[mugdecoffee]

I have an Opus DAC which currently only has the USB input module though I'm planning on adding the SPDIF reciever. Rather than spending lots of money and using lots of case space (my case is small) on the multiplexer module, I figured I could just use a simple switch instead.

First, on the multiplexer board, every input is transformer-coupled but the input to the SPDIF reciever is not. What are the benefits of transformer coupling? I feel like I've read somewhere that they eliminate ground loops...

Secondly, if I choose not to isolate the different sources, is it ok to connect all the grounds for the input signals and only switch the data line?

Finally, is there an easy way to add an LED source indicator without just using a switch with an additional pole?

 

[JamesL]

Can't you just use the Y1 board for spdif input?

Small footprint, simpler power supply, a transformer coupled coax input, spdif receiver and mux, i2s output to your OPUS.

I could be completely off though, but I think it should work. E-config.. leave out analog section, hard-wire coax inputs.

 

[linuxworks]

Quote:

Originally Posted by mugdecoffee /img/forum/go_quote.gif First, on the multiplexer board, every input is transformer-coupled but the input to the SPDIF reciever is not. What are the benefits of transformer coupling? I feel like I've read somewhere that they eliminate ground loops...

its good practice to isolate the spdif coax line on BOTH the ground and center pin wires.

here is a very old schematic that I often refer back to as a design guide:

it shows, pretty clearly, the protocol used for wiring input pulse trafos and output, as well.

input is cap blocked but output is not. you want low-Z 'drive' on the source but you want only the ac components on the receive side (so the blocking cap is there).


Quote:

Secondly, if I choose not to isolate the different sources, is it ok to connect all the grounds for the input signals and only switch the data line?

its better to follow that schematic's idea. AFTER the transformer, you can ground one end, for the circuit. most circuits are not balanced, internally, and so the trafo converts that (to TTL level, in my case). but once the signal leaves the box, no spdif wire should be 'common grounded'.

Quote:

Finally, is there an easy way to add an LED source indicator without just using a switch with an additional pole?

are you going to use actual mech. switches or use logic gates to control the spdif?

I'm putting together a low-end design for an spdif switch - maybe you can wait until I have the whole thing ready to publish?

 

[Beefy]

Quote:

Originally Posted by mugdecoffee /img/forum/go_quote.gif I have an Opus DAC which currently only has the USB input module though I'm planning on adding the SPDIF reciever. Rather than spending lots of money and using lots of case space (my case is small) on the multiplexer module, I figured I could just use a simple switch instead.

You seem to be under the impression that you need both the receiver and the MUX?

In reality, you only need one or the other. The MUX replicates all key functions of the SPDIF receiver, with the benefit of switching, transformer isolation etc.

 

[mugdecoffee]

Quote:

Originally Posted by Beefy /img/forum/go_quote.gif You seem to be under the impression that you need both the receiver and the MUX? In reality, you only need one or the other. The MUX replicates all key functions of the SPDIF receiver, with the benefit of switching, transformer isolation etc.

I forgot the MUX outputted I2S

That would make a lot of sense. The original reason I didn't want to use the MUX was because of the cost of pulse transformers which cost almost as much as parts for the whole reciever board.

Linuxworks: I've actually been following all of your work for a while and getting ideas

On the other hand, I don't need a remove or microcontroller support so I was hoping to just use a mechanical switch. What I'd like to ideally do is connect the spdif wires to a mechanical switch and connect that output to the input of the receiver board which doesn't have a pulse transformer (cheaper). I'm pretty sure this would work; its analogous to unplugging a source and replacing it with another. If I did this, would I be able to switch just the data lines and tied the grounds together or no?

 

[linuxworks]

one reason people like to avoid mech switches for data lines is that they can be noisy (mech noise) and high bounce rate on a switch can interfere with high data rates (like digital audio).

its really a better idea to logic-switch the data flow and maybe have a switch front-end that.

all you really need is a simple chip to gate which of the inputs goes to the output. THAT can be a switch but I would not really use a mech switch for the data-passing part of it.

see if this article is helpful to you:

Remote Controlled AV Switch with S-Video, Composite Video, and Audio

and this:

web-ee.com/images/VSwitch_Connectors.gif

that kind of gave me the general idea of using a 4051 or 4052 chip for a data-passing switch.

he used that chip for pure analog audio. I tried it with spdif digital. works fine! very cheap and easy to get (4051).

you then select which input line you want via a 2 or 3 bit (A B) address. if you use an 8-input switch chip, then you can cheap-out and avoid binary and just take the 1's lines (line 1, 2, 4) and have an 'easy to select' 3 port switch and not have to do binary stuff at all.

you can also front-end that binary address stuff with a 'priority encoder' which will take 1 of 'n' lines and output the binary code that it 'saw' the highest line #.

finally, add in a latch so you can press a button and have it 'stick' to that 1-of-n line thing and that will complete the whole design.

the 4051 style chip, while unorthodox for spdif, DOES work pretty well!

 

[linuxworks]

link to my other thread that also deals with this subject:

http://www.head-fi.org/forums/f6/ic-...itcher-398737/