Saucy box - what is this part doing?

[maertz13]

I’ve built it. It’s cool. But that unfortunately didn’t lead to me understanding this hard clipping arrangement. The resistor and cap, what are they doing? I’ve never seen this before.

 

[nooneknows]

At first sight it seems a trick to limit to some frequency band the clipping, but I made some calculation and, even considering the 1uF at the opamp output, you have at max - 3dB at 25 Hz, so I'm waiting for other considerations by much more expert people too

 

[ICTRock]

fairly stout series resistance into LED hard clipping to ground would make them more like attenuators than anything else ... a way to lose some dbs between gain stages without a lot of added compression or clipping and I'm pretty welcome to being wrong on that but that's what it looks like when you factor in those 100K resistors that would be part of the voltage divider

 

[darwin999]

Here's my take. Look at it as a voltage divider, including the 100K to the opamps vs the dynamic voltage-dependent resistance of the LEDs to ground. As noted above, the 10µF is so large that at normal guitar frequencies it just blocks DC. The LEDs are very high resistance until you reach sufficient voltage. So the net result is compression, with the series 6.2k and 10k somewhat adjusting the aggressiveness of the LED's rounding over of the voltage.

 

[giovanni]

I don’t see a voltage divider with the 100k resistor unless I’m missing something obvious? If the 100k resistor was before the other network then you would have a divider, but as is you get a tiny voltage drop across the 100k resistor but not much more?

The RC plus diodes networks do ground most frequencies (the cutoff is around 2-3Hz if I did the math right) but only for signals above the diodes forward voltage (one should also consider the voltage drop across the 6.2K and 10k resistors but that’s probably small). So they should clip the peaks of the signal but probably a bit smoother than hard clippers? I’m not an expert, I would have to think about it a bit more.

 

[darwin999]

@giovanni -

What I meant by the voltage divider (probably poor choice of words) was there are 2 paths, thru the 100k to the opamps or thru the LEDs to ground. The signal will flow thru the path of least resistance. For audible frequencies, (1) low voltage signals pass onto the opamp, while (2) when the when the instantaneous voltage exceeds the threshold voltage of the LEDs, then the signal shunts to ground. There is rounding of the corner there set by the voltage-dependent LED resistance (∂V/∂I) + the series 6.2k or 10k resistor.

And by compression, I meant it compressed the peaks once the LED threshold voltage was approached. It's not an abrupt clip, but is rounded over a small range of voltages which is equivalent to nonlinear compression. You called it clipping, same thing.

Edited - struck through a couple of repeated words...

 

[giovanni]

@giovanni -

What I meant by the voltage divider (probably poor choice of words) was there are 2 paths, thru the 100k to the opamps or thru the LEDs to ground. The signal will flow thru the path of least resistance. For audible frequencies, (1) low voltage signals pass onto the opamp, while (2) when the when the instantaneous voltage exceeds the threshold voltage of the LEDs, then the signal shunts to ground. There is rounding of the corner there set by the voltage-dependent LED resistance (∂V/∂I) + the series 6.2k or 10k resistor.

And by compression, I meant it compressed the peaks once the LED threshold voltage was approached. It's not an abrupt clip, but is rounded over a small range of voltages which is equivalent to nonlinear compression. You called it clipping, same thing.

Ah I see what you mean. I think your explanation makes sense. I kinda wanna run a spice sim of this…

 

[darwin999]

@giovanni -

So I think we agree, although your terminology was probably better.

Share the SPICE sim, if you do it. I'm curious why they clip it twice in a row. Those 2 opamp stages have nominally unity gain at low frequencies, and fall off at higher frequencies 1/2πRC ~ 7kHz. Right now I'm guessing that the 2nd clipper removes any excess high frequencies generated by the 1st round of clipping.

 

[ICTRock]

I think it's important to look at the whole schematic, that said I don't know if this is the original saucy box or the hard clip version. I'd imagine the hard clip version shrinks the series resistance to the diodes to get more clipping from the LEDs in the overdrive side of the circuit

https://effectslayouts.com/wp-content/uploads/2023/12/Spice-Runner-Build-Doc.pdf

it's important to note that the clean path is boosting the signal which would explain the larger resistor in the second attenuator and that the first gain stage has soft clipping using diodes with a lower Vf and only a 100K gain pot so the LEDs aren't going to be doing a ton of clipping

amz has a series of posts about saturation controls that might help

 

[maertz13]

I highlighted the gain side because the clean side is understood and I can technically bypass it.

Yes, I clipped that from the “spice runner” documentation.

I’d also really like to see a spice somulation

 

[ICTRock]

https://www.diystompboxes.com/smfforum/index.php?topic=129390.0

Mark Hammer might save you some simulating

let's consider the king of tone where it is possible to have both the negative feedback diodes and shunting diodes in circuit at the same time. in that scenario, the forward voltage of the negative feedback diodes (2 pair MA856 Vf ~1.5V) is more than double that of the shunting diodes(1 pair 1S1588 Vf ~700mV). you get more compression and harmonic distortion because the shunting diodes will clip before the negative feedback diodes do and you're likely to clip the signal twice. in the saucy box, that's reversed because the negative feedback diodes have less than half the forward voltage of the shunting diodes and the shunting diodes are unlikely to clip at all.