Re: LRM schematics
Hi John,
Many thanks for sharing the schematic. I know we always could email you to ask the schematic but I think the schematic you just attached make this thread complete with your schematics.
Now I will start to solder the components which I have sourced here to the pcbs that you've sent to me and then find this UTOS will be more usefull than USBDI2S so I will try to make a board from UTOS schematic and to build it at the first try instead USBDI2S.
Any special consideration to use CS8416 instead using CS8414? It will be more easy to source the latter here.
Then I assume that I can use the oscillator you've sent together with the kits? Is that right?
I have one more question.
I have some op-amp based super regulator pcbs which ready to use now. I've bought them three months ago and planned to use one of them to supply the master clock instead shunt regulator. I have read from other thread in this forum that op-amp based super regulator is not to be worth in some critical circuit. What's your experience about this issue?
Thank you.
Hi John,
Many thanks for sharing the schematic. I know we always could email you to ask the schematic but I think the schematic you just attached make this thread complete with your schematics.
Now I will start to solder the components which I have sourced here to the pcbs that you've sent to me and then find this UTOS will be more usefull than USBDI2S so I will try to make a board from UTOS schematic and to build it at the first try instead USBDI2S.
-ecdesigns- said:
Any special consideration to use CS8416 instead using CS8414? It will be more easy to source the latter here.
-ecdesigns- said:
Then I assume that I can use the oscillator you've sent together with the kits? Is that right?
I have one more question.
I have some op-amp based super regulator pcbs which ready to use now. I've bought them three months ago and planned to use one of them to supply the master clock instead shunt regulator. I have read from other thread in this forum that op-amp based super regulator is not to be worth in some critical circuit. What's your experience about this issue?
Thank you.
Hi -ecdesigns-,
Thanks for your previous "compensation" tip for opamps and the AD link: extremely interesting. I am using your noise manipulation advice and it works great, as all your advices do (I have to buy some metal-film resistors, though, to optimize it).
Hi Ims,
It appears as superregs could oscillate with low ESR bypass caps downstream, so one of two option: or you don't install bypass caps near components and power from SR (at your own risk), or, preferably, use stock power supply. I recommend try first stock option.
Cheers,
M
Thanks for your previous "compensation" tip for opamps and the AD link: extremely interesting. I am using your noise manipulation advice and it works great, as all your advices do (I have to buy some metal-film resistors, though, to optimize it).
Hi Ims,
It appears as superregs could oscillate with low ESR bypass caps downstream, so one of two option: or you don't install bypass caps near components and power from SR (at your own risk), or, preferably, use stock power supply. I recommend try first stock option.
Cheers,
M
Ubuntu vs XP
Hi all
This seems a bit off topic BUT :
Ubuntu (studio) vs XP (pro) on a laptop ----> DAC DI8M (usb or spdif)
Ubuntu wins far out ! Why ???
And then :
Disable in device manager both 2 drivers under "battery" . ( ac driver and acpi battery manager)
Who wins now ?? In my case XP on four (4) different laptops !!
I use kernel streaming usb -foobar or spdif with a vxpocket pcmcia card on an external clock)
Pse try this out and let us now if gates or I suck.
Regards
Onno
Hi all
This seems a bit off topic BUT :
Ubuntu (studio) vs XP (pro) on a laptop ----> DAC DI8M (usb or spdif)
Ubuntu wins far out ! Why ???
And then :
Disable in device manager both 2 drivers under "battery" . ( ac driver and acpi battery manager)
Who wins now ?? In my case XP on four (4) different laptops !!
I use kernel streaming usb -foobar or spdif with a vxpocket pcmcia card on an external clock)
Pse try this out and let us now if gates or I suck.
Regards
Onno
An interesting article about digital sources; PC V/S transport.
http://www.positive-feedback.com/Issue22/nugent.htm
We are on the right path.
Cheers,
M
http://www.positive-feedback.com/Issue22/nugent.htm
We are on the right path.
Cheers,
M
tubee said:
Tubee
Peter at XXhighend is about to issue pre-production version 9u which he thinks might be "a revolution in playback of digital sources" using the Vista operating system.
Whether Peter is right or wrong, it may be worth taking an open-minded approach to borrowing a Vista PC and trialling the software.
http://www.phasure.com/index.php?topic=365.msg2379;topicseen#msg2379
Cheers
Frank
Sorry for offtopicness.
I see a lot of people in a daily basis, i can tell you for sure they have a lot of comments about vista, it is not the appropiate upgrade on xp as it should be. The moment i let win slip into the ravine here at home was the best decicion i made on pc-matter. For instance i have on an old winxp ready machine 1.8G/ 1st sdram 756mb/ 128 nvidia astonishing 3d graphical effects vista can only dream of, running on linux then ofcoarse.
I see a lot of people in a daily basis, i can tell you for sure they have a lot of comments about vista, it is not the appropiate upgrade on xp as it should be. The moment i let win slip into the ravine here at home was the best decicion i made on pc-matter. For instance i have on an old winxp ready machine 1.8G/ 1st sdram 756mb/ 128 nvidia astonishing 3d graphical effects vista can only dream of, running on linux then ofcoarse.
tubee said:Sorry for offtopicness.
I see a lot of people in a daily basis, i can tell you for sure they have a lot of comments about vista, it is not the appropiate upgrade on xp as it should be. The moment i let win slip into the ravine here at home was the best decicion i made on pc-matter. For instance i have on an old winxp ready machine 1.8G/ 1st sdram 756mb/ 128 nvidia astonishing 3d graphical effects vista can only dream of, running on linux then ofcoarse.
I understand that that machine will run sluggish under Vista, but for the best possible SQ I dont think that linux can match it with either asio or xxhe.
I do not understand why such a firm like MS can earn so much money with unsatisfying OS stuff and programs. My kids have to work at school on w2k machine, because they are forced to do so. The teacher complains why she has to update all the adobe stuff every time. Whe have entered with in a new era by now, where in a company like MS is not acepted, at least not by me and a lot of others, but we stay all on win because we have learned to "work" with it. When i started with linux i discovered again the pc can work for me.
For the sound quality: there is an other thread here, http://www.diyaudio.com/forums/showthread.php?s=&threadid=93315&highlight=
I am sure for the earlyer mentioned reasons, it should be posseble to create an audiophile linux music server pc, with outstanding results. Better then apple (readed it in Audio german mag,) I have planned that too but never came so far.
For the sound quality: there is an other thread here, http://www.diyaudio.com/forums/showthread.php?s=&threadid=93315&highlight=
I am sure for the earlyer mentioned reasons, it should be posseble to create an audiophile linux music server pc, with outstanding results. Better then apple (readed it in Audio german mag,) I have planned that too but never came so far.
Hi xp9433,
Highest / ultimate sound quality is achieved with bit-perfect data and very low jitter. Manipulations in the digital domain are NOT required / desired for High-End applications.
There should be no significant sound differences between computer or OS, if the digital audio data is bit perfect, and jitter amplitude is very low.
About higher sample rates and bit depth:
With a given jitter amplitude (every clock has jitter), resolving higher sample rates / bit depth requires (very) low clock jitter amplitude.
Lets have a look at some examples with an ultra low jitter clock of 2ps RMS, and see what sample rate / bit depth combinations pass. I used the calculation method suggested by Kusunoki.
44.1 KHz/16 bits NOS: 1 / 44,100 / 2^16 / 2 = 173ps (OK)
44.1 KHz/16 bits at 8 * OS: 1 / (44,100 * 8) / 2^16 / 2 = 21.6ps RMS (OK)
48 KHz/16 bits at 8 * OS: 1 / (48,000 * 8) / 2^16 / 2 = 19.8ps RMS (OK)
> 44.1 KHz/19 bits at 8 * OS: 1 / (44,100 * 8) / 2^19 / 2 = 2.7ps RMS (OK)
44.1 KHz/24 bits at 8 * OS: 1 / (44,100 * 8) / 2^24 / 2 = 84.4fs or 0.084ps RMS (FAIL)
96 KHz/24 bits at 8 * OS: 1 / (96,000 * 8) / 2^24 / 2 = 38fs or 0.038ps RMS (FAIL)
96 KHz/19 bits at 8 * OS: 1 / (96,000 * 8) / 2^19 / 2 = 1.2ps RMS (FAIL)
96 KHz/18 bits at 8 * OS: 1 / (96,000 * 8) / 2^18 / 2 = 2.48ps RMS (PASS)
192 KHz/24 bits at 8 * OS: 1 / (192,000 * 8) / 2^24 / 2 = 19.4fs or 0.019ps RMS (FAIL)
192 KHz/19 bits at 8 * OS: 1 / (192,000 * 8) / 2^19 / 2 = 0.62ps RMS (FAIL)
So these examples show how difficult it is to resolve higher bit depth / sample rate. It also shows that it takes approx. 100 times lower clock jitter than the excellent Tent clocks to resolve 24 bit / 192 KHz. Then there is the problem of keeping jitter that low during distribution of the clock signal.
Both SPDIF and USB interface receivers have specified clock jitter between 75 and 200+ ps RMS. If no reclocker or slaved transport is used, this barely qualifies to drive a 16 bit 44.1 KHz NOS DAC (173ps RMS).
> The DI8M has 352.8 KHz sample rate and max. 19 bits resolution. Clock jitter needs to be below 2.7ps RMS, this can be achieved with real-world clocks and careful tuning.
Highest / ultimate sound quality is achieved with bit-perfect data and very low jitter. Manipulations in the digital domain are NOT required / desired for High-End applications.
There should be no significant sound differences between computer or OS, if the digital audio data is bit perfect, and jitter amplitude is very low.
About higher sample rates and bit depth:
With a given jitter amplitude (every clock has jitter), resolving higher sample rates / bit depth requires (very) low clock jitter amplitude.
Lets have a look at some examples with an ultra low jitter clock of 2ps RMS, and see what sample rate / bit depth combinations pass. I used the calculation method suggested by Kusunoki.
44.1 KHz/16 bits NOS: 1 / 44,100 / 2^16 / 2 = 173ps (OK)
44.1 KHz/16 bits at 8 * OS: 1 / (44,100 * 8) / 2^16 / 2 = 21.6ps RMS (OK)
48 KHz/16 bits at 8 * OS: 1 / (48,000 * 8) / 2^16 / 2 = 19.8ps RMS (OK)
> 44.1 KHz/19 bits at 8 * OS: 1 / (44,100 * 8) / 2^19 / 2 = 2.7ps RMS (OK)
44.1 KHz/24 bits at 8 * OS: 1 / (44,100 * 8) / 2^24 / 2 = 84.4fs or 0.084ps RMS (FAIL)
96 KHz/24 bits at 8 * OS: 1 / (96,000 * 8) / 2^24 / 2 = 38fs or 0.038ps RMS (FAIL)
96 KHz/19 bits at 8 * OS: 1 / (96,000 * 8) / 2^19 / 2 = 1.2ps RMS (FAIL)
96 KHz/18 bits at 8 * OS: 1 / (96,000 * 8) / 2^18 / 2 = 2.48ps RMS (PASS)
192 KHz/24 bits at 8 * OS: 1 / (192,000 * 8) / 2^24 / 2 = 19.4fs or 0.019ps RMS (FAIL)
192 KHz/19 bits at 8 * OS: 1 / (192,000 * 8) / 2^19 / 2 = 0.62ps RMS (FAIL)
So these examples show how difficult it is to resolve higher bit depth / sample rate. It also shows that it takes approx. 100 times lower clock jitter than the excellent Tent clocks to resolve 24 bit / 192 KHz. Then there is the problem of keeping jitter that low during distribution of the clock signal.
Both SPDIF and USB interface receivers have specified clock jitter between 75 and 200+ ps RMS. If no reclocker or slaved transport is used, this barely qualifies to drive a 16 bit 44.1 KHz NOS DAC (173ps RMS).
> The DI8M has 352.8 KHz sample rate and max. 19 bits resolution. Clock jitter needs to be below 2.7ps RMS, this can be achieved with real-world clocks and careful tuning.
This post is such a great source of idea
I still work with my single CS8414/1541A/SAA7220 + D1 DAC to learn.
I implemented the DEM Clock. I understood that the attenuator (680/180) was here to give 1 Vrms pp. I tried to remove it and let only the 470 pF caps... This has effect on sound mainly louder bass and so more bodied sound. So the question was why 1 VRms should be used ?
I know fit a DIR9001 and is planning to use the last LRM as described here too. With the SAA7220, i will have BCK à 5.6448 MHz which is not at least a multiple of 48 Mhz... To keep the same ratio (x17) will I have to use some 96 Mhz Xo ?
I still work with my single CS8414/1541A/SAA7220 + D1 DAC to learn.
I implemented the DEM Clock. I understood that the attenuator (680/180) was here to give 1 Vrms pp. I tried to remove it and let only the 470 pF caps... This has effect on sound mainly louder bass and so more bodied sound. So the question was why 1 VRms should be used ?
I know fit a DIR9001 and is planning to use the last LRM as described here too. With the SAA7220, i will have BCK à 5.6448 MHz which is not at least a multiple of 48 Mhz... To keep the same ratio (x17) will I have to use some 96 Mhz Xo ?
I have published now the my new DAC better than any TDA1541
http://www.audiodesignguide.com/DacEnd/index.html
http://www.audiodesignguide.com/DacEnd/index.html