Drving diodes with a bigger cap and then cutting some bass

[John Lyons]

Picture this.
(Diode to ground clipper) .1 cap feeding symetrical diodes to ground then into a .02 cap to volume pot, out.
Is the .02 cap in series with the .1 just dividing capacitance or is the larger cap (.1) driving the diodes
harder (more being clipped) and then the .02 dropping a little bass and then out?
What I'm trying to do is drive the diodes with a full freq signal (for the most part) and then drop a little
bass by using a smaller output cap. Is this doing what I hope it is?

[R.G.]

Picture this.
(Diode to ground clipper) .1 cap feeding symetrical diodes to ground then into a .02 cap to volume pot, out.
Is the .02 cap in series with the .1 just dividing capacitance or is the larger cap (.1) driving the diodes
harder (more being clipped) and then the .02 dropping a little bass and then out?

Well, the 0.1uF cap probably has something behind it driving it. That has some source impedance, yes? It may be linear or nonlinear, but it's there. An opamp with no resistor may have an output impedance of 10 ohms or lower. A transistor collector probably has about 1K to 100K (the value of the collector resistor). An emitter follower has an odd output impedance. So that part intrudes.

The diodes are both nonlinear resistors. They are (kind of) open circuits til the voltage hits knee, then it may change over 50mV down to a few ohms. The current into them is no limited by the capacitor much at audio frequencies, but is instead limited by the driving impedance. The output 0.02uF is loaded by the volume pot, so the 0.02uF in series with the volume pot appears in parallel with the diodes for determining what current gets put through the 0.1uF cap.

There is an unfortunate dichotomy here between the frequency domain response and the time domain/transient response. In time domain, the action of the driving cap changes when the diodes start to conduct, and the cap starts filling from the low impedance while the diodes conduct. It stop this when the currently conducting (!?)diode stops, holds that charge and starts discharging it into the opposite diode when the driving voltage makes that one conduct. What this does in the frequency domain is ... well, that's how clipping harmonics are generated.

What I'm trying to do is drive the diodes with a full freq signal (for the most part) and then drop a little
bass by using a smaller output cap. Is this doing what I hope it is?

Kind of. The value of the volume pot gets into both loading down the 0.1uF driving cap, and in the bass rolloff of the 0.02.

This kind of thing is really messy if you don't have a good handle on the driving impedance before the 0.1uF. That's why a lot of pedals put a 10K in series with the diodes.

[John Lyons]

Thanks for the rundown RG.

There is a transistor with a 3k3 collector resisance driving the .1 into the diodes.
Then a .02 output cap to a 100k volume pot.
So are you say9ing that the .1 doesn't drive the diodes harder (more low end signal)
than if I just used a .02 cap into the diodes?

John

[R.G.]

There is a transistor with a 3k3 collector resisance driving the .1 into the diodes.

That will drive the 0.1uF into the diodes the same amount as an opamp with a 3.3K resistor in series with its output.

This is a pretty straightforward real-world swap of a Norton-equivalent circuit (transistor collector) for a Thevenin-equivalent circuit (opamp output). Smart guys, Thevenin and Norton.

Then a .02 output cap to a 100k volume pot.

That looks like a 100K resistor at frequencies above F = 1/(2*pi*.02uF*100K) = 79Hz as long as the loading on the pot wiper is large compared to the 100K pot value. If the loading on the pot wiper is significant compared to the 100k, then the loading on the feed to the pot wiper is significant and changes the bass rolloff up in frequency as the loading increases - that is, as the volume pot is turned up. The 0.02 *might* choke off some bass if it were loaded down more. It's that old RC rolloff thing.

So are you say9ing that the .1 doesn't drive the diodes harder (more low end signal)
than if I just used a .02 cap into the diodes?

Not exactly.

With a resistor in series with the cap - which you have, in effect, in the 3.3K collector resistor - then you can't really say it drives the diodes harder. What it does is lower the frequency where the 3.3K output impedance stops being able to drive the diodes well.

The bigger the resistance in series with the 0.1, the less the 0.1uF matters as changing things when the diodes start/stop conducting.

With a 0.1uF and a 3.3K series resistance, at frequencies above F = 1/(2*pi*R*C) = 482Hz, the resistor is what matters most, not the cap. Below that as frequency goes down the cap's impedance goes up (in frequency domain) and the cap limits what current goes into the diodes. It does this increasingly as the frequency goes down. In the time domain, the cap does that changing from charging to not charging thing I mentioned as the diodes conduct and stop conducting.

Interesting, isn't it? The 0.1uF cap limits bass response more than the 0.02.

If you load down the 0.2 more, perhaps with a 10K across the volume pot or just use a 10K volume pot, the bass cut on the output gets more pronounced.

However, this divide-and-conquer approach does neglect the changes that happen when the diodes are both off. In that case, the diodes are out of circuit, and the two capacitors look like a series combination; but the bigger 0.1uF dominates there.

There are a few principles which guide the understanding of this kind of thing.

1.  Order of magnitude of effect; if something has a 10% or less effect than something else, then you can ignore the smaller contribution with only small errors. Very good for a first look at what's happening, terrible if you have to provide an answer to two decimal places.
2. R-C frequency rolloff, which is the AC version of Ohm's law applied to series elements.
3. Frequency domain versus time domain duality.
4. Nonlinearity effects from diode turn on/off in both domains.

[John Lyons]

Thanks RG once again. This one pretty much clips
all the time as it's a medium gain overdrive.
Also the volume is at about 3/4 or more up all the time.
I'll have to digest your writing...

John