word clock vs superclock 3 in modded squeezebox
Mar 22, 2005 at 1:32 AM Thread Starter Post #1 of 35

kmcdonou

100+ Head-Fier
Joined
Oct 15, 2004
Posts
376
Likes
10
I am considering getting the new Squeezebox modded. One guy in France who does these mods uses a word clock in the Squeezebox. I am know there are different types of clocks, but I do not understand the differences between them. I suspect if I get this modded in the states and wanted a clock upgrade they would use a superclock 3.

What are the differences between using a word clock in the SB vs a Superclock 3. Would one give significantly different performance over the other? When I have asked the guy in France (Dan Belitty) about his clock he responded "the input clock system we use in SB is our own development. (All is done in a microprocessor we program). We use a buffer multiplicator to use an external clock reference. We do not try to improve clock accuracy but to use the SAME clock signal for all digital unit of the audio system."

I am also curious as whether a clock upgrade would significantly improved the performance of the new SB 2 (when fed into a Ack Dack 2.0)? Sean Adams doesn't seem to think I would need a clock upgrade based on their new design. Just curious what the collective thought.
 
Mar 22, 2005 at 2:15 AM Post #2 of 35
Quote:

Originally Posted by kmcdonou
I am considering getting the new Squeezebox modded.



kmcdonou,

What's the goal here? Are you trying to get higher clock precision, or do you need to add external word clock (aka "house sync") input?
 
Mar 22, 2005 at 2:30 AM Post #3 of 35
My goal is to get world class sound from a $250 box.
icon10.gif


I obviously am not an expert on modding equipment, but I generally see people getting significant improvements in their CDP (when used as a transport) through upgrades in the PSU, digital output, and clock input.

Two different people who will mod the SB use different clock techniques. I am just trying to find out the difference between them and whether one would give better performance over the other (or are any of them needed now with the improved design)?

So, to repeat myself, my goal is to get the best possible sound (within $1000) from the SB 2. If the new design would not significantly benefit from a clock mod, then so be it.
 
Mar 22, 2005 at 5:41 AM Post #5 of 35
What DAC are you trying to drive with the Squeezebox?

If you have a DAC that outouts a word clock or you want to drive both the DAC and the digital output of the SB with an external wordclock than this will give you jitter quality as good as the word clock.

A superclock will simply improve the digital output which might or might not give you better sound depending on your DAC.

If someone can really mod the SB to have a wordclock input then a combination with a Univeral Audio 2192 DAC could be a killer combination. I have no experience with the SB but at least the digital output to DAC would be world class.

Cheers

Thomas
 
Mar 22, 2005 at 7:36 AM Post #6 of 35
Quote:

Originally Posted by kmcdonou
My goal is to get world class sound from a $250 box.
icon10.gif



OK, I hear you.
smily_headphones1.gif


My first advice is listen to it first. Compare it (double-blind) to the most expensive CD transport you can find, and decide what you want to improve. This is going to come off as a bit cocky, but I really don't think there's a lot of room for improvement of the s/pdif output quality in Squeezebox2. Dollars don't buy better sound per se - it comes from careful design and extensive testing.

In Squeezebox2 we are using fixed-frequency oscillators, one dedicated for 44.1 and one dedicated for 48Khz. In our new design, the only source of drift is in the oscillator itself, since all the logic is driven directly from these clocks. That means the only room for improvement is in replacing these clocks, which is easy to do if you want to. Just disconnect the crystal and attach an aftermarket clock. The problem is, you will need a $50,000 timing analyzer to measure the difference (at best a couple picoseconds of deviation). Otherwise you'll have to trust your ears that the $400 spent on the upgrade was worth it.

If you want my advice on the best hack, I would suggest eliminating the s/pdif transport. The _receiving_ end of the s/pdif line is where the clock is recovered from the manchester-encoded waveform, and this is the only really measurable source of clock noise in a "digital source -> offboard DAC" system. You could do this in Squeezebox2 by tapping the MCLK/BICK/SDATA/LRCK lines going into our DAC, and running them over short lengths of coax directly into your DAC's DAC chip (bypassing its s/pdif receiver).

Or you could just use our built-in DAC, which is a decent one. I have verified better-than-data-sheet performance (96-97dB THD+N) from this DAC in our actual hardware.

Again though, I would make sure you're equipped to measure any changes, and certainly I would identify what you want to improve before changing anything.
 
Mar 22, 2005 at 7:43 AM Post #7 of 35
What is wrong then with a masterclock input? This way the S/PDIF receiver in the DAC does not have to recover the clock from the Spueezebox output however good or bad it might be?

At this point the quality will only be determined by the quality of the clock in the DAC. You are not really proposing to bypass the DACs clock and drive it via the Squeezebox clock output signal?

Cheers

Thomas
 
Mar 22, 2005 at 7:58 AM Post #8 of 35
Quote:

Originally Posted by thomaspf
You are not really proposing to bypass the DACs clock and drive it via the Squeezebox clock output signal?


What DAC's clock??

The DAC does not drive the clock at all, the source does. The DAC is clocked by its s/pdif receiver, which recovers the clock generated by the source (ie Squeezebox2).

I'm suggesting that the "lowest hanging fruit" in terms of clock improvement is to bypass this clock recovery process (the s/pdif receiver) by driving the clock straight through. That's all.
 
Mar 22, 2005 at 8:17 AM Post #9 of 35
thomas is talking about having the master clock in the outboard DAC and feeding word clock derived from it into sqeezebox in order for the squeezebox to lock on it, hence you don't have to use recovered clock in the outboard DAC, just receive data from squeezebox that is in sync to the outboard DAC's own masterclock.. this way you eliminate running masterclock over long run of coax cable.. it doesn't have to be word clock per se, much better and easier approach is to put S/PDIF transmitter to the outboard DAC and S/PDIF receiver into squeezebox and use the recovered clock from s/pdif input instead of those fixed oscillators you have there.. again you're in sync and the jitter is limited just to the implementation of clocking circuitry in the outboard DAC itself, which can be done very well considering the proximity of it's own oscillator to it's DAC chip..
 
Mar 22, 2005 at 8:20 AM Post #10 of 35
Quote:

Originally Posted by Glassman
thomas is talking about having the master clock in the outboard DAC and feeding word clock derived from it into sqeezebox in order for the squeezebox to lock on it, hence you don't have to use recovered clock in the outboard DAC, just receive data from it that is in sync to the outboard DAC's own masterclock..


What DAC supports this? (Not a rhetorical question - I'm honestly asking)
 
Mar 22, 2005 at 8:22 AM Post #11 of 35
Quote:

Originally Posted by seanadams
What DAC supports this? (Not a rhetorical question - I'm honestly asking)


BTW more useful than a word clock would be the "master clock" - 256 x Fs or something.
 
Mar 22, 2005 at 8:23 AM Post #12 of 35
pretty much all proffesional DACs do use word clock syncing as well as proffesional soundcards do.. but yes there is no consumer device of such kind because there is no appropriate source device to connect to this way.. but if someone's taking the modding route, one can do anything, even unusual but absolute solutions..
 
Mar 22, 2005 at 8:24 AM Post #13 of 35
Quote:

Originally Posted by seanadams
BTW more useful than a word clock would be the "master clock" - 256 x Fs or something.


read my post again, I've edited it a few seconds before you posting this
wink.gif
 
Mar 22, 2005 at 3:29 PM Post #14 of 35
Quote:

Originally Posted by Glassman
pretty much all proffesional DACs do use word clock syncing as well as proffesional soundcards do.. but yes there is no consumer device of such kind because there is no appropriate source device to connect to this way.. but if someone's taking the modding route, one can do anything, even unusual but absolute solutions..


I was just looking at the Universal 2192 you mentioned.... ok, so it has _internal_ clock sources from which you can send a word sync out to the source(s).

Anyway adding word clock inputs on Squeezebox2 is a relatively trivial matter of just driving the internal MCLK off of a multiplier (or drive MCLK directly if you have it). I say "relatively trivial" as in "relative to Squeezebox1" where this was not a straightforward hack due to the clock generation being embedded in the DSP chip.

Here's some more info about the new design which may be of interest:

http://article.gmane.org/gmane.music...tch=house+sync

Sean
 
Mar 22, 2005 at 5:32 PM Post #15 of 35
Wow, I guess that would be one of the cheapest digital sources with a word clock input.

A few more DACs with word clock output below. There are more with word clock input which then would require an external clock like the Apogee BigBen or the DCS Verona to drive both the Squeezebox and the DAC.

Universal Audio 2192 (mentioned above)
Apogee Digital Rosetta 200
Apogee Digital Rosetta 800
DCS Delius DAC
DCS Elgar Plus DAC
emmLabs DAC6
emmLabs DAC8
emmLabs DCC2

and I am sure I am missing others.

Cheers

Thomas
 

Users who are viewing this thread

Back
Top