How does the output stage of the Dark Rift Delay work?

[niektb]

I've been trying to wrap my head around the schematic of the Dark Rift Delay: https://www.pedalpcb.com/docs/DarkRiftDelay.pdf
But I can't figure out what's going on at the output! I would've expected a non-inverting summing topology but that is not what this is... Can anybody explain to me how this works?

[antonis]

It's a differential topology..



When V1 & V2 are 180^o out of phase, it can work as absolute values "totalizer"..

[niektb]

Whoops, I was staring so much at summing topologies, I forgot to look at others 

So basically what you're saying is that they assumed that the signal out of the PT2399 was out-of-phase and at the op-amp this signal (at the minus pin) is inverted?

[anotherjim]

The wet signal is variably delayed - you don't know what the phase shift might be and the delay time is always greater than the signal waveform period, so mixing them out of phase is irrelevant.

In this design, the input is treated to a non-inverting bufferstage, sending the dry to the output also non-inverting means the pedal doesn't invert the signal polarity.

Couple of errors maybe in the designs PDF.
Unused PT2399 amplifier pins 13 & 14 should be shorted together by a link.
Clipping diodes on the repeats D2 & D3 should be opposing each other - not the same way around?

[niektb]

So you're saying that I could as well replace the differential stage with a non-inverting summing stage?

[anotherjim]

You could mix that way but why? It doesn't matter if the sound out of the delay chip gets inverted as it is. And looking a different way, some popular designs on the same theme use inverting mixing and so the clean signal also gets inverted on the way through. And some others use passive resistor output mixing and invert the clean signal with an inverting input buffer so clean goes through inverted.
You might be thinking that the non-inverted summing is somehow better/purer - it isn't because there is some interaction via the summing resistors that reflects back to the sources. The inverted summing mixer puts a virtual ground on the junction to the inverting amp input which prevents that interaction.

The only "problem" with the mixing in the Dark Rift is that the non-inverting feed gets the Rf/Rin gain + 1 while the inverting input only gets Rf/Rin gain, but that's easily corrected by adjusting levels of the input signals.

[ElectricDruid]

+1 agree with what the others have said.

It's a differential op-amp, being used as a mixer. The Diff Amp is a nice tool to have in the toolbox, 'cos there are times when it does (neatly) stuff that would take two op-amps otherwise.

Here, like Jim said, it's being used to avoid inverting the dry signal, so that (as far as is possible with a delay) the output is in phase with the input. The fact that the delayed signal is inverted doesn't matter because it's so far delayed in terms of phase anyway, you've no idea whether it'll be in or out.