k ahead of them instead of the 10k.

[R.G.]

Is the .1 cap and 10K redundant after the diodes?

No, not at all.

It seems that I could use a 20K / .047 cap before the
diodes and achieve the same thing.
(caps in series divide, resistors in series add. )

What's going on is that the first 10K limits currents in the diodes when they conduct to cause clipping. The second 10K is setting the gain of the second opamp. The two capacitors are breaking the DC paths from the first opamp and into the second opamp. All of the parts are needed.

If you lumped R10 into R9, the circuit would never clip because the inverting input node of an opamp is a virtual ground - it only moves by the error voltage, which is signal divide by the open loop gain of the opamp. So the signal across the diodes will never cause them to conduct. In fact, many opamps have back to back diodes across the input pins to protect the inputs. The diodes never conduct. R10 gives the diodes "room" to conduct. Neither capacitor can be removed if the opamps are not biased at the same voltage as the bottoms of the diodes, and the symbols don't look like they are. Leaving out C4 would let IC2 offset the diodes so they conducted unevenly, possibly wildly so. Leaving out C5 makes the second opamp have a DC gain of five for signals big enough to make the diode conduct, and it's gain would get modulated wildy by the diodes conducting. That might be interesting, but it's not what this circuit is intended to do.

The clipping diodes and DC breaking caps make all the difference.

[John Lyons]

Ok, I can see what your are say9ing R.G.
Thanks!

I didn't know about the virtual ground of the inverting input.
I was looking at it as a redundant coupling cap and not taking the op amp into consideration.


John

[Brymus]

Yeah that was my thought to John. A redundant coupling cap.
I learned something here too.
I like reading RGs answers even when they arent my questions.
What circuit is the snippet from ?

[TELEFUNKON]

R8 looks unneccessary to me.

[John Lyons]

From what I know R8 is a fail-safe. Having some resistance on the output of an op amp.
Something about loading down of the the op amp...I'm not sure really....

The snippet is from a Maxon OD-880

[Brymus]

Thanks John

[R.G.]

From what I know R8 is a fail-safe. Having some resistance on the output of an op amp.
Something about loading down of the the op amp...I'm not sure really....

No, you have it right. Some opamps don't like funny loads on their outputs - high capacitance loads and loads which change resistance with drive, like clipping diodes. A little resistance in series with the output keeps the opamp happy and stable. In many cases it's a just-make-sure component, but then that's what just-make-sure parts are all about.

[JKowalski]

Is the reason that people use ground as the clipping reference rather then Vref that the circuit is already biased at only because the Vref is usually a relatively high impedance source (compared to ground) when made with the common voltage divider method? Or is there something else?

[TELEFUNKON]

From what I know R8 is a fail-safe. Having some resistance on the output of an op amp.
Something about loading down of the the op amp...I'm not sure really....

No, you have it right. Some opamps don't like funny loads on their outputs - high capacitance loads and loads which change resistance with drive, like clipping diodes. A little resistance in series with the output keeps the opamp happy and stable. In many cases it's a just-make-sure component, but then that's what just-make-sure parts are all about.

Isn`t R9 and R7 fail-safe enough?

[R.G.]

Isn`t R9 and R7 fail-safe enough?

To keep it working with the world outside the circuit board? Yes.

But some opamps are funny about things like how much capacitance are on their output pins, and whether their load changes dramatically - like driving a capacitor directly into a diode to ground. This situation can, possibly, sometimes, occasionally make the opamp oscillate, either all the time or in bursts at certain points in a waveform. Not all opamps do this, and many fewer of them do it today than back when opamps were high tech.

But a competent EE designing something to be SURE that there were not field problems in 100,000 units when warranty costs could eat up all the profit for the run might want to put in a $0.02 resistor just to eliminate even the possibility. (Notice that the preceding statement could equivalently read "... a responsible and skilled EE would want to be SURE that not a single buyer would get a squirrelly acting pedal and have to go through the frustration and disappointment of his pedal not working right." They're equivalent statements, viewed from two different angles.)

In that sense, it really is "just make sure" and not "fail safe"; those are slightly different meanings. "Just make sure" stops even the possibility of certain failures. "Fail safe" means it may fail, but the consequences of the failure do not propagate and cause more harm than the original failure itself. At least that's how I'd interpret that.