Build a dedicated Audio PC

[jkeny]

The MK2 & MK3 devices connect via SPDIF to a DAC so imediately this complicates matters. Furthermore. the MK2 uses an external PS for powering an on-board chip. If this PS is grounded, this further complicates matters. Maybe Kyrill can say what he uses for this PS (i suspect batteries :))

OK, so let's try to keep it simple & focus on the MK3. If the DAC that it connects to is grounded then we could have a ground loop between the DAC & amplifier. But here's my take on this:
- the ground loops are very small current on the ground & translate into maybe mV of voltage or lower.
- these small voltages will only have an audible effect on the power supplies to very sensitive functions, like the power supply to clocks which time the digital audio output
- if a ground loop exists between the MK3 & DAC it's audible interference is probably mitigated by the fact that the ground connection is from the DAC's analogue output to the amplifiers input. This seems to isolate it from the more sensitive areas of the DAC.

A final piece in the puzzle - I experimented with introducing a grounded device into the MK3 by connecting a PS to it which powers the battery charger board. Running this while listening to music, I expected to hear it have a detrimental effect on the music - it didn't. I could switch this PS in & out while listening to very quiet music & there was absolutely no audible effect. So what we have here is a grounded PS powering a switched mode (SMPS) battery charger - a recipe for sonic mayhem, conventional wisdom would suggest. But it doesn't. Don't ya love the real world, it trips you up if your not watching carefully. So how is this explained - I don't know? Maybe the batteries act as huge capacitors & filter out any noise coming from the PS or SMPS charger board charger.

So there you have it - more than you ever wanted to know about grounds & noise. But I believe that the handling of ground noise is one of the last frontiers in audio. I also believe that some of the success underlying the great sound from Jplay has got to do with reducing ground noise coming from the PC by reducing the number of concurrent processing happening while playing back audio.

Anyway, back to the start - do we need an isolator (on my devices) if what I say above holds up? The proof is in the pudding but at $400 (Adnaco) it's an expensive pudding.

Can this isolation work between HDD & PC & give audible improvement. According to the above, no but I'd like to hear from Kyrill & see what his setup is & what his grounding configuration is.

But, it is a very useful technique for grounded devices, I suspect.

Now, if we could build this Audio PC without grounding............

[jaybee]

is there any benefit to moving from USB to FireWire or thunderbolt??

You mean for HDD or for audio?

both, from a theoretical perspective that is, not necessarily in terms of what's available...

[james]

9) Install windows 8 consumer edition ( actually a beta) it is free and sounds more musical than 7

Does this beta version expire ? [If it does expire then you would have to factor in the cost of windows 7 or 8 which is about 150 euros for a full -- non-bundled -- oem version]

james

[jkeny]

is there any benefit to moving from USB to FireWire or thunderbolt??

You mean for HDD or for audio?

both, from a theoretical perspective that is, not necessarily in terms of what's available...

I don't know the details of Thunderbolt or a lot about Firewire but I'm not sure where the advantages would be!

[kyrill]

Guys,
Before spending a lot of money on isolation devices, let's investigate this a vit in theory first & the in practise.
Very good

This is tentative information based on experiments & a little bit of knowledge. However my conclusions could be wrong so please jump in if you see something that doesn't make sense. Also, the real world doesn't always work according to theory - that's why it's fun :)

My DAcs are completely isolated from noise carried via the USB cable because they use battery power (no USB 5V power for me, thank you :)). That leaves the USB ground wire as a possible source for noise transmission.
yep regardless of grnd loops
If the PC is also ungrounded (which is the recommended configuration i.e laptop) then we have a system where there are no ground loops (CM noise) because only one device, the amplifier is connected to ground - so no loop possible.

I have experimented with this & when I use a laptop that has a charger which connects the laptop to ground then I can hear some haze & noise at the speaker output - mind you only when I turn the volume up high. However this noise effects the music at normal volume & robs it of the blackness that is usually there. This is one of those issues that I think you have to train your ear to hear in normal listening. I'm not saying that it is borderline requiring a double blind test but it is something that afflicts the whole audio spectrum & therefore not something that jumps out at you like non-linear distortion.

Now, I believe the benefits of these isolators is to remove this CM noise.
That is the technical reason, but there is/seems to be much more at hand Ground should ideally be marked + and the data lines - The Mother of electrons is ground. They "travel form grnd to plus. So the "active" stream of particles (electrons) travel from grnd to "plus" They should be uncoloured. Nasty Dirty grnd makes that highly unlikely. Especially floated grnd cannot "dump" these nasties into mother earth directly. Only grnd directly near the 15mtr thick copper pole 14mtr deep in wet earth is "infinite" and absorbs all "plus" variations in whatever form so its voltage is true deep zero..and onlt there grnd is c l e a n. I think ( but I am not an EE) floating ground is not zero volts. It is relative zero volts to data and plus lines. Compared with true grnd voltage fluctuations exists. Th grnd of the wall socket is also not zero volts. In fact in my house gnd is 1,5V-2 V different than the neutral of the wall socket, while they are both connected to each other with this copper pole in the entrance of my house. There the difference is zero further up the lines into the house they start to deviate from each other... .The (floating) grnd of a PC is definitely contaminated as I see it

However, you can see that I contend it should not make a difference with my DACs, right?
I think it will make a big difference with yr dac beacus the not so clean grn cannot present their "dirt" into the dac. However, the real world can throw you a looper & I have been wrong about different USB cables effecting the sound. I thought my devices should be immune to USB cable for all the above reasons & because the devices use asynchronous USB - I was wrong!! What is the mechanism by which USB cables do make a difference in this scenario? The jury is out but I have my suspicions! The same explanation, when found will explain hw grnd delivers hearable gremlins independent of loops.

The problem is that I can't use the ADUM (OLIMEX) devices on Hi-speed USB 2.o (480mbps) which is what my devices use. YEp That is why I use it only between music stored SSD and PC And I'm too cheap to spend $400 on the Adnaco as an experiment which it seems will handle this speed. You can rely on Marcin and the guys at South Korean high end forum. They even updated the clcok on the Adnaco with expected beautiful results. Also The Adnaco wil be used between the 2 SSD and the PC. But can be used between DAC and PCThe next best thing is to find someone who has an Adnaco & one of my devices. Marcin is probably the only one in the world that falls into this category - he had one of my devices on trial but I don't know if he had the Adnaco at the time. I'll ask him about this! No he did not as far as I know
By the way also the USB datalines coming from PC presumably more than its grnd is fully contaminated by the PC. Galvanic isolation cleans them up I have a metaphor for it:

The digital stream is a soup made of water and it may contain nice square blocks of frozen water made of that same water, that for this story cannot melt unless when instructed to. All lines are tubes all memory are cups Water alone (in a given set time frame) is "0", water with an ice cube is a "1". the nature of the water does not count as a third variable in digital land, because itis not digital, it may be pure and transparent or coloured and transparent or hazy to a point. too muddy or hazy does not work as too many cubes will be undetectable. The X factor is the nature of the water. when it is absent the water is pure
Only at the dac side the nature of the water becomes paramount as it passes it on to or influences the analogue side. the conversion part needs cubes made of and clean transparent water, a condition undetectable for the computer or anybody who can only think in the purity of zero's and ones


Now let's talk about the MK2, MK3 in the next post as this is getting kinda long.

[Rocker]

Crikey! I am away from my computer for a few days, to come back and find a detailed discussion on building a dedicated PC for audio. And I have not got a clue what you guys are talking about. To date the only dedicated computer for audio I ever heard was in Ken Morelands. Impressive too. So Ken if you are building one of these devices, let me know. I could be 'persuaded' to buy your old box...........

[HiFiFan]

Just a quick note on Firewire. The Weiss DAC202S (well respected and reviewed) use Firewire and there are a number of articles from the designer as to why Firewire is superior over USB. This was one of their stronger marketing points. However they have recently just added USB to their DAC (Firewire is still there) and this is a good indication of what a previous poster said about popularity and the ubiquitousness of USB.

[Ken Moreland]

Good man Tom, I'll build the price into my investment plan.
KM

[Ivor]

Just a quick note on Firewire. The Weiss DAC202S (well respected and reviewed) use Firewire and there are a number of articles from the designer as to why Firewire is superior over USB. This was one of their stronger marketing points. However they have recently just added USB to their DAC (Firewire is still there) and this is a good indication of what a previous poster said about popularity and the ubiquitousness of USB.

I think firewire fell out of favour simply because Microsoft (i.e. Windows) stopped supporting it.

[jkeny]

Kyrill,
Thanks for the answer/info.

Ground, however is a tricky concept. All it is is a reference. It is not necessarily 0 volts. In fact it hardly ever is 0volts as your mains measurement shows.

The important thing with references is that they are stable. If the reference fluctuates, it will cause anything using this reference to also fluctuate. This is the problem to avoid. So avoiding current flowing on the grounds between two chips or two devices is crucial. Controlling this between chips is tricky but not impossible as it is under the control of the pcb designer. In a lot of cases using a star ground ensures that return currents in a chip don't contaminate the ground of another chip.

Controlling ground between devices is more difficult as it isn't always under the control of the designer. So concepts like star ground don't apply. This is where the problems usually arise. If the ground of one device is at a different potential to the ground at another device then current will flow between devices & a voltage potential will develop.

So when you say that a PC's ground is noisy, I believe this noise has nowhere to go if the PC's ground is floating & the USB devices's ground is also floating. This noise only becomes a voltage when it goes to ground, hopefully at the amplifier, where such low voltage noise will not have much audible effect.