Let’s talk about Clipping

[PBE6]

+1 amptramp!

That sounds fascinating, I will have to try out those clamp circuits. Thanks for the handy list of circuit modules as well.

[thehallofshields]

Hey, I've got a good question.

In Diode->Ground configurations; does Source Impedance affect clipping?

I would think that its purely based off Signal Voltage, but I've noticed that Fuzzes followed by Diode Shunt Clippers get Farty  in a way the something like the Dist+ never seems to. To me; the Bosstone, Percolator and Univox Squarewave all sound much better with the Diodes lifted. Sure, its probably the Bass getting clipped twice, but I've suspected the higher output impedance of BJT's is partially to blame.

[teemuk]

In Diode->Ground configurations; does Source Impedance affect clipping?

Yes. It always does.

[R.G.]

In Diode->Ground configurations; does Source Impedance affect clipping?

Yes. It always does.

... and dramatically.

Artificially adding some resistance in series from a low impedance signal source, like an opamp output, to clipping diodes to ground, causes a noticeable change in distortion. Even better, add some resistance between the signal take-off point at the "top" of the diode pair and the diodes. Putting in a 10K pot used as a variable resistor is very instructive.

[teemuk]

Think about it: If diodes are placed in feedback loop of a gain stage the signal amplitude gets limited to forward voltage of the diodes, this simultaneously also limits diode current to a specific limit. If clipping diodes are instead placed in "shunt" to the gain stage's output the gain stage will try to swing to its usual voltage limits and precedingly the signal gets clipped by the diodes; more current flows through the diodes because amplitude of the signal fed to them isn't limited like in the first scheme.

That's main reason why people refer to these setups as "soft" or "hard" clipping. The lower current flow in feedback clipping scheme simply results to softer clipping.

Now think about adding that source resistance; it's limiting the diode current and the higher the resistance the lower the current and softer the clipping.

Most gain stages also have a practical output current limit they can't exceed, for example, a typical opamp can usually supply no more than 20 mA to a load. A shunt diode without series resistance practically resembles a short circuit so the gain stage will muster all of it's current capability to drive it. Opamps that have current-limited outputs will simply clip hard at decreased headroom, some more sensitive circuits without current limiting may even fail. Due to that it is generally adviseable to introduce at least some source resistance before shunt diodes even if the goal wasn't soft clipping.

[amptramp]

With a Tube Screamer, the diodes are in the feedback path but the stage is non-inverting, so with the diodes conducting fully, you still get unity gain for the incoming signal, just more gain before the clipping starts.  This gives the effect of softening the clipping in contrast to a DOD250 type clipper with diodes to ground after a gain stage.  Diode characteristics dominate in a DOD250 because there is no unity gain path and clipping seems harder.

[thehallofshields]

Most gain stages also have a practical output current limit they can't exceed, for example, a typical opamp can usually supply no more than 20 mA to a load. A shunt diode without series resistance practically resembles a short circuit so the gain stage will muster all of it's current capability to drive it. Opamps that have current-limited outputs will simply clip hard at decreased headroom, some more sensitive circuits without current limiting may even fail. Due to that it is generally adviseable to introduce at least some source resistance before shunt diodes even if the goal wasn't soft clipping.

OK, except that in Diode Shunt configurations we have a DC blocking capacitor so current limits shouldn't really be reached by our AC signal shorting to ground right?

Maybe a more practical question to ask is how Source Impedance affects clipping compared to Load Impedance vs Resistance between Diodes and Ground. Is one going to effect Threshold while another effects the knee?

[anotherjim]

The blocking cap won't limit current much at all -  it will only have significant impedance at low frequency. Since it will have been chosen to pass all frequencies of interest, it will act like a wire to our AC audio signal. So the amp will still hit current limit when the diodes to ground conduct.

Also, if there is no limit resistor - isn't it making the capacitor the primary load and cause the amp to become unstable?

Does anyone know of a commercial pedal that doesn't fit limit resistance before hard clip diodes?

[PBE6]

The Foxx Tone Machine doesn't.

[anotherjim]

Ok, I mean any opamp driven clipping diodes without limit resistor.

The Fox is running the diodes off a transistor stage that has emitter and collector resistors, so it does have limiting resistance, just not provided solely to limit the clipping diode current.

[PRR]

> running the diodes off a transistor stage that has emitter and collector resistors

Emitter resistor has almost no effect; anyway is bypassed for audio.

The naked collector is 200K or more, so the stage output is (as you say) pretty much the 10K collector resistor. Somewhat reduced by tone-network loading. IAC, significant resistance, just not explicit.

[acehobojoe]

yea, I definitely noticed the "harder" clipping with a dod250 circuit. I still don't know where to start with the clamping, but I will get it one day.

[thehallofshields]

Ah yeah. I now see that the Rat uses a 1k series resistor, and the Dist+ and Dod250 use 10k.

So the lower Resistance should result in more current getting dumped through the diodes and the Opamp being worked harder... resulting in a... harder knee? Lower clipping threshold?

[PRR]

> harder knee? Lower clipping threshold?

The shape differences are less than a gnat's toenail.

Going from 100K to 10K to 1K does cause peak level to come up a factor of 1.5, but we *always* have gain adjustments after a clipper and routinely make larger changes than this.

[thehallofshields]

Makes sense. Thanks PRR.

[anotherjim]

If you want to play with it, put in limiting of a 1k in series with a 100k trimmer. Small expense if the changes prove they aren't worth a panel control.