 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
82.156.128.190
In Reply to: RE: Big SSD? Or, small SSD with: 1. LAN enabled? 2. external SSD/HDD? posted by theob on February 17, 2011 at 08:26:53
Hi Theo,Here’s some info on my current setup.
* computer parts.
Mobo: Gigabyty G41M-ES2LProcessor: intel core 2duo E7300
Runs on: 1 volt. (so not lowest possible)SSD: Kingstone 8Gb
Ram: 1Gb Kingstone HyperX
CAS Latency Time: 3
Command Rate(CMD): 1
Rest of timings: auto
Runs on: 2 volts (so: + .2 volt)
* DC Power supply.
According too Cics idea, I divided the power supply into a ‘dirty’ section and a ‘clean’ section.
But I DON’T do that by using two standard Antex Earthwatts ATX. I do this through using 2 linear driven pico’s.
- ‘Dirty’ pico
Velleman linear PSU : > P4 + pico PW-200-M > USB bracket + SSD (+ dvd only when needed)
smoothing caps are on linear 12V lines too pico and P4.
smoothing caps are on the 5v lines coming from the pico.
- ‘Clean’ pico
EA linear PS 12V/5A: > pico 160-xt > P24 (I think the 160-xt sounds better than the 150-xt)
- smoothing caps.
smoothing caps on there 12V lines too the pico.* Sound Card.
- ESI juli@ digital part only.
Smooting/cleaning caps placed on 3,3V and 5 volt connections pins too (not used) analog part
- optical out used with quality toslink cable: van den Hul Optocoupler MkII
Cics is oh so right: optical out can sound very, very good!
But only when a high quality toslink cable is used (I first refused too pay 20 times more than for an average plastic one, but the difference is beyond compare)* AC supply 230 volts.
- The linear PSU’s to the computer are on a separate 230 spur.
The Linear PSU’s are behind an Epcos mains filter too prevent them spoiling the mains or vice versa.
- The Lavry has an switcher inside.
So the Lavry is also behind an Epcos mains filter.
- The active Klein & Hummel speakers are also behind two Epcos mains filters.
For them I use 16 amp Epcos mains filter.
Because each O300 has three dedicated power amps inside which consume a hefty total of 435 watt per speaker.* Grounding.
Al cases (linear PSU’s and computer case) are star-grounded to a real earth connection.* Use of LAN.
Use of 1 Gb LAN connection for accessing files on a NAS, doesn’t degrade sound quality in my setup.* under clocking.
For some reasons this setup can run on the lowest possible bus speed of 100 mHz.
I never could go any lover than 135 mHz.
But this setup runs on 100 mHz bus speed !* No upsampling or decoding used.
- I don’t use any up-sampling.
Neither real-time nor off-line.
I don’t hear any real quality improvement.
I also don’t understand why and how real-time computer up-sampling from 44.1 to 96 and than feeding this (real-time) up-sampled material to my Lavry DAC, will make it sound any better.
It still stay’s the same recording.
- I don’t use FLAC
I decode FLAC off-line back to WAV and than store the wav-files.
HDD space is very, very cheap these days.
I see no reason anymore too use compressed formats.
Up-sampling or decoding FLAC’s will take processor capacity.
I rather use this capacity to drive my LAN connection.* USB EMU 0404
I never loved my EMU 0404 very much.
It didn’t sound very good 2 years ago in a non cMP optimized PC using a RCA coax connection. So almost 2 years ago I lent it too a friend of mine.
1 week ago, I got I back.
When using the EMU 0404 now as digital interface only and using the optical out with van den Hull Opto coupler and the 04040 is connected too the separately powered USB bracket, this time the EMU USB 0404 also sounds really perfect (!) through AES pro optical in my setup.
But now it sounds as good as the Juli@. Although the EMU uses 20 % (!!) of the CPU resources, while ESI Juli@ is only using 2% max.* XP tuning
I do all XP tuning as per cics recommendations.
Even though I have no CDROM connected (only temporarily when needed), I still do this one extra, as RME recommends it on their website:
Disable CDROM Autoplay
Hkey_local_machine> system>currentcontrolset>services>cdrom. Set autorun to 0.If you have any question, do ask.
Mark
fully optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Edits: 02/17/11Follow Ups:
This is great detail... thank you! I have a Kingston ssd on order. I have a gigabyte G31... mobo ... what is the 41?s it a variation of the 31?
Hi Theo
The Intel G41 chipset is the successor of the Intel G31 chipset.
The G31 supports 4 GB ram
The G41 supports 8 GB ram
So when I couldn't buy the G31 anymore,
there was only it's successor left.
Soundwise there no difference (I think).
The Mobo's also look physically the same.
Only the G41 Mobo costs $ 12,- more.
See also:
http://en.wikipedia.org/wiki/List_of_Intel_chipsets
Mark
thanks. why did you run high voltage but a real low bus speed? I'm not criticizing I'm trying to understand your logic.
Hi Theo,
Don’t worry I don’t feel offended so easily.
Discussing things also provides me with new insights.
Keep in kind that I do the following reasoning without any real in-depth knowledge of how things really work.
But here are my considerations:
I just look at the currency meters of my Velleman and TTi linear PSU.
I try too get best of both worlds: stable power and accurate processing/timing.
- Stable power (with no spiky ups and downs) is imho not equal to: lowest power consumption.
- Although low power consumption certainly helps too get stable power from a PSU.
- Also internal clocks like a steady and stable power.
- A steady voltage level, also increases accuracy while identifying the threshold value’s for when something is a ‘one’ or when something is a ‘zero’.
- Reducing speed reduces power consumption but also reduces errors (increases accuracy)
- Stable power is also maintained when there is a steady and stable workload.
* Low and steady workload.
There is less workload when there is no up-sampling needed and/or no decoding needed.
Although these are very steady workloads, not using it, is the most simple and straight forward way too reduce workload.
* reducing speed.
I think this is also very straight forward.
In modern computers high speeds are not needed for simple tasks like processing a stream of audio bits too an external DAC.
So reducing speed for audio stream processing can be done without immediate trade offs, because one easily stays far above the minimum needed speed levels. Reducing speed however significantly saves power. Look at the currency meter!
I also think it improves accuracy
* reducing voltage.
To me reducing voltage is a less straight forward thing. I have a feeling (no real knowledge!) that one should not let a system work on the edge of a breakdown.
I like too stay well above the absolute minimum needed voltage levels too keep it working properly and accurate. May be this comparison isn’t valid, but overclockers raze voltages too get more accuracy at higher speeds.
Also look at the currency meter, there isn’t so much more too win on power, by going al the way down and letting the system work on the edge of breakdown.
So I consider voltage reducing, too be more of a balancing act. While reducing speed and reducing workload are more straight forward.
None of this is based on any real in-depth knowledge.
It’s just me guessing, speculating and conceptualising.
Mark
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:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]