PDA

View Full Version : Modifying squeezebox clock



Neil Hastie
2005-03-02, 08:15
Hi All,

I have some further information regarding this problem.
Late last year I was trying to get a freinds MK1 Audio Synthesis DAX to lock
to my Squeezebox. We saw the same problem, A good lock with MP3 streams,
no lock at all with FLAC.

The Mk1 DAX has a very narrow PLL clock capture window specified as +/-
200ppm.
(Verified by the designer of the DAX).
Subsequent revisions of the DAX product line have multiple capture windows
to allow
locking to a wider range of transports, albeit at a lower quality.

The DAX2 has a dual stage lock window at +/- 200pmm and +/- 1000 (ppm).
The wider lock windows have been reported as working.

With help from Triode I tried tweeking the programmed crystal value to try
and find a
suitable lock (basically a modification to the pitch function) with no luck.

It is unclear to me why there should be a data dependent difference in SPDIF
clock accuracy.
(In fact activating a DSP to perform MP3 decoding would seem to be far more
likely to increase
jitter due to noise injection). The Micronas data sheet offers no insight
that I can find.

Recently another friend reported the same lock issue with the latest model
Cyrus DAC-X.
So it appears that this is not an issue confined to a single high-end
manufacturer.

I can guarantee the sale of another squeezebox if this issue can be resolved
!!

Neil


>From: "Julian Alden-Salter" <julian.alden-salter (AT) ntlworld (DOT) com>
>Reply-To: Slim Devices Discussion <discuss (AT) lists (DOT) slimdevices.com>
>To: "'Slim Devices Discussion'" <discuss (AT) lists (DOT) slimdevices.com>
>Subject: [slim] Modifying squeezebox clock
>Date: Wed, 2 Mar 2005 11:11:31 -0000
>
>Hi,
>This is the stuff I found on audio synthesis' website...
>
>Following input selection two PLLs are used to extract and purify the word
>clock. An electrically and mechanically isolated crystal oscillator forms
>part of a sophisticated second order PLL, tightly bandpass filtering the
>recovered clock before applying it to the digital filter and DACs. A third
>digital loop is used to isolate jitter generated in the digital filter -
>all
>DAC loading and sample & hold timing is derived from this third digital
>PLL.
>
>Inputs are reclocked on multiple occasions before reaching the DAC itself.
>Low frequency jitter is the most difficult to eliminate so the three PLLs
>have been carefully optimised to boast a sub 1Hz jitter cut off frequency
>With the aid of multiple PLLs and intelligent muting we have made locking
>to
>jitter-ridden or off frequency transports routine, whilst maintaining
>automatic silent crystal locking for higher quality sources.
>
> >From here: http://www.audiosynthesis.co.uk/dax_decade.htm
>
>So much phlogiston to me though.
>
>I've actually upgraded recently to the decade from a dax-2 from the same
>company - this previous dac also exhibited similar behaviour with mp3 /
>pcm/flacs.
>
>I take on board the buffer stuff phil mentioned but I don't get stuttering
>and can't see how an intermittent buffer underrun would cause the dac to
>fail to xlock (crystal lock from above).
>
>Cheers
>
>
>Julian.
>
>
>
>-----Original Message-----
>From: Sean Adams [mailto:sadams (AT) slimdevices (DOT) com]
>Sent: 02 March 2005 05:09
>To: Slim Devices Discussion
>Subject: [slim] Modifying squeezebox clock
>
>
>On Mar 1, 2005, at 8:08 PM, Phil Karn wrote:
>
> > Julian Alden-Salter wrote:
> >
> >> 1) The fact that my dac locks on with different qualities of lock
> >> when mp3
> >> and flacs are played back. Suggesting that there is indeed some
> >> difference
> >> in the spdif data stream between the two formats.
> >>
>
>There is a "just for your information" bit for clock precision in the
>s/pdif channel status data - that may be what it is. If so, it has
>nothing to do with actual clock precision, it's just a bit that says "I
>think I'm a high precision clock" or not. I'm not sure what we send
>for this bit but I could check it on my analyzer - at any rate, it does
>not affect functionality/performance in any way.
>
>Does your receiver (maybe in the instruction manual) say exactly what
>it's reporting?
>
>
>
>
>
>

Phil Karn
2005-03-02, 15:31
Neil Hastie wrote:
> Hi All,
>
> I have some further information regarding this problem.
> Late last year I was trying to get a freinds MK1 Audio Synthesis DAX to
> lock
> to my Squeezebox. We saw the same problem, A good lock with MP3 streams,
> no lock at all with FLAC.

Are you *sure* you aren't seeing any buffer underruns in FLAC/WAV mode?
The buffer is a lot shorter in raw PCM mode, which makes underruns far
more likely than in MP3 mode. Even a brief underrun might cause a glitch
in the S/P-DIF clock that would case your external DAC to have trouble
re-locking for a while, especially given its tight PLL loop.

Or you could simply connect the analog outputs on your Squeezebox
directly to your amplifier and enjoy the music. :-)

Phil