|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
85.147.82.48
In Reply to: RE: why so different? posted by EmailTim on January 05, 2011 at 08:25:18
Hi Tim,Okay, now your choices are more clear too me.
1) I tried al kind of OS-es XP, win7, Linux, OS-X SL (hackintoch) on the cMP hardware.
I found that Windows and OS-X have exactly the same sound quality. There is no difference between those OS-es, the difference is in the audio-software/drivers/player that are used on Windows or OS-X.
I expect the same for Linux. Although when I tried linux Mint10 lately with some of the standard audioplayers that come with that distro, but I found that all the players sound somewhat ‘dark’ / ‘brown’. Not being my taste. But may be there are also better sounding linux players outthere, that fit my taste.
5) Since you want too do a small form factor experiment, have you considered an socket 1156 mITX MoBo from Gigabyte? Chipsets, sockets and processortypes, all do have their own specific sonic fingerprint and sonic quality. I don’t know the sound od the Intel NM10 chipset and Intel Atom D525 processor, but I know that the new I-core format definitely sounds better than the old core-2-duo / socket 775 format.However, Cics doesn’t recommend the use of mITX MoBo’s because they only just have one PCI(e) slot that is directly coupled too the processor (or something …). You better ask Cics for some background info on this.
7) PSU.
Power supply, power consumption and keeping the power as clean as possible, are main topics in the cMP project.
I stopped using 230 Volt AC ATX PSU altogether. Almost all ATX PSU throw back lots of pollution and distortion into the 230 V AC power lines that feed them, and thus spoiling the power for your amps, DAC and other auxiliaries. Also at the other DC end, ATX PSU’s also throw lots of pollution and distortion into the 3.3, 5 and 12 volt DC lines and ground lines.I followed the concept of a ‘clean’ and a ‘dirty’ PSU by using 2 Pico PSU’s.
I know Pico’s are also ‘switching’ PSU’s, but the switching technique is only used for making 3.3 V DC and 5 volt DC out of the (linear) supplied 12 V DC.
My hope is that the switching-technique used to get 3.3 V DC and 5 V DC out of 12 V DC results in less pollution and distortion than the switching-technique used to get 3.3, 5, 12 V DC from 230 AC.
Just a thought. I don’t know if this is true.With maximum sound quality in mind I would avoid feeding a Pico PSU with a SMTP brick. Instead I would use a linear PSU to feed it. This way you already avoid 1 noisy and polluting 230 AC switching PSU in your audio-setup.
I follow the concept using two PSU’s. One for dirty components and one for ‘clean’ components. The idea is in preventing that dirty components spoil the power for clean components.
So I use two Pico PSU’s with each have their own dedicated linear PSU that feeds the Pico PSU’s 12 V DC.
One Pico feeds all the ‘dirty’ components: USB, HDD/SSD, DVD and the P4.
One Pico feeds the P24.* Does the concept of separating ‘clean’ and ‘dirty’ work?
I don’t know. I have no scope. But I do know that pollution and distortion can freely travel around on all wires, cables, parts and cards that somehow are connected too the MoBo.
Also in a proper circuit design the real earth is not connected and/or used as a ground. But I think in PC’s the real earth and the ground ( 0 volt) are connected.So probably ‘the dirt’ of ‘the dirty parts’ will somehow reach the clean section and spoil it.
So my main strategy is to avoid dirty components everywhere as much as possible (if no dirt is generated, it can’t spread). Or use a less dirty one (less dirt). Dirty components that easily can be avoided are: 230 AC ATX switching PSU’s and the use of motors in fans and HDD’s.Two months ago I did an experiment. I put 3.3 Volt DC on a 5 Volt pin in the P24 connector. This 3.3 Volt DC showed up everywhere on the MoBo, HDD, cards, casing, ect. No matter where I put my Multimeter the 3.3 Volt DC was found everywhere.
So know I see some relevance in RME stressing in their product specs that their soundcards use ground free transformer coupled digital output connectors. And also in Cics recommendation too preferably use a high quality toslink connection (and an interface + DAC with a high quality optical sending and receiving units)
* Grounding
So to answer your grounding questions and taking in account the above, that’s a really though one. There are so many unknown grounding routes and unkown real earth connections.
For me it’s impossible too give a specific tip or advice.
After ready this post, I hope some other more knowledgeable inmates have some more thoughts about grounding.
But i can tell you what I do on grounding in my setup.
1) I connect the casings of both linear PSU’s together with an wire.2) I try to sink ‘dirt’ on the DC lines and on the 0 Volt lines (gnd). I do this by placing capacitors, which (hopefully) shunt high frequency AC too a real earth connection. (as they also do in mains filter circuit designs)
I place this capacitance before Pico PSU’s on the 12 + 0 Volt DC lines, and after the Pico PSU’s on the 3.3, 5 Volt and 0 Volt DC lines.3) Since dirty AC high frequency also can travel trough a transformer, I decided too start using a high quality toslink connection. This way too be sure too have a real good galvanic isolation.
Hope this helped somehow.
Mark
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Edits: 01/05/11 01/06/11Follow Ups:
Hi Mark,
How do you judge/pick a linear power supply?
I have found these Acopian PS but don't know what a good ripple specs are.
Their High Performance A model ripple numbers, are 0.25 mV RMS with a regulation of +/-0.005%.
Thanks,
Tim
Hi Tim,
I’m far from an expert on this matter. So i’m not the right person too answer your question with some real and in depht knowledge.
I know from comparison that only just 0.25mV is a very good spec! and 0.005% is also very good. But than again, these are so good, that I suspect them only too be applicable for some optimized and limited working range/situation.
I also understand that ripple and regulation are important specs, but also that they unfortunately only show just small part of the bigger picture.
From what I understand it’s also very important that the DC output is free from AC noise (50-60 Hz power hum + there harmonics and also free from (very) high frequency noise + there harmonics).
But unfortunately no manufacturer or vendor gives any useful information on this. Even very respectable manufacturers like TTi (Thurlby Thandar Instruments). I bought a second hand TTi PL330TP as a reference. But I also bought 3 other linear PSU’s ranging from 34 euro’s till 85 euro’s. I think (!!) the cheapest one, sounds just a little less. But I can’t hear any differences between the 85 euro one and the 670 euro’s TTi.
So I chose them on reputation (from what I read from other inmates). But that is also a not so reliable source. Many just repeat what they have heard from others. And also every inmate has his own acquired taste on what sounds good. Many inmates like an ‘airy’, ‘creamy’, ‘silky’, ‘tube-like’ sound character. Which I don’t like. IMHO it has little too do with accurate reproduction. But I must admit, it sounds very enjoyable.
I think the Velleman PS1503SB, at which Cics pointed a while ago, is a good choice.
The specs are not spectaculair, but to me it sounds the same as the TTi.
Mark
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
hi mark,
thank you for this very informative posting.
also I have now got the XT150 PICO PSU that I hope to install very soon, but I wonder why you use two PICO´s ?
my impression was that USB is powered via the motherboard.
the way I plan to implement my new PICO is to power the P4 from a linear 12V supply; feed the harddisk from a regular external unit like the granite; and another linear supply for the PICO in the P24 slot.
I use only USB for the wireless mouse.
does this sound sensible ?
kind regards
cMP2 Computer (XP minlogon) Intel E7400 cPlay039>Allocator>Lynx Two B. /192kHz throughout. 2x AcousticReality Ref. 202 & 2x AcousticReality Ref. 601´s ICEpower. Magnepan MG3.3R beechwood frames & custom stands. Miller chokes on Ribbons.
Hi Playmate,“Does this sound sensible ?”
Well, it’s not 100% by the cMP-concept. The cMP concept recommends too use an extra USB-bracket. By doing so the USB-ports on the bracket can be powered separately.
See item # 13 here: http://www.cicsmemoryplayer.com/index.php?n=CMP.05Components#mobo.
( On the software side of things: I also disable all not used Universal USB Host Controllers)
Using 2 linear PSU sounds sensible too me.
I also used too power my MoBo that way.
But a don’t like the 90-260 Volt AC powered Granite’s digital PSU inside my PC case.
I really cannot see why some one would want 90-260 Volt (unshielded) AC wires running inside their PC alongside with delicate 3.3, 5 and 12 Volt DC power lines.So that is why I started to use the extra Pico.
But using this extra Pico, also allowed me too follow Cics concept of putting all ‘dirty’ components on a separate PSU and using another dedicated PSU (the so called ‘clean’ PSU) too exclusively power the P24.Cics describes this concept by using two conventional 230 V AC ATX PSU’s.
But since ATX PSU’s have a notorious reputation of throwing a lot of ‘dirt’ around. Into the 3.3, 5 and 12 Volt DC lines, as well as back into the 230 AC lines that feed them, I try too avoid the use of these ATX PSU’s.
Instead, I implement this scenario with the use of two PICO PSU’s. The PICO’s still are also switchers, but I hope they don’t throw so much ‘dirt’ around, as there bigger standard 230 AC ATX PC PSU cousins do.But please keep in mind I have no education or degree in electronics ( I’m just a passionate hobbyist). I just ‘creatively’ follow this recipe and share it at this forum. Hoping that more knowledgeable inmates comment on it and/ or give useful suggestions.
This way I end up with two linear powered XT150 PICO PSU.
These XT150 leave the 12 v DC line untouched and just pass it on. One can check this easily by varying the 12 V DC input. The 12 V DC output varies exactly in the same way.* the power supply chain for ‘dirty’ components than looks like this:
Linear PSU -> XT150 PICO PSU -> HDD or SSD
......................................................-> DVD
......................................................-> USB Bracket
......................................................-> P4 (but I prefer too tap the 12 V DC tap directly from the linear PSU)* the power supply chain for ‘clean’ components (P24) than looks like this:
Linear PSU -> XT150 PICO PSU -> P24About ‘smooting’ & ‘filtering/ sinking’ dirt.
On both supply power lines I use smoothing caps like they are used with regular ATX PSU.
But the XT150 PICO PSU’s don’t tolerate high value’s.
I also try to ‘sink’ High Frequencies on 3.3, 5 and 12 Volt DC lines and 0 Volt lines too a real earth. (for instance like this is done in mains filters)Right now I study how too use LC style filters that can be used on the 3.3, 5 and 12 Volt DC lines and 0 Volt lines. But lacking advanced knowledge about electronics , this is extremely complicated for me. I cannot find much information on how too keep the LC filter resistance (R) extremely low and also on how too effectively prevent the filter from resonating.
Using Google I haven’t found any circuit examples on how too use such LC style filters when powering the P24 socket (so on the 3.3, 5 and 12 Volt DC lines, 0 Volt lines).
Also on this forum there not much response too this approach. Most knowledgeable and skillful inmates, (understandably) all go for the ‘full linear route / approach”.
But for less skilled inmates it should be not too hard too build an LC filter and use that between the PSU and the MoBo.
The only thing I need is a well designed LC circuit for this purpose.Your approach sounds sensible too me.
Accept for not using the recommended separately powered USB bracket.
I also would not like a 230 AC volts powered Granite inside my PC case
(although it is recommended in the cMP concept).Mark
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Edits: 01/07/11
Hi Mark,
Thanks for your feedback.
I have been considered trying a 1156 mobo next after experimenting with this one. I see the SandyBridges will be using a 1155 socket and is supposed to support higher bandwidth on the PCIe bus than the 1156 which could reduce some contention. Time will tell.
The Granite Power Supplies should be arriving tomorrow so I will be testing them next.
I have 120 power here and use Equi=Tech balanced power conditioners on my audio equipment. If I find suitable 3-prong 12V linear power supplies to subsititue for the PICO brick(s), the combination may prove fruitful.
Sincerely,
Tim
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: