Audio Switching

[JRJR]

Thanks for all your help guys!

To sum it up:
1. Relay could work (with small clicks) if a snubber helps to reduce the voltage spike in the relay.

2. Opto Switching works great too. Same goes for FETs.

3. A VCA would be a very clean solution.

4. Devices like ADG1411 are also a way to go.

I'm ordering parts now. Will breadboard all options and see what works and feels best.

Thanks a lot! I'll keep you postet.

Cheers
J

 

[JRJR]

So finally I made it work with just some relays & diodes.
The only question left is:
Should I short unconnected connections to ground?
Because somehow (I'm not sure why) there is still some leakage from the cut signal present after I opened the relay.
I'm not sure why that is. Probably because the overall thing is so complex and just breadboarded I get some feed trough somewhere.
But just as a general question, should I?

Thanks!

 

[Martin Griffith]

This part of Doug Self's tome is informative  on FET switches
https://books.google.es/books?id=X9OoCgAAQBAJ&pg=PA129&lpg=PA129&dq=j111+circuits&source=bl&ots=xfVZzJV3IF&sig=dc8IwSpUaNXezsrmUKyyWDAXpAg&hl=en&sa=X&ved=0ahUKEwjrh_-08bjOAhVMthQKHYKcAr0Q6AEIODAF#v=onepage&q=j111%20circuits&f=false

 

[Bo Deadly]

Because somehow (I'm not sure why) there is still some leakage from the cut signal present after I opened the relay.
I'm not sure why that is. Probably because the overall thing is so complex and just breadboarded I get some feed trough somewhere.

That's almost certainly ground line bleed. You can't just connect a ground anywhere. Even though it's the same network, there is a tiny resistance and inductance. So if you ground a signal at any random point, it can easily bleed back into another circuit. The ideal arrangement is to make a "star ground" where all grounds run back to the power supply filter caps. But 90% of the time that is just impractical and, more important, not necessary. In practice ground lines would be more like a tree. Or a more accurate analogy would be like a subway map where you have a few lines converging in one spot and things connect to them at various points. You could probably fix your breadboard by just trying to move important ground connections to a common point next to the filter cap of the supply.

 

[abbey road d enfer]

So finally I made it work with just some relays & diodes.
The only question left is:
Should I short unconnected connections to ground?
Because somehow (I'm not sure why) there is still some leakage from the cut signal present after I opened the relay.
I'm not sure why that is. Probably because the overall thing is so complex and just breadboarded I get some feed trough somewhere.
But just as a general question, should I?

Thanks!

You're probably experiencing capacitive leakage. When open, a relay still has a capacitive path, of a few pF. In order to minimize the resulting x-talk, the impedance of the circuit must be very low. Since the actual impedance of the circuit is that of the connected source, you must reduce its impedance.
In the absence of a connected source, x-talk is to be expected.

 

[dogears]

. Meaning try adding large resistors to ground (like 22K on each input and 1M on each of the other switch contacts).

Can you sketch what you mean here?

 

[Brian Roth]

Indeed, without seeing a schematic of the design there is no way for us to guess what the problem might be,

 

[JRJR]

Thanks guys! That adds a lot of knowledge to my brain.

After rewiring the ground 5 times I found a cold solder joint at the switch.
It's all working now. Thanks a lot!
The occasional clicks are totally acceptable.

JR

 

[scott2000]

sorry to pick here to drop a change but, Anyone familiar with the Kramer matrix switchers? I picked up one to use as a spdif switcher and the seller sent me the wrong unit, although they do look identical to be fair.

The one I have has R,G,B connections, it's the one on the left of specs image. The one on right is what I was supposed to get. Out of curiosity and the seller's ok, I used one of the rgb connections and it does seem to work switching spdif but, I'm pretty sure there's a loss of quality in the audio compared to not using the switcher but I haven't really dug into checking . Just unplugging and listening with the several second memory I have of the sound.

Any reason using these video connections wouldn't work to switch spdif? I'm not familiar with spdif or what it takes as far as components. These particular units are active, the audio looks like it's switched using some analog devices 75019 and has opamps for the ins and outs... wasn't sure if that would dictate how the video side of things worked....can't see that far into it......
Maybe just jitter I'm hearing or just a bad idea to try using it for something it's not made for??
Thanks for any insight

 

[abbey road d enfer]

I'm pretty sure there's a loss of quality in the audio compared to not using the switcher but I haven't really dug into checking . Just unplugging and listening with the several second memory I have of the sound.

I think you're the victim of cognitive bias.
Corrupting a digital signal typically results in go/no go, not in subtle modifications.

 

[scott2000]

I think you're the victim of cognitive bias

I think you are right.

 

[john12ax7]

It might not just be cognitive bias. Digital can and does sometimes sound different. Errors are not always corrected and jitter is not always suppressed.

 

[abbey road d enfer]

Digital can and does sometimes sound different.

It happens in the A/D and D/A conversions, not in the transmission of digital flux.

Errors are not always corrected and jitter is not always suppressed.

I takes a lot of insane degradation to make jitter high enough to create issues; it usually comes with non-subtle audible defects.
Let's exclude problems due to error recovery that happen in CD and digital tape.

 

[scott2000]

Here are a couple of clips from a cd recorded with and without the kramer...24/44 files

one is cd spdif out to kramer video in/out to dac spdif in/ analog out to ad to computer
other is cd spdif out to dac spdif in/ analog out to ad to computer

 

[john12ax7]

It takes a lot of insane degradation to make jitter high enough to create issues; it usually comes with non-subtle audible defects.

I don't agree. If, for example, you try a bunch of external clocks you will eventually hear a difference. The differences can be subtle and not to the level of something is "broken".

You can do an ABX test and pick out the differences, even though you might theoretically think the PLL would make them inaudible.

 

[scott2000]

I forgot to clock the computer interface to the ad in the earlier files....... ....here are new files...same routing.

one is cd spdif out to kramer video in/out to dac spdif in/ analog out to ad to computer
other is cd spdif out to dac spdif in/ analog out to ad to computer

..turned up the dac a little too

 

[swpaskett]

Irrelevant, but I'll say it anyway. Back when I was 18 I bought my mom a tabletop radio. She said it sounded really good and I laughed. I told her it sounded like a phone - no highs, no lows, and lots of distortion. Now that I am the same age as she was then, everything sounds the same as that radio - no highs, no lows, and lots of distortion. So, if you've got golden ears, take good care of them. I miss so much of what I used to hear. (don't mean to be a thread snatcher)

 

[zamproject]

here are new files

you have a "time" shift, phase it turning about 90° each 10 sec...

 

[scott2000]

you have a "time" shift, phase it turning about 90° each 10 sec...

Can you explain further?Both tracks have this? I do know the tracks aren't exactly lined up because I just recorded them and didn't align them but that isn't what you are talking about.

 

[abbey road d enfer]

one is cd spdif out to kramer video in/out to dac spdif in/ analog out to ad to computer
other is cd spdif out to dac spdif in/ analog out to ad to computer

I would suggest any difference (personally, I can't hear one) is due to the final A/D conversion. Even done with the same pieces of gear, distortion changes particularly with temperature.
There's a quite interesting thread on the Cosmos Discord channel, where the designer of the Cosmos ADC explains how distortion changes; he offers a thermostatic option for his converter.

..turned up the dac a little too

Did no one tell you that only one parameter may change in a comparative test?