function and influence of an external word clock (master clock)?

0 Members and 1 Guest are viewing this topic. Read 3704 times.

kyrill

Hi Steve
Once you were a kind of DEQX expert.. :)
I hope you may answer my questions as I think a digital expert is needed and maybe you can point me to advises.
I posted these question first in the LAB section but I presume tio have here more "digital knowledge"
If it is inappropriate for this forum I expect you to feel free to remove this post.

I find the deqx PDC-2.6 though to have some minor min points after their HD upgrade. Their internal analogue side can still be fully updated with better film output caps (Teflon for instance)
and I am already years searching for a better clock. in order to remove jitter. I updated the clock finally last year in the USA with a precision clock  and dedicated own pws. It made the sound analogue to me and some more. 
Then 7 months later the unit did not finish its start up sequence. A DEQX  technician almost got angry at me and blamed the new clock for the problems. He removed the superclock and got the original clock  back, a clock based on ideas about clocks 8 years ago ( that is how "old" my DEQX is) . I still consider it very strange to blame the new clock, as the unit worked flawlessly for almost 8 months, 5 days a week, 2-5 hrs a day.

So my question is, is it possible if you see on the board a canned crystal clock, to only replace it with a better clock ( same oscillation freq. of course) without  affecting the rest of the schematics as if the system still "thinks" it is the original clock only more precise? This at least is suggested by all those "replacement clock manufacturers. But what about Z-in and Z-out of the clock compared to the replacement "module"? Does that play a role?

However I cannot leave the analogue musicality of the DEQX behind by accepting DEQX standard clock solution. ( which was acceptable good but not to the standards of for instance http://www.newclassd.com/index.php?page=36 with their super clock and ditto pws

So I though maybe a whole different solution is DEQX digital out board, which bypasses the PDC 3 internal dacs and analogue volume control to be replaced by 3 external dacs. This digital out board has  an AES "IN" for an external word clock. I assumed faulty that this outer clock would switch off the on board clock and take over. According to this technician this external link is only used to "reclock" the output  to the ext. freq. So if DEQX processes a 16 bit 44.1khz red book file, and the ext clock is set to 96 khz, the output of the DEQX will be this 96 khz 24 bit. At least this is how I interpret the technician's message:
"Kyrill,

The sync input [of the digital board] has nothing to do with the PDC  clocking!!! The sync is only to control the sample-rate convertors of the Digital output so you can set the sample-rate I.e. 44.1, 48, 88.2 or 96k "

I always thought when an ADC or DAC accepts an external word clock, it also accepts the precision or lack of it of that particular ext. clock. Is that right? To what extent means "..to control the sample-rate converters"?
Is the DEQX only reading which freq (44.1-96khz) but it resamples it on the the board independent of the quality of the ext. clock? the technician 's email reads that way.. If so, the jitter quality of the external word clock  has no influence whatsoever on the jitter output of  DEQX re-samplers?

I hoped that if the last stage prior to DEQX's  output was fully controlled by an ext clock, then by choosing a high quality clock to have less jitter at DEQX's output and so can leave the internal PDC clock alone.

All questions related on how to precise the clock freq for DEQX in order to lower the accumulated jitter.

Any ideas?
(3 Pace clockers for the DEQX 3 digital outputs (bass/mid/highs) would solve my problems, but beyond my budget)

audioengr

Hi Steve
Once you were a kind of DEQX expert.. :)
I hope you may answer my questions as I think a digital expert is needed and maybe you can point me to advises.
I posted these question first in the LAB section but I presume tio have here more "digital knowledge"
If it is inappropriate for this forum I expect you to feel free to remove this post.

I find the deqx PDC-2.6 though to have some minor min points after their HD upgrade. Their internal analogue side can still be fully updated with better film output caps (Teflon for instance)
and I am already years searching for a better clock. in order to remove jitter. I updated the clock finally last year in the USA with a precision clock  and dedicated own pws. It made the sound analogue to me and some more. 
Then 7 months later the unit did not finish its start up sequence. A DEQX  technician almost got angry at me and blamed the new clock for the problems. He removed the superclock and got the original clock  back, a clock based on ideas about clocks 8 years ago ( that is how "old" my DEQX is) . I still consider it very strange to blame the new clock, as the unit worked flawlessly for almost 8 months, 5 days a week, 2-5 hrs a day.

So my question is, is it possible if you see on the board a canned crystal clock, to only replace it with a better clock ( same oscillation freq. of course) without  affecting the rest of the schematics as if the system still "thinks" it is the original clock only more precise? This at least is suggested by all those "replacement clock manufacturers. But what about Z-in and Z-out of the clock compared to the replacement "module"? Does that play a role?

There are a few issues with installing/replacing a clock like this.  Most novice DIYers don't have al of the education/experience to do this right.  It really requires a fast scope (1GHz min) to verify the result:

1) the power and the clock output on the clock module must be kept separate, usually two twisted-pairs or a twisted-pair and a coax cable
2) impedance matching to the cable is important.  You cannot simply drop the clock in the circuit.  You may have to adjust a series resistive terminator on the module and then add a parallel resistive terminator at the designation DEQX board.  The Terminations must match the coax or twisted-pair cable being used.  If you use a 50 ohm coax, then the parallel teermination will be 50 ohms to ground.  If it is also a series termination, then the drive output impedance, usually around 10 ohms plus the resistor must equel 50 ohms.  Twisted-pairs are usually 90-110 ohms charachteristic impedance.
3) signal connection points on the DEQX board are critical, both for the power and ground.  They must be close to the load device and the ground wire must be connected very close.  The parallel terminator can be connected after the load pin, but not by much.  Series resistors on the module must be within .1" of the driver pin.
4) the Signal Level required by the DEQX circuit must be the same.  If the clock is driving a 3.3V circuit, then the clock must have 3.3V voltage levels.  If it is diring a 5V circuit, then the signal level must be 5V.  If there is both series and parallel termination, this will divide the signal in half, so you can usually start with 5V signal and then end-up with 2.5V driving a 3.3V circuit with good results.

Quote
However I cannot leave the analogue musicality of the DEQX behind by accepting DEQX standard clock solution. ( which was acceptable good but not to the standards of for instance http://www.newclassd.com/index.php?page=36 with their super clock and ditto pws

So I though maybe a whole different solution is DEQX digital out board, which bypasses the PDC 3 internal dacs and analogue volume control to be replaced by 3 external dacs. This digital out board has  an AES "IN" for an external word clock. I assumed faulty that this outer clock would switch off the on board clock and take over. According to this technician this external link is only used to "reclock" the output  to the ext. freq. So if DEQX processes a 16 bit 44.1khz red book file, and the ext clock is set to 96 khz, the output of the DEQX will be this 96 khz 24 bit. At least this is how I interpret the technician's message:
"Kyrill,

AES/EBU input on the DEQX is simply a pro digital input, not a word-clock or master-clock input.

Quote
The sync input [of the digital board] has nothing to do with the PDC  clocking!!! The sync is only to control the sample-rate convertors of the Digital output so you can set the sample-rate I.e. 44.1, 48, 88.2 or 96k "

The input signals have the basic clocks imbedded.  There is an internal high-speed clock that synchronizes to them using software buffering.  This is why changing that crystal osc. improves the SQ.

Quote
I always thought when an ADC or DAC accepts an external word clock, it also accepts the precision or lack of it of that particular ext. clock. Is that right?

A clock is always sent to a D/A chip, whether it's enbedded in the data or a discrete seperate clock.  This clock is the master clock, usually 256 times the frequency of the word-clock.  Sometimes the manuf. puts the master clock inside the DAC and has a word-clock output divided down from this clock.  This can usually be selected optionally.  Control of the frequency of the source clock is required with this internal clock, so this is not used much as there are not too many source devices that allow control of their clocks.

Quote
To what extent means "..to control the sample-rate converters"?
Is the DEQX only reading which freq (44.1-96khz) but it resamples it on the the board independent of the quality of the ext. clock? the technician 's email reads that way.. If so, the jitter quality of the external word clock  has no influence whatsoever on the jitter output of  DEQX re-samplers?

I am not aware that the DEQX upsamples.  You can set it for 48 or 96.  A signal at 44.1 stays at 44.1.  A signal at 96 stays at 96.  If it is set for 96 and you send it 44.1, no music.

Quote
I hoped that if the last stage prior to DEQX's  output was fully controlled by an ext clock, then by choosing a high quality clock to have less jitter at DEQX's output and so can leave the internal PDC clock alone.

You were on the right track.  I think it was maybe a implementation problem only.

Quote
All questions related on how to precise the clock freq for DEQX in order to lower the accumulated jitter.

Nothing to do with precision.  Low jitter clocks are designed to be low jitter, but not always accurate.  Accuracy is not particularly important in audio streaming.


Quote
(3 Pace clockers for the DEQX 3 digital outputs (bass/mid/highs) would solve my problems, but beyond my budget)

I used this at RMAF2008.  Worked pretty well, although I discovered that even with flat response, the DEQX adds compression.  Their DSP algorithm could be improved I think.  Volume control of 6 channels is a bit of a challenge.  We did it partly with analog and partly with digital.  Here is the 2-way system that I used at RMAF:





Steve N.
« Last Edit: 25 Sep 2011, 08:40 pm by audioengr »

kyrill

thx very much
Steve

I would never try doing this myself replacing clocks and so on

yr  2-way solution looks impressive.. so how was it received?

I see on yr schematics you used DEQX spdif out, only DEQX's  digital out board can do that. Only on that board are 6 spdif out; their info on this is:
"The digital output module provides three stereo AES/EBU outputs via three XLR connectors, while a fourth female XLR will accept an AES sync input that will asynchronously sync all digital outputs to the input clock reference up to 100kHz sample rate, regardless of the PDC’s input or processing sample rate. The unit also provides three stereo S/PDIF outputs via RCA connectors."

reading this it really suggests that a possible (very) low jitter quality of the out board word clock will not positively affect DEQX re-sampling quality. Am I wrong or pessimistic? What does it really means: to the input clock reference only the freq. or next to it the jitter quality as well?

audioengr

thx very much
Steve

I would never try doing this myself replacing clocks and so on

yr  2-way solution looks impressive.. so how was it received?

It was so complicated that people, even reviewers had no idea what was going on.  Sounded good though.

Quote
I see on yr schematics you used DEQX spdif out, only DEQX's  digital out board can do that. Only on that board are 6 spdif out; their info on this is:
"The digital output module provides three stereo AES/EBU outputs via three XLR connectors, while a fourth female XLR will accept an AES sync input that will asynchronously sync all digital outputs to the input clock reference up to 100kHz sample rate, regardless of the PDC’s input or processing sample rate. The unit also provides three stereo S/PDIF outputs via RCA connectors."

reading this it really suggests that a possible (very) low jitter quality of the out board word clock will not positively affect DEQX re-sampling quality. Am I wrong or pessimistic?

The digital output is quite jittery IMO.  It needed a reclocker to sound good.

Quote
What does it really means: to the input clock reference only the freq. or next to it the jitter quality as well?

The input signal jitter matters in all cases.  It affects SQ.  This is why I reclocked it and established a new master clock externally in one of the two Pace-Cars.

Using the 3 digital outputs from the optional board seems easy, but it is really a lot more frought with problems than you think.  Unless you are very technical, I would avoid it.  It is more straightforward to use the internal DACs and just drive the S/PDIF input with a really low jitter source like the Off-Ramp 4.  Mod the DACs and analog output sections if you want it to sound good.  I don't do these mods anymore, but you may be able to pick one up used on Audiogon if you watch carefully.

Steve N.

Steve N.

kyrill

thx a lot Steve :D

kyrill

Hi Steve
could Cullen circuits make the DEQX a slave to the Pace Car?
cullen circuits.com seems not to exist anymore? I get the ISP homepage Webwidgets...

audioengr

Hi Steve
could Cullen circuits make the DEQX a slave to the Pace Car?
cullen circuits.com seems not to exist anymore? I get the ISP homepage Webwidgets...

They cannot, and neither can I.  What you can do is use a Pace-Car after the DEQX and then a source device to drive DEQX with word-clock input, such as the Fireface400.

Steve N.