For the Carbon I am building, the user would like to be able to bypass the volume pot, but would like it switchable for when the source doesn't have a volume control. I had initially told him no problem, this will be easy. The more I thought about it, I realized it isn't quite as easy as I had surmised. Simply shorting the input to the output (wiper) would still result in potentially bad things depending upon where the pot was in its rotation.

Since I had already done the pot adapter boards for the gb awhile ago, I started with one of those for pcb layout (the pot I am going to use is a TKD CP2500 4 channel).

My initial thought was to take the pot completely out of the circuit by switching both the input and the output (wiper) out? Or would it be sufficient to only switch the output (wiper) out? I had decided to use a relay approach since switching both out would require an 8PDT switch which would be a nightmare to wire up, and the cheapest one that Mouser stocks is $600.

I had a similar quandary in the past in a tube amp trying to bypass the "evil(?)" Alps Blue velvet.  Used the amp with the pot just bypassed completely.  Later, Alps was changed to better-reputed TKD for sources that needed volume control.  These days, I mostly use the TKD with volume open all the way to "effectively" bypass it. 

To be honest, I wouldn't want to bet any amount of money I could consistently identify sound quality differences in blind testing, in any mode above. 😥 

Thinking about this, just bypassing the wiper may work the best. The input would see the same resistance whether the pot was switched in or not (pot R || input impedance). Plus I could do balanced with a 4PDT (or 2 dual relays). Let me know what you think.