S/P-DIF Input Selector Solution

[tiefbassuebertr]

I am looking arround to a solution for an easy to build SPDIF source select. Certainly here on diyaudiio there are a lot of threads to this topic, but I am not able to find this.
This IC could be a appropriate device therefore:
UA9637 Datasheet pdf - DIFFERENTIALE VERDOPPELUNGLINIE EMPFÄNGER - Texas Instruments

[tiefbassuebertr]

No advices?
I have found a wide range of threads here on diyaudio, but no of them are really helpful. Here some examples:
Digital input selector
Digital Audio Input Selector ?????
Digital source selector?
Digital source switching

In the attachement I have create a simplified schematic with tri-state buffers (line receivers).
Therefore are a lot of integrated circuits available like the follow:

http://www.national.com/ds/DS/DS34C86T.pdf
http://focus.ti.com/lit/ds/symlink/ua9637a.pdf
http://www.datasheetcatalog.org/data...F125_126_2.pdf
http://focus.ti.com/lit/ds/sbfs019a/sbfs019a.pdf
http://www.onsemi.com/pub_link/Colla...MC10E164-D.PDF

Which of this ICs is to prefer, if there a reclocking unit is in use in front of the DAC and which without such a reclocking unit?

[karmik]

I'm the lowliest of the newbies here but a simple question comes to mind.

Why are you trying to overcomplicate your setup? Use a 2-poles electrical switch with as many positions as sources you want to feed to your dac/re-clock unit.

An SPDIF signal is, at heart, an electrical signal running on a coaxial cable just like any other (well, except it's a digital signal but it doesn't really make a difference for this matter).

I can't stress enough how much of a newbie i am with this matters, but i don't see why a simple eletrical switch tied to one spdif input on your dac wouldn't work?

This other option wouldn't work with a reclock unit between your source selector and your dac, but if you want to err on the "neat" side, just use the builtin multiplexer in your SPDIF receiver, i believe most units have it.

Twister Pear audio has a very good example of this layout:4:1 Receiver

P.S.:
I'm not 100% positive on what's the policy for external links, especially ones with a commercial side to them. If this link goes against diyaudio policies please remove it mods or drop me a PM and i will do so promptly.

[SY]

The issue is bandwidth and impedance matching. spdif signals, unlike audio signals, are transmitted with matched source, load, and cable characteristic impedance. If there's a significant mismatch, you get ringing and reflections. Will a normal switch cause enough of a blip to have an effect on the eventual audio signal? Possibly. An RF coaxial switch would be a safer choice- they're commonly used for antenna switching.

[tiefbassuebertr]

Originally Posted by karmik

I'm the lowliest of the newbies here but a simple question comes to mind.

Why are you trying to overcomplicate your setup? Use a 2-poles electrical switch with as many positions as sources you want to feed to your dac/re-clock unit.

An SPDIF signal is, at heart, an electrical signal running on a coaxial cable just like any other (well, except it's a digital signal but it doesn't really make a difference for this matter).

I can't stress enough how much of a newbie i am with this matters, but i don't see why a simple eletrical switch tied to one spdif input on your dac wouldn't work?

This other option wouldn't work with a reclock unit between your source selector and your dac, but if you want to err on the "neat" side, just use the builtin multiplexer in your SPDIF receiver, i believe most units have it.

Twister Pear audio has a very good example of this layout:4:1 Receiver

Indeed - an ordinary rotary swich will work here. Also by switching other RF signals (e. g. various LNB's for sattelite tv). But this isn't the royal way. Especially in case of unwanted contact resistances through oxydation you will get unwanted audible effects. I want to have electronic switch solutions therefore.

Originally Posted by SY

The issue is bandwidth and impedance matching. spdif signals, unlike audio signals, are transmitted with matched source, load, and cable characteristic impedance. If there's a significant mismatch, you get ringing and reflections. Will a normal switch cause enough of a blip to have an effect on the eventual audio signal? Possibly. An RF coaxial switch would be a safer choice- they're commonly used for antenna switching.

Yes, or RF relais, like used by Nelson Pass DAC D1. But I prefer an electronic solution, because I want to switch also various things inside of the DAC (e. g. between oversampling and NOS mode).

Originally Posted by karmik

P.S.:
I'm not 100% positive on what's the policy for external links, especially ones with a commercial side to them. If this link goes against diyaudio policies please remove it mods or drop me a PM and i will do so promptly.

This I don't understand. The adding of external web addresses in context to the topic is for me helpful and I cannot see here a disadvantage.

I have forget the attachement mentioned by post #1 - sorry

[Javin5]

Feeding the SPDIF signals directly over a mechanical rotary selector is not a good idea. However, the TPA 4:1 Mux is a neat solution; in addition, each source has its own input transformer to eliminate ground loops. You can still use a rotary selector, but it only provides the DC control signals. I have the TPA module and for me it works very well.

[karmik]

Very good point, the RF switch solution is still much easier and quicker than building a whole pcb for this. Are RF switches very expensive?

The impedance of a common coaxial SPDIF cable is 75 ohm right? Are there RF switches with this impedance? i thought RF equipment was commonly 50 ohm?

Thanks for your insight so far, i'm gonna be needing something like this in the near future when i decide to cough up the money for a proper dac kit

[SY]

Originally Posted by karmik

Are RF switches very expensive?

The impedance of a common coaxial SPDIF cable is 75 ohm right?

$1.98 buys one at the electronics shop nearby. 75 ohms. Good for several hundred megahertz. The only complication is that it uses F connectors instead of BNC, but it's not like that's an insuperable problem.

[karmik]

Well it seems the Original Poster wants a more versatile solution, time to fire up your cad software then!

The TPA solution looks very good and i think i will pursue that way when the time comes. Only two problems come to mind:

A) I quickly looked up the transformer, am i reading right? 10 dollars each? That's hefty.

B) Since it's a fully digital switch happening inside the IC, can it be done "live" (with the DAC turned on)?