Differential mux approach to rotary input switching

[atavacron]

Has anyone done a differential mux with an encoder and multiple 8:1 x 2 analog switches? Seems like the ADG1407 is the best candidate on dual 15V rails, though TI might have something that matches its specs by now. I’m aware of 16:1 x 2 chips, but haven’t seen anything with as low THD+N for pro line level. [EDIT: TI has nothing in 16:1 x 2 and Analog has one part but it’s for dual 5V.]

This would be for an input select switch; I’m not sure how many sources yet but let’s say 12 to 16 positions, so there’s a transition mid-travel where you need to switch which mux chip you’re controlling. The control is a knob on a console monitor section that selects from a bunch of stereo bus/aux/line returns and feeds to a single trim. Load is trivial - over 100K, non-inverting. I figured it out in strictly physical switching terms and it’s too many decks / impractical - 6P12T kind of stuff if you include solo bus DC.

I have some reading to do, so if there are good links to check out with reference designs, LMK.

 

[atavacron]

I asked something similar here:

Some differential audio mux questions - Q&A - Switches/Multiplexers - EngineerZone

…we’ll see what they say.

 

[FredForssell]

With bipolar 15 VDC supplies, you will be limited to a max signal level of about +22 dBu (15*0.707=10.6 Vrms).

For a monitor controller with sources like the 2-mix stereo output or an external DAC, this probably will not be a problem. But for line level returns and buss outputs, it could be an issue if you don't attenuated the signal before the ADG1407.

 

[atavacron]

Thanks Fred. Supplies could be up to dual 18V, but the general purpose amp in the system is OPA2210 (whose THD takes off after 7V RMS), so I’m generally working with +19dBu SE / +25dBu differential downstream. It’s a bit of a headroom tradeoff for really, really great flexibility in a DC coupled environment. You can do stuff with that chip that is tough with most bipolars and too precision oriented for most FETs…a console full of AD4898/OPA2211 and AD4625/OPA828 would be pretty “cool” but run pretty hot — and would be cost-prohibitive I think. I’m relying on nobody having an ADC that accepts more than the SMPTE max of +24dBu differential, and meters per channel that are red above +16 and stop at +24.

For context, the project is a 12 bus / 8 aux inline console with a workflow that allows for traditional multitrack tape routing or more direct output reliant modern DAW routing. It looks like a early-aughts Cadac and a Neumann N20 had a rackmountable baby. This particular control would accept stereo pairs of bus or aux, and a following switch would select odd or even. A stereo monitor strip would forego the following switch but use the same selection method.

 

[atavacron]

So, I figured out how to make a 16-input differential switch with a 16-position BCD encoder and two ADG1407. That’ll let me bring 8 buses and 8 auxes into any given mono monitor channel. Same tactic, doubled, will work for stereo monitor channels….in that scenario a 10-position encoder will cover all. Grayhill 26 series all around.

 

[Newmarket]

With bipolar 15 VDC supplies, you will be limited to a max signal level of about +22 dBu (15*0.707=10.6 Vrms).

dBu reference voltage is 0.775V