[Wijnand]

One more note.

Roon chose this track to play

Watching the Waves (Acoustic Version) from David Gray.

The Album GOld in a Brass Age (Ground Control Acoustic Sessions)

The singer has an incredibly clear and natural voice. It’s just an internet stream

[Wijnand]

Wifi is not possible for me. I think I should stick with the current situation that works. It’s still on an incredible level. Really on top streamer levels. There seem to be limits in one of the parameters. I think it’s the clock banging.

[Wijnand]

In my situation I use a Jcat which has a galvanic isolation to the PC’s powersupply. It uses an external powersupply. Which is a Farad in my situation.

Beside this. Still after 5 years. I would be the only one that has problems with these streamers. I also think I’m probably the only one with a very clean powernet, has a dedicated network for audio and upgraded each an every powersupply on the network

[Wijnand]

I have measured the PC parts in the Pink Faun streamer. Indeed they are not measuring good. It is still a PC.

On the ohter hand when people listen to the sound it produces they still pick it above the GrimmS

[Wijnand]

Due to the fact the AQvox became a problem. Which is connected between the Audio PC and the Streamer / DAC. I think the clock banging is the most logical explanation. Even more because even replacing the fuse in the Super 10 by a QSA Red will cause the same effect when using the normal switch.

Still what I don’t get is when I connect several passive LAN  isolators in series to reduce the signal, which should be near perfect. That the sound will not only clear up, but it will also sharpen and lose strength, body, natural colors etc.  An explanation could be that the DAC must work hard to restore the weak signal

[Wijnand]

In those day’s I didn’t suspect the ethernet yet. So it was most certainly connected

[Wijnand]

Somehow I think my stream is too strong or some noise will be highlighted when other noise is cancelled or the network clocks are banging into the audio clocks.

For the record. All my clocks are standard at the moment. No OCXO’s or whatsoever

[Wijnand]

But also when the DAC/ Streamer is connected directly to the Router. It will become dark. Even when a CD player is hooked up to the system, it will remain dark.
There’s one thing they all had in common and that is the ethernet was connected to the system. Actually when I replace the Cat5e SFTP with CAT8 the problem will become worse. Also when I put a few powernet filters in the chain, the problem will become worse.
Yesterday it seemed that even the AQvox switch was too much. The system went dark later in the evening. It cleared up after I had put the normal D’Link DGS108 back in the chain. Actually the AQvox and the DGS108 are the same switches, but the AQvox has higher noise reduction

[Wijnand]

Good morning Tobias, Yes that is what I think as well, the PC. I remember that for a test I also chose my wired tablet as an output. It was dark as well, but also the television when using Netflix while connected with Toslink to the DAC. Also when using AES/EBU from my Pink Faun streamer to the DAC. The motherboard of the Pink Faun streamer was replaced 4 times, due to this issue. Also I2S is used and USB output. Those were all high grade audio modules.

I also have a NAD M12 dac of which I replaced the stock VCXO clock with an ultra OCXO. The OCXO was fed by a Farad powersupply. When the NAD became dark back in 2019 I could switch off the clock and the sound didn’t change. So the NAD seemed to play on its TTL, but when I reconnected the clock after 45 minutes you could hear everything cleared up for a while.

Also last week I built a completely new PC, Asus Prime, intel I5 with a Seasonic Fanless power supply which is a SPS, but a pretty good one. This PC is compared with my other PC which was built 7 months ago. It has a Seasonic as well, MSI motherboard and Intel I9 one of the very fast. They all suffer the same problem. In the near Future there will be a Farad ATX which shall replace a Seasonic. This Farad ATX will outperform the Pink Faun powersupply. Both are from the same hands.

 

[Wijnand]

One more thing. My current dac also plays for about 5 seconds after I disconnected the ethernet. The sound never changes at that moment. The DAC always keeps the same quality as which the buffer is filled.

This is a really difficult thing to connect with the noise aspect

[Wijnand]

Thanks! I will think about this deeply for the coming days.

I just have done some test’s and processed some thoughts regarding noise aspects. My first thought is that when I only use switch mode power supplies. That there’s no trouble with dark sound. The sound is really not that good with switch mode power supplies, but there’s no trouble.

Theoretically it could be possible that a high grade powersupply could magnify some sort of noise which  already came with the stream from elsewhere. Also the ethernet signal itself is some kind of noise for an audiosystem. When it becomes a really stable blockwave who says the DAC can handle that. In the end the audio clock runs at another speed when it would drift up towards the 25Mhz would be a problem.

Well there are already much more toughts that I could write down, but I will give it some time first. Because many experiences still doesn’t fit. 🙂

[Wijnand]

Here I will trust your knowledge about ethernet. Still the theorie could work. When we see what is happening in the network as a side effect. I was grabbing in my memory. Where we tested a R2R / NOS dac and some I2S dac’s. These dac work on an external clock which of course also can suffer and be influenced by noise as well.

When the only thing that can cause for example dark sound is the DAC itself. And let me first tell there must be several kinds of noise of which at least a few aint the problem, because they are canceled. Everything is always very pure, the only thing that troubles is the dark sound. This must mean the data is due to jitter only half transferred to analog. For example half way the clock cycle.

Then the question is why will this only happen after several hours and never immediately. Then it must be a clock that is slowly drifting and cutting data in parts and only convert half bits to analog.

I think a clock will drift due to another clock that’s driving. This seem to be a clock in the ethernet system

 

 

[Wijnand]

Thanks I will do some more homework! 🙂

[Wijnand]

Well that doesn’t work correct than. This is tried with the Perfect Wave Dac 2. They called it the memory player. But still when you disconnect the ethernet after the data was loaded. The buffer sounded just as bad as with the connected ethernet. So there’s something not right.

[Wijnand]

Then there’s no problem 🙂 and we are back in 2010 (just kidding)

I think these components are called transistors and capacitors. They don’t understand 1 or 0. They work with a voltage. From a certain level it will open and close when the glass is empty. So it’s an analog process which we call digital

[Author]

Posts