Building a decent input switcher

[RubenV]

Hi,

Thanks for reading! I want to make an input selector since my active speakers have only one input and using my surround processor degrades the sound of my dac (which has a built in preamp). The dac direct connected to the speakers sounds a lot better but switching cables all the time is not convenient for me.

The only requirements are: 2 inputs, one output with the best soundquality possible.

I looked at commercial available equipment like the dodocus (but it degrades the sound) or manley skipjack (too expensive and to many options I don't use).

I do want to keep things rather simple and have something like this in mind:

An externally hosted image should be here but it was not working when we last tested it.

Since I want the input selector to be as transparant as possible I am looking for recommendations of parts I should use.

Does something like a DACT CT3 degrade the sound? Should I use another kind of switch? What kind of bindingposts and internal cables are recommended?

 

[AndrewT]

you will need to shield that box to prevent interference affecting all the open loops.

All the seams in the shielded box will need to be in electrical contact, otherwise the interference still leaks through.

 

[DF96]

I can't see any signal ground connections in that picture. Maybe it uses the chassis as the ground? If so, not the best way to do it. Ground should follow the signal. However, for many purposes you can get away with doing it 'wrong'.

 

[RubenV]

you will need to shield that box to prevent interference affecting all the open loops.

All the seams in the shielded box will need to be in electrical contact, otherwise the interference still leaks through.

Thanks I will take this in to account when I start building it. I can welt the side together and use screws for the top or something like that.

I can't see any signal ground connections in that picture. Maybe it uses the chassis as the ground? If so, not the best way to do it. Ground should follow the signal. However, for many purposes you can get away with doing it 'wrong'.

How does this grounding works? Can it be done inside, like a grounding wire to the chassis? Or does the chassis as a whole needs grounding?

Any thoughts on the internal wiring (shielding perhaps or a different layout) or the switch itself?

 

[rayma]

I want to make an input selector Since I want the input selector to be as transparant as possible
I am looking for recommendations of parts I should use.

http://glass-ware.stores.yahoo.net/atandsise.htmlhttp://glass-ware.stores.yahoo.net/atandsise.html

 

[AllenB]

I can welt the side together and use screws for the top or something like that.

Multiple screws through overlapping joints using shakeproof washers that make contact through the paint should work.

 

[DF96]

Use coax or twisted pair for the wiring, so the ground follows the signal. The sockets should be isolated from the box. The grounds should be connected together near the switch. In addition, there should be a connection from here to the metal box.

 

[AndrewT]

..................to prevent interference affecting all the open loops..............

...........................Any thoughts on the internal wiring (shielding perhaps or a different layout) or the switch itself?

Use coax or twisted pair for the wiring, so the ground follows the signal. .................

Close coupled Flow and Return of the signal wires is very important. Open loops must be avoided.

 

[bentsnake]

'
If I understand the requirement, it seems a single DPDT (double pole, double throw) switch is all that's needed for switching.

Apparently two low level, low impedance audio signals are to be switched. Essential these are line level signals, not especially susceptible to RF (radio frequency) or other interference. In a living room environment the circuit below could probably be assembled in a plastic box.

Of course a metal enclosure is always better, in which case the enclosure should be grounded as shown. Note the single-point grounding near the output jack.

Additional notes:

1. Mounting nuts should not be relied on for a ground connection. These nuts can loosen slightly over time, causing static, hum ("ground loops"), and other problems. Instead run separate ground conductors as shown. Insulated jacks could be used, as was already suggested.

2. In this application conductors should not be twisted into "twisted pairs." This is because there are no feed-and-return pairs, the left and right channels are different audio signals. Simply keeping conductors as short as possible is sufficient.

3. Shielded stereo patch cables are used for connecting devices. "Patch cables" by definition have a male connector on each end. These are available on eBay and elsewhere in a variety of lengths.
.

 

[M Gregg]

There is more to it than this,

It depends how far you want to go, there is not enough current to maintain a good connection across a switch and you might find reliability issues.
The contacts can get dirty and lose contact..

You talked about transparency and then change to a simple switch..

As DF96 pointed out the grounding is as important as the signal live. The signal travels down both lines..

However if you just want a quick change over then go with the switch..
You forgot what you are talking about is basically part of a passive pre-amp ie the selector without the volume pot..

People argue about SS switching and relay switching and then the contact type and metal contacts..

Here is an example of relay input..However you can just go with a switch with silver contacts..its up to you..

Chassis ground and signal ground should be separate and not part of the signal path..
Phono sockets should be isolated from the chassis and the ground for the chassis connected to the output..why because CD players and ipod etc don't have a ground..some people switch both ground and signal..however it depends how far you want to go!

So a passive selector can be passive but powered..relay or SS..or you can just go selector<<as long as you are happy with it!
As simple or complicated as you want..🙂

That's without saying things like whats the point of expensive cables then using different cable inside a selector?
So you see its where do you stop..

Regards
M. Gregg

 

[BillWojo]

Hi, I picked up a Kramer AV switch, model VS-4X4YC at a recycle yard. It has 4 inputs and 4 outputs. Does anybody know anything about this? It's pro AV, likely out of a studio so it must be decent. Inside it's filled with a CB and very nicely made. It doesn't use a rotary switch but instead has 4 input and 4 output buttons along with a readout of what has been selected. Can the video switching side be used for digital audio? It has BNC connectors for that and regular RCA connectors for audio.
Rack mountable (1U) so it has the same form factor as my Carver gear.
Thanks

BillWojo

 

[bentsnake]

...there is not enough current to maintain a good connection across a switch and you might find reliability issues.

That isn't exactly how it works. The "wetting current" you're referring to applies to establishing electrical continuity, not to maintaining it. It's generally of concern with very delicate, very low contact pressure switches--such as sensor switches--and in digital circuits.

This is the opposite of any DPDT switch available from a supplier. These switches have very high contact pressure, and a switching action that can best be described as violent.

Otherwise, your comments are certainly valid, but in my view can be sidestepped in this particular circuit--which by the way is not a "passive preamp" because there's no such thing.
.

 

[bentsnake]

...whats the point of expensive cables then using different cable inside a selector?

For that matter, what's the point of expensive cables at all?

NOTE to those who can hear the difference between gold and silver speaker cables: Congratulations on your auditory transcendence, but I'm not going to argue about it.
.

 

[DF96]

In this application conductors should not be twisted into "twisted pairs." This is because there are no feed-and-return pairs, the left and right channels are different audio signals. Simply keeping conductors as short as possible is sufficient.

There always is a feed and return pair, as signal currents always flow in a loop. Each signal should be twisted with its own ground. The signals are switched by the switch. The grounds are connected at the switch. However, it may be that the left and right grounds can be kept separate in the switch box - they will be joined elsewhere so no need to introduce more loops.

 

[M Gregg]

Otherwise, your comments are certainly valid, but in my view can be sidestepped in this particular circuit--which by the way is not a "passive preamp" because there's no such thing.
.

Its terminology,

Icon Audio Passive Line Pre Amplifier for £349.96 in at Audio Affair

NB "we" had to put resistors to draw current in control systems to prevent contact resistance..for 27 years + in industrial control..in heavy duty contactor switch gear and control systems..including data acquisition. In the end we ended up with mercury wetted contacts or reed switches which had their own problems due to carbon build up creating resistive conduction.

Quad used there own ideas..
Just for interest..

http://www.angelfire.com/sd/paulkemble/sound9c.html

YMMV..

Regards
M. Gregg

 

[bentsnake]

There always is a feed and return pair, as signal currents always flow in a loop. Each signal should be twisted with its own ground. The signals are switched by the switch. The grounds are connected at the switch. However, it may be that the left and right grounds can be kept separate in the switch box - they will be joined elsewhere so no need to introduce more loops.

Actually no, there isn't always a pair, for instance now. The ideal scenario you describe is unfortunately impossible because the the left and right channels of a stereo patch cord share a common ground/return: the shield. It's really hard for me to believe you didn't already know this.

Of course one could get into multi-conductor cables, but I don't see that as being appropriate to the project described.
.

 

[bentsnake]

Its terminology,

..."we" had to put resistors to draw current in control systems to prevent contact resistance..for 27 years + in industrial control..in heavy duty contactor switch gear and control systems..including data acquisition....

No it's not, it's reality.

Yes, fritt current is applicable to industrial controls, as are 200 amp motor starters and and 440 volt transformers. Equally applicable to industry are grain silos and 3 ton cranes. No argument here.
.

 

[M Gregg]

No it's not, it's reality.

Yes, fritt current is applicable to industrial controls, as are 200 amp motor starters and and 440 volt transformers. Equally applicable to industry are grain silos and 3 ton cranes. No argument here.
.

I have never worked on grain silos..😀

Remember Decatron counters what a PITA they were..😀

http://en.wikipedia.org/wiki/Dekatron

Regards
M. Gregg

 

[M Gregg]

Reliance, Philbrick Nexus..a blast from the past..Teledyne a few names to conjure with..
Allen Bradley and IPC..what a nightmare..

I could write a book but the fault location guide would be bigger than the encyclopaedia Britannica..😀

Off topic..back to input selectors..

Yes whack a toggle switch in job done..🙂

Regards
M. Gregg

 

[AndrewT]

Actually no, there isn't always a pair, for instance now. The ideal scenario you describe is unfortunately impossible because the the left and right channels of a stereo patch cord share a common ground/return: the shield. It's really hard for me to believe you didn't already know this.

Of course one could get into multi-conductor cables, but I don't see that as being appropriate to the project described.
.

A circuit cannot exist unless there is both a flow and a return for the signal to pass from the SOURCE and return to the SOURCE.

All currents flow from their source and each of these currents MUST RETURN to their source.

All signals come from a source. The signal current can only flow if a return route is provided.
The flow and return must be close coupled, as they are in a coaxial or twisted pair, if one wants to minimise interference..