Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › soundcard clock modding to improve S/PDIF jitter?
New Posts  All Forums:Forum Nav:

soundcard clock modding to improve S/PDIF jitter?  

post #1 of 51
2/21/10 at 1:01am
Thread Starter 
hi there,

there's many cheap CMI8738/8768 cards around, and they all work w/ these amazing bit-perfect/automatically bitmatched windows drivers(KS/WASAPI/etc): cmediadrivers

all I care for atm is TOSLINK, and I see that my DAC has a DIR9001 chip(given for 50ps clock recovery).

I've got a bunch of these cards around, and they all carry a 14.18318Mhz clock...one of them has "NSK 5J 14.31818AXN" written on the chip, but I haven't been able to find out what its resolution is on the NSK website...prolly 50/100ppm anyway.



could I swap these clocks for this 10ppm version as a drop-in replacement? Digi-Key - 887-1241-ND (Manufacturer - 9B-14.31818MEEJ-B)

would that improve anything whatsoever over TOSLINK? jitter is measured in ps, so what does the ppm resolution have to do w/ anything? OTOH these crystals cost $0.5 a pop..

and I see that this chip has a "load capacitance" of 18 pF, is that bad?

M2Tech says that jitter is so bad over TOSLINK that they decided to go coax only...so would that be futile

thanks!
post #2 of 51
2/21/10 at 1:15am
wow, this opens a whole can of worms if you must know.

I'd recommend two things, for what they're worth.

Ignore jitter specifications and don't use toslink.
post #3 of 51
2/21/10 at 1:21am
Thread Starter 
well, I need toslink....because as I've recently been told: "toslink has the advantage of galvanically isolating sender and receiver, and this is why I prefer it over coax."

I've got uber-sensitive headphones and computer ground loops must cease...I've used toslink in the past to transfer DAT tapes, and it was fine

I see M2Tech hates toslink, but Musiland has it on all their "Monitor" USB devices.
post #4 of 51
2/21/10 at 1:24am
Quote:
Originally Posted by leeperry View Post
galvanically isolating sender and receiver,
Isn't that why people use transformers in their coax signal chain, to galvanically isolate?
post #5 of 51
2/21/10 at 1:27am
Thread Starter 
Quote:
Originally Posted by cobaltmute View Post
Isn't that why people use transformers in their coax signal chain, to galvanically isolate?
apparently it's always done on AES/EBU, but often skipped on coax S/PDIF to save costs

http://www.google.com/search?hl=en&s...&aq=f&aqi=&oq=
Quote:
ADAT and S/PDIF optical both use fibre-optic cables, giving true galvanic isolation, just like a transformer, so they can never result in ground loops.
http://www.kdvelectronics.eu/DVB-J2/DVB-S_J2.html
Quote:
Optical SPDIF output (Toslink) was chosen instead of coax to have galvanic isolation between the set-top box and any audio equipment
post #6 of 51
2/21/10 at 8:09am
Thread Starter 
ok, anyway I see that crystals age around 5/10ppm per year, and some of these cards are 4-5 years old...so going from 100/150ppm to 10ppm should be audible I think

still curious about the "load capacitance" spec, but I guess that'd work fine.
post #7 of 51
2/21/10 at 8:11am
Toslink is worse off than coax, so there is no point in upgrading your clock if you are going to shove it down some crappy plastic

14.18318Mhz isn't an integer ratio of any of the audio sample rates, so there has to be some stuff going on to make it to the same rate and that is going to destroy any jitter improvements you add to the crystal

The accuracy of the crystal has nothing to do with jitter performance of it, so don't look at the ppm readings as they are useless.

Now lets say that you did go to the trouble of getting a good clock at the right frequency to output spdif over coax, you now need to modify the output to be a true 75ohms output and you can't use RCA's, it needs to be a real 75ohm BNC or you are just wasting your time.

You still want to go all the way with this? I hope I haven't crushed all your dreams so far.
post #8 of 51
2/21/10 at 8:22am
This is what you need: Antelope Audio
post #9 of 51
2/21/10 at 8:26am
And this: Antelope Audio
post #10 of 51
2/21/10 at 8:31am
Thread Starter 
Quote:
Originally Posted by DaKi][er View Post
14.18318Mhz isn't an integer ratio of any of the audio sample rates, so there has to be some stuff going on to make it to the same rate and that is going to destroy any jitter improvements you add to the crystal

The accuracy of the crystal has nothing to do with jitter performance of it, so don't look at the ppm readings as they are useless.

Now lets say that you did go to the trouble of getting a good clock at the right frequency to output spdif over coax, you now need to modify the output to be a true 75ohms output and you can't use RCA's, it needs to be a real 75ohm BNC or you are just wasting your time.
ok, you nailed it about the crystal frequency! the DSP ASIC must be doing some clock conditioning of some sort.

hah, so now even RCA coax is worthless? BNC or nothing

I see my DAC has a "Virtual-Ground power supply": http://www.firestone.idv.tw/cgi-bin/...sp?pdtseqnm=27

so maybe they're actually filtering the S/PDIF coax ground...I'll run some tests from the mobo coax RCA out before getting a M2Tech Hiface or Musiland Monitor 01USD of some sort.
post #11 of 51
2/21/10 at 11:04am
Thread Starter 
http://www.head-fi.org/forums/f46/di...nswers-352889/
post #12 of 51
2/21/10 at 11:51am
I don't think anyone any DIY forum is going to hold much stock in an argument in a computer audio forum at Head-Fi. Plus, some of the most notorious vapid posters here at Head-Fi weighed in that thread.

Anyways, just because people don't something to cut cost, doesn't mean it isn't the best way to do things. I mean, look what you're suggesting, replacing a clock with a more expensive / "better" one.

Furthermore, it's been my experience that topologies changes bring about better results than blind part swapping.
post #13 of 51
2/21/10 at 1:38pm
Thread Starter 
ouh there's a lot of missing words in your last post I'm afraid...anyway, ok I'll ask my DAC manufacturer whether they've added a transformer in the coax signal chain...and if they actually did, I might grab a M2tech Hiface

I still wonder why many firewire/USB devices don't have this kind of ground filtering....

OTOH, Bob Katz thinks that optical is better than coax hah.. http://www.head-fi.org/forums/f21/pl...2/#post1812131
Quote:
"It [glass cable] also has superior bandwidth and therefore causes fewer interface jitter problems, jitter as low as as any good copper connection. Glass fibre connections can have even lower interface jitter than unbalanced copper connections becuase they eliminate ground loops and EMI sensitivity."
and tbh, in the pro audio world, most ppl will pick toslink over coax
post #14 of 51
2/21/10 at 2:04pm
I've found the best way to find out something is to do it yourself and listen for yourself. The differences in hearing and perception are as varied as recording quality and tastes in music.
post #15 of 51
2/21/10 at 4:36pm
Thread Starter 
well, most likely only the high end stuff runs transformers on the coax, so consumer grade gear would give groundloops from a computer...so I still stand by toslink

and indeed, I guess switching my soundcard clock to a 10ppm chip prolly wouldn't matter...I still have asked for samples, if they ever show up I might swap them for the hell of it

I know some ppl modded the Asus STX clock to a 1ppm chip and claimed improvements over S/PDIF...OTOH it was a 24.576Mhz clock(which divided by 512 gives 48kHz)

6moons.com - audio reviews: A Toslink vs. RCA digital cable comparison

LITE AUDIO DAC AM REVIEW
Quote:
The Mitsubishi POF cable sounded dull and lifeless in comparison to the BlueJeans coax. I spoke to the folks at Pacific Valve and they said that they try to use good quality coax with this unit as they don’t like to invest in expensive optical cables. I found a glass optical cable and tried it out. I then had a series of double blind shootouts with these cables. The short version is that the BlueJeans coax and a Dayton GOC-3 from PartsExpress came out the clear winners with the number of folks preferring the Dayton cable beating the BlueJeans by a margin smaller than the margin of error. There was a difference but folks could not agree on which was better. In my opinion, the Dayton glass optical cable was a little smoother and warmer than the coax but with better sounding highs and imaging. Others preferred the livelier BlueJeans coax but nobody felt real strongly about the difference between these too. We all agreed that the POF cable was the loser of the bunch in this setup.
New Posts  All Forums:Forum Nav:
  Return Home
This thread is locked  
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › soundcard clock modding to improve S/PDIF jitter?