Build a dedicated Audio PC

[DaveF]

So the only way to evaluate A & B is really to live with them in your system over a period of time, giving each one enough time for relaxed evaluation. Sure A/B auditioning will give us a handle on obvious differences, hopefully!

The counter argument would be that you get used to or accustomed to the new component over this evaluation time as audio memory isnt great or fades quickly. I think this happens in some cases.
That said I remember it took me about 2 weeks to realise that a pair of Kharma monoblocks were giving me listening fatique. Didnt show up in initial listening but became more apparent as those 2 weeks went by. It was more subtle than obvious...I just began not enjoying the music for whatever reason.

[tony]

I'd be more interested to see if transports make a difference due to the ground plane noise coming down the cable. Whether the Devialet is susceptible to it is another story. I need a transport to replace the on loan one at some point.

That is something you will have to test and figure out yourself I think John is the only person who will be able to work with you on testing that. If you want to do any tests on what I have just drop me a pm.

[jkeny]

I'd be more interested to see if transports make a difference due to the ground plane noise coming down the cable. Whether the Devialet is susceptible to it is another story. I need a transport to replace the on loan one at some point.

Sure, I can lend you a Ciúnas converter. Give me a week or so.

[DaveF]

The one chink in your example & one that is always used as a counter argument is the "properly designed" mantra i.e the non-ideal pcb layout in your example was causing your issues, the DAC itself was not shielded adequately from noise, other noise sources were not "competently" dealt with - you know how it goes?. And the conclusion that with a "competently designed" system there will be no problems in audio, either!

There is certainly some truth in this but it's an idealists view that doesn't cater for the existing reality of where we are in this venture. We don't understand all the mechanisms involved, we are interfacing a number of different pieces of complicated equipment together, one of which is a very complicated system (PC & OS) not designed to operate in a real-time environment that is audio reproduction.

So it's a highly complex engineering analysis which is not being addressed by the real scientists with budgets/equipment for doing this sort of research.

In my example, the board was one we adapted to suit our needs. It became apparent after a while that we getting all these tones all over the place. So it was a process of trying to find and isolate any sources of this noise. Of course sometimes it was difficult to do so as the noise tones where buried in the noise of the laser itself.
Anyway it was a lessons learned and we then had the knowledge to more competently design a board for our needs. Like you said above, putting shielding around the DACs, and around the power regulators, keeping the analog and dig circuits away from one another where possible, PCB layout to reduce crosstalk, more ground layers in the PCB stackup and so on.
We still had issues with this board when it came back but nothing we couldnt handle. At one point we were picking up a 98MHz tone but when u touched the laser heatsink with your hand the tone would disappear. The heatsink hadnt been grounded properly! It was actually picking up 98FM!

[jkeny]

So the only way to evaluate A & B is really to live with them in your system over a period of time, giving each one enough time for relaxed evaluation. Sure A/B auditioning will give us a handle on obvious differences, hopefully!

The counter argument would be that you get used to or accustomed to the new component over this evaluation time as audio memory isnt great or fades quickly. I think this happens in some cases.

Yes, to a certain extent we can become accustomed to a new sound & not notice it's flaws - that's why I said have both A & B to evaluate over time & see which you prefer over long-term listening.
When you say "audio memory fades quickly", I'm not sure we are talking about the same thing - yes the exact sonic details cannot easily be remembered over a short period - this is the sort of focussed, detailed listening I'm taking about - it is problematic to retain the memory of the details of the sonics of A & compare it to the same details of the sonics of B. Often instant switching isn't possible or available so time intervals between comparisons happen. Also even with instant switching, we are not comparing the exact same piece of music between A & B - when we switch we are hearing the next bit of the music, not the last bit that we just listened to.

That said I remember it took me about 2 weeks to realise that a pair of Kharma monoblocks were giving me listening fatique. Didnt show up in initial listening but became more apparent as those 2 weeks went by. It was more subtle than obvious...I just began not enjoying the music for whatever reason.

Yes, if our normal usage for our audio systems is to enjoy them over a long period of time then we need to evaluate any new device over a period of time. As you say, we can often miss issues with a short audition

[tony]

John Quote "Yes, if our normal usage for our audio systems is to enjoy them over a long period of time then we need to evaluate any new device over a period of time. As you say, we can often miss issues with a short audition"

I had a listen to a CAD Dac and expensive Focal and Martin Logan system driven by burmeister and mark levenison stuff it all cost a fortune. Impossible to judge on a demo how good bad or ordinary the dac was.
System sounded great but not an instant like that the avantegarde's in the Listening Suite gave.
It has been said here a lot and no doubt you need to get this stuff home and listen to it for a couple of weeks to make a definitive judgement otherwise it is just pot luck.

[Gazjam]

Hi, new member 1st post. :)

This thread caught my eye as I'm looking into linear PSU's for my music server and there's some great info here.
My question is to those who are running two Maplins 7A linear supplies into their servers...?

Got a couple today - boy they are big, no chance of hiding them behind the equipment rack!
My server build is I think quite similar to a few here, in that I have a PicoPsu fitted and wish to replace the power brick with the Maplins linear.
Also, I wish to hook my second supply up to the 12v cpu socket on the board.

Easy enough to hook up the Pico supply I guess?
Two cables 12v and ground terminated in a DC plug hooks up to the DC jack on the back.

I'm not exactly sure about hooking up the 4Pin 12V cpu socket on the motherboard though...?
4 wires from the board; 2 x 12v and 2 x ground... into 2(not 4) connections on the Maplins linear.
Common sense kinda suggests just connect the two +12V together and the two GND together and connect the two pairs to the psu.
Just wanted to see if thats right and what the rest of you guys did when settng it up.

Thanks very much. :)

Gaz.

[Sligolad]

Welcome Gaz,
You are correct in everything you posted regarding connections.
Just converted to the CPU supply at the weekend myself, I switch it on before firing up the PC and switch it off after the PC.
I read on Paul Pangs site that he is happy to leave it on all the time but I am sure this is just wasting power but I have not run a test on the current draw with the PC off yet.

I suppose at some stage I could get a small relay fired off the Pico to switch in and out but that is for another day.
Cheers, Pearse.

[nige2000]

yea just join the two +12v and the two grounds, thats what happens anyway in a 4 pin plug
good job that you didnt ask about a 8 pin cpu plug as thats twice as complicated. :)

be careful you can damage stuff if you make a mistake
yea best use the three seperate supplies if possible for best results, every step is worth it.

nige

[Gazjam]

Cheers guys, thats good to know and thanks for the welcome.
I'm already using an external linear for my soundcard so thats covered.

One thing, the mains cable on the Maplins linear looks a bit weedy...is it worth rewiring it?
The rest of my kit is fed from a balanced mains transformer setup using proper "foo" cable, and I found this area of mains pays off as well.
(Detailed here: http://theartofsound.net/forum/showthre ... -Heavy!%29)

Wondering if theres more to be had feeding the Maplins linear direct hard wired from the balanced mains...I like hardcore me. :)

Also, regarding connecting the Pico to the + and - binding posts... little devil on my shoulder is saying hardwire the DC cable directly on to the board!
That said the internal wiring looks of good quality but it would be a break in the cable using the binding posts...assuming it will make any difference and not just my audiophile OCD at work!

Just wondering if theres other ways to maximise the quality of the servers sound and if anyone else has gone down this road.
Does it sound awesome as it is anyway? :)
Expect it will be a jump up from the standard power brick.

ta.