Adding stereo-in to the Boss Dimension C (DC-2) circuit

[aion]

I am DIYing the venerable and formidable DC-2, and that unused half of the compander on the input has successfully tempted me into adding a second input stage. With two completely separate delay lines and signal paths, it's a perfect candidate for stereo-in-stereo-out (and I believe the original Roland Dimension D rack unit was in fact full-stereo).

Here is the original DC-2 schematic for reference:
http://www.freeinfosociety.com/electronics/schematics/audio/pictures/bossdc2.jpg (JPG)

I love the way they implement the stereo output, with the presence of the jack in output "B" closing a FET switch that would otherwise mix the signal down to mono and send everything out through "A". Brilliant and very usable. I am trying to do the same thing on the input, in such a way that they are fully independent from each other (so technically it would even allow stereo-in-mono-out if the jacks were connected that way).

Here is the schematic I've come up with - basically duplicating the input section, adding the input jack "sensor", and moving to the classic discrete flip-flop since the BA634 IC is impossible to find.
https://dl.dropboxusercontent.com/u/10003621/dc2_stereo_sch.pdf

I haven't double-checked all the connections, so fair warning that there may be some errors here and there, and it needs to be cleaned up a bit for readability. I'm mainly looking for feedback on the implementation of the FET switches.

1. Quickie: Over on the upper-right, do R113 and R114 need to be there? I just copied what was done with R90 and R91 a little down and to the left, but I'm not sure under what circumstances the FET switch would need to have these resistors to 1/2V - none of the flip-flop FETs have them. What's the theory here?

2. The bigger question: I'm concerned about noise injection from still allowing the "B" input to feed into the circuit with nothing connected to the jack. Is this worth worrying about? My understanding is that the jack-sensing implementation can really only "close" gates and not open them (at least without some added complexity). Ideally we'd be able to cut off the "B" signal path at the two points where the FETs mix the channels together (near Q14 and Q15), but I don't know if this is solving a real problem or not. It's a different type of problem than the output jack sensor, though, since that one is going out of the circuit rather than coming into it, so added noise is not a factor.

If it's something that needs to be dealt with, I figure there are two ways to handle it:

- Use a switched jack for the "B" input wired in such a way that it grounds the input of the circuit when nothing is inserted. The ground "signal" still runs through a bunch of circuitry after that point, though, so it's susceptible to the inherent noise of any of the active devices along the way.

- Take it one step further and also using that switched ground connection to drive a second jack sensor that is the opposite polarity, opening the gates of FETs that would be added next to the existing Q14 and Q15 (effectively making SPDT switches instead of the current SPSTs).

What's the best approach here?

[Mark Hammer]

You'll want to take a look at this.  http://www.birthofasynth.com/Scott_Stites/Pages/dimc_main.html

[aion]

Yep, been all over that and read through the Electro Music thread where Fromel did a DIY project with Scott back in like '07. Superb research by those two. He uses a physical toggle switch for the stereo setup though which I was hoping to avoid.