Digital Audio Input Selector ?????

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Look here to see how I solved the problem using gates:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=12545

Which solution is best depends on what comes after the input selector. If it's a ASRC or if you use reclocking in your DAC, it doesn't matter that much to pass the signal through another couple of gates. One problem with the Schmitt trigger input of the 74HCU04 is that it reflects crap back on the S/PDIF line. To solve that problem, you can use a fast comparator at the input.

A simple and yet very good solution is to use relais to switch the inputs. Make sure all inputs are terminated properly, even if not selected. To avoid grounding problems, it might be best to use transformers on all inputs (if you don't use optical connectors, that is)
 
I also thought of the idea do use relays at all inputs.
I will be using optical inputs but at the output I want to connect it to the DAC input which has one optical input and one coaxial input.

I will disconnect the optical input and connect internaly the output of the selector to the coax input of hte DAC.
Should this connection be done with Cinch or direct with twisted pair?

What do you think of the relay solution. What would be the best way to do this?
 
Originally posted by promitheus
I will be using optical inputs but at the output I want to connect it to the DAC input which has one optical input and one coaxial input. I will disconnect the optical input and connect internaly the output of the selector to the coax input of hte DAC.

Sorry, I can't make sense of that. Can you clarify your setup? What will you connect to deliver the digital signal?
 
ok sorry about that.

my DAC has one input only.
I can choose either Coaxial or Optical. I have the DAC 2000 from the Elektor magazine.

I have at least 3 devices with optical output. So I need an input selector.

I was thinking of making a 4 optical input selector with relays or digital ICs
The output of the selector I want should be coaxial SPDIF level so I can connect it to my DAC using the coax option. This connection will be made inside the box I will use. Do you think this connection between PCBs should be coaxial or can I use normal twisted pair cable?
 
I am building a stereo system, and asked at the Nuts and Volts forum about how to make a selector switch, and was kindly given this schematic. I haven't tested it yet, but it allows 4 inputs, and an LED to indicate what input is selected. Hope this helps.

I have a higher quality picture, but I couldn't attach it here. If you want it, I can e-mail it to you.

-Mike
 
If you have optical inputs only, then it would be better to use gates instead of relays.

The 74HC125 will do the job. It's a tri-state buffer. I connected all the outputs together as the "output bus" and used a 74HC42 BCD-to-decimal decoder to select which of the tri-state buffers should drive the bus. The Toslink receivers will deliver TTL level and I suggest driving the input of your receiver through a small capacitor.

You build the Elektor DAC2000? Don't solder in L2, IC2, C9, C8, R2, JP1. You don't need them for your setup. Alternatively, remove JP1.

Now, if the connection between input selector and DAC is longer than, say, 10cm, I'd recommend using a coaxial connection. You need to use 75 Ohm cable and put a 75 Ohm resistor between input selector output and cable (to match impedance).

If the connection is short, use twisted pair, but remove R1.
 
Sorry, i forgot to attach it. here it is.
 

Attachments

  • audio switch.gif
    audio switch.gif
    90.1 KB · Views: 3,580
Originally posted by soundNERD
I haven't tested it yet, but it allows 4 inputs, and an LED to indicate what input is selected. Hope this helps.

Looks quite complicated to me. For four inputs, you need only one 74HC125. The BCD-to-decimal decoder is optional (to save port lines for use with a µC), you could also switch the buffers using a rotary switch.
 
Yeah I know what you mean.

Actualy I am using the upsampling converter which has a CS8420 but the input looks exactly like the original DAC does. It has a TORX173 which I won´t be needing and a coax input terminated with 75 ohm. and of course the jumper to select input.

Actually the cable can be less then 5 cm if I can make a pcb to be placed next to the upsampling converter.

I will check the ICs you gave and see what I can think of.

Is there something I should be carefull of when making a pcb?
The frequencies are high till 5MHz, right?
 
Originally posted by promitheus
Is there something I should be carefull of when making a pcb?
The frequencies are high till 5MHz, right?

Proper PS decoupling, short signal traces, a ground plane. Not too difficult.

Can't think of why you would need an inverter (rather two) between Toslink and CS8420 input. Have a look at the upsampler board and its optical input. We want exactly that, only the selector between it. The buffers are non-inverting, btw.
 
Just realized that I used inverters, too. One between Toslink and 74HC125 and another one between 74HC125 and output. Don't remember why, probably only to buffer the signal after the 74HC125. To restore phase, I added another inverter before the 74HC125. I think you don't need them. I had the 74HCU04 on board anyway (as "amplifier" for the coax inputs) and that way I could use all six gates.

If you use a rotary switch, you don't even need the 74HC42 :cool:
 
i have here an elektor circuit that uses a 74HC4051 multiplexer. It has one inverter in the input and then one in the output. I guess they use two to keep correct phase and as buffers for the inpts and outputs of the multiplexer.

If the 74HC125 is a good buffer then there is I think no need for the inverters.

With a rotary switch I think the best would be to use 100K resistors tighing the output enable inputs to high so the output impendance is high. The rotary switch should only have ground potential connecting the wanted resistor to ground giving that output enable input a low.
 
He did :(

Jocko is right, of course. Keep on posting crap in the early morning hours.

Don't care about the phase of a biphase-mark encoded signal. Could have thought of that. Why is everyone using those inverters anyway...

And the 74HCU04 has no Schmitt trigger input. The CS8412/14 has. Mixed that up, too.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.