Reclocking CDPro, TubeDac and I2S, Newbie asks questions!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all

Please excuse the following ramble and stream of thoughts and questions but not sure what to do next and would welcome any thoughts (dismissive or otherwise) and a little enlightenment. Be warned though, I am a complete novice with little experience wandering foolishly into territory I have no business in, but I'm afraid that doesn't seem to stop me and I'm curious about some of this digital stuff!

I just finished building a transport kit based on the Philips CDPro2 mechanism (not too difficult as display and power supply PCB's already made up). This is hooked up to the Scott Nixon Tubedac I built last year. I had no idea how the DAC worked when I built it (but now beginning to get a vague idea) but it worked first time and I am pleased with it. The CDPro transport replaces my ancient Marantz CD63KI, which is starting to show its age. It is definitely an improvement on the Marantz but not as much as I would like (I expected this would be the case before I started but was hell bent on building one anyway). I did though reclock the Marantz, with a Trichord clock II, which methinks must be giving it an advantage.

I am thinking therefore of reclocking the CDPro. I have seen some good instructions (I need clear instructions!) on the Homeoptics site for fitting a Tentlabs XO3 clock to the CDPro. This seems like a good clock for a reasonable price, does anyone have experience of this? The XO3 also seems to have a second reclocking circuit to improve the SPDIF output of the transport (not sure how this works, is it simply a second clock feed fed into the SPDIF output somehow?), which seems a good thing. I could therefore go ahead and fit the XO3. There is also an XO-DAC but this is getting expensive!

If I went this way, I thought I'd mount the CDPRo2 and DAC right next to each other, and replace the coax interconnect with v short unscreened leads. Tubedac presently has a 75ohm shunt resistor across its input, to match the characteristic impedance of the SPDIF input cable I believe. Is this resistor necessary if you use unscreened wire?

My attention is though drawn to the three pin I2S output on my CDPro2 (I believe the mechanism may have been modified slightly to provide this). Most seem to think this is a better interface than SPDIF which everyone seems to hate. I understand it has a separate clock output/feed for reduced jitter. Is this at the same frequency as the master clock in the transport?

Could I connect my Tubedac to I2S rather than SPDIF and would it be better? Some searches suggest I would need to bypass the CS8412 (does this chip essentially turn SPDIF into I2S anyway?) in Scotts DAC and go directly to the TDA1543. The data sheet indicates the TDA accepts I2S on pins 1-3. Will this work? is it this simple? does there need to be any decoupling (capacitor/transformer) like in SPDIF or can I simply wire the two together? Sorry so many questions!

Would the SPDIF reclocking circuit on the XO3 be redundant if I used I2S, and therefore should I buy the cheaper XO2 (or similar clock) that lacks this feature, or can it also improve the I2S output? I wonder though if the I2S clock feed comes straight from the master clock of the transport anyway, and is therefore going to be just as good?. When a transport is reclocked with a simpler clock such as the XO2, are there still benefits when using an external DAC. Methinks if the transports internal clock is better then surely the clock component of its SPDIF output has got to be better, even without the extra XO3 thingy.

Another possibility I understand is a master clock with two separate feeds (not sure if the X03 can be used in this way) to supply the DAC and transport separately, but do both need to have the same clock speed? I don't know the clock speed of the CDPro yet, and what is the clock speed of TubeDac? I've read that the CS8412 works on a 11.2896Mhz clock, but I'm confused, I thought it just extracted its clock from the SPDIF input?

On a more general not, I'm a bit vague on how the clock signal works in my DAC (or any DAC for that matter) and how it controls the TDA1543 and would welcome some explanation. I think the CS8412 recovers it from the SPDIF input using some phased lock loop thingy and then sends it out to the TDA on pin 12. Is the clock freq. output by the CS chip the same as that received through the SPDIF link. How does the TDA use the clock? Are several clock cycles used to feed one audio sample from one channel (that the CS8412 has kindly provided), through the dac chip bit by bit? I presume the word select input then changes state and a sample from the other channel is then fed through?

Apologies if you are all slapping your foreheads in despair at my ignorance!

Hoping you might have some suggestions on what way to go.

Thanks Rupert
 
I connected TDA1543 directly to CD PRO2 I2S out with a short cable/ 10cm and it works without any problem. I´m no reclocking addicted and can´t tell anything about that. At the moment i´m listening to balanced AD1865N-K DAC with D1 stage feed with CS8412 spdif from CD PRO2. Works also no prob.
 
Hi Rupert

Being relatively new to all of this as well, I will try to pass on my (maybe dubious) understanding of this subject. If I get it wrong, maybe someone with more knowledge will set us straight.

The XO2 is designed for use in a one-box CDP. The XO3 is designed for use in a CDP used as a transport for an external DAC. It has an additional circuit which optimises the signal to be sent by SPDIF to the external DAC. The XO DAC is designed to be placed in an external DAC. It reclocks the incomming SPDIF and sends a clock signal back to the transport - which is probably the best option for an external DAC. The choice of clock module depends on your desired application and future flexibility requirements.

The previous replies are sensible - why bother going through SPDIF just to convert it back to I2S - and introduce a lot of jitter in the process? Of course, if you want a DAC that you can easily move around and connect to other transports - then that's a different story.


Could I connect my Tubedac to I2S rather than SPDIF and would it be better? Some searches suggest I would need to bypass the CS8412 (does this chip essentially turn SPDIF into I2S anyway?) in Scotts DAC and go directly to the TDA1543. The data sheet indicates the TDA accepts I2S on pins 1-3. Will this work? is it this simple? does there need to be any decoupling (capacitor/transformer) like in SPDIF or can I simply wire the two together? Sorry so many questions!

Yes, yes, yes, yes, yes, no and yes.

It is that simple and, therefore, it should be better.

I'm not sure what's included in the CDPro2 - so I also suggest you get some specific advice from Guido on which clock module to use - after you have decided which way you want to connect the dac to the transport.
 
Fin said:
Hi Rupert

Being relatively new to all of this as well, I will try to pass on my (maybe dubious) understanding of this subject. If I get it wrong, maybe someone with more knowledge will set us straight.

The XO2 is designed for use in a one-box CDP. The XO3 is designed for use in a CDP used as a transport for an external DAC. It has an additional circuit which optimises the signal to be sent by SPDIF to the external DAC. The XO DAC is designed to be placed in an external DAC. It reclocks the incomming SPDIF and sends a clock signal back to the transport - which is probably the best option for an external DAC. The choice of clock module depends on your desired application and future flexibility requirements.

The previous replies are sensible - why bother going through SPDIF just to convert it back to I2S - and introduce a lot of jitter in the process? Of course, if you want a DAC that you can easily move around and connect to other transports - then that's a different story.




Yes, yes, yes, yes, yes, no and yes.

It is that simple and, therefore, it should be better.

I'm not sure what's included in the CDPro2 - so I also suggest you get some specific advice from Guido on which clock module to use - after you have decided which way you want to connect the dac to the transport.


Hi

I'd suggest XO3, it leaves all options open, price difference is small

I prefer I2S. When using that, consider direct clocking of the DAC chips, using the extra clock outputs from either XO2 or XO3

cheers
 
till, what do you prefer, the TDA1543 running through I2S or the AD1865 dac?



I prefer the (more parts count, not that simple) AD1865 DAC. I Have not made enough experiments to tell why. I will make another TDA1543 Dac (parts allready here).

Also i will run the AD DAC directly from I2S in future, if i manage to programm the CD PRO2 for the right output format.
 
I knew there was some life out there!

I2S is definatly one to try then, if all I need is three bits of wire then what is there to loose.

I agree Guido, the extra flexibilty of XO3 for small extra cost would seem sensible. Even if the reclocking circuit is redundant when using I2S, a future DAC may bring me back to SPDIF.

I am very interesting in clocking the DAC chip directly from one of the extra clock outputs of XO2/XO3, as I've heard this is perhaps the best way of doing things, but don't know if the TDA1543 in Scotts DAC can be driven from the XO2/XO3 (at CDPro freq.) in this way. I'm reluctant to bug Scott just yet as I've only just emailed him with some other questions.

I have enough though to get started on, so I should perhaps post some questions on this another day.

Many thanks for your collective advice.

Rupert
 
Guido Tent said:
Hi

I'd suggest XO3, it leaves all options open, price difference is small

I prefer I2S. When using that, consider direct clocking of the DAC chips, using the extra clock outputs from either XO2 or XO3

cheers


Guido - does the "extra clock outputs" mean that the XO2 and XO3 have a clock distribution included, or is this still something to consider when installing them?
 
Guido Tent said:
Hi Fin,

Yes, 3 outputs are present, each have their 47 ohm series resistor integrated on the XO2/3 board.

regards


Hi Guido

Does this mean there is no advantage it building a 74HC04 gate based distribution - like in the higher end CDPs?

I assume one clock output goes to to the decoder, one to the filter and one to the dac (or filter/dac in the case of Bitstream)?

Should any of the original clock traces between these chips be altered?

You can tell I haven't even opened the package containing my XO2 yet!!
 
Fin said:



Hi Guido

Does this mean there is no advantage it building a 74HC04 gate based distribution - like in the higher end CDPs?

I assume one clock output goes to to the decoder, one to the filter and one to the dac (or filter/dac in the case of Bitstream)?

Should any of the original clock traces between these chips be altered?

You can tell I haven't even opened the package containing my XO2 yet!!


If gates are not required, do not use them as they add jitter

XO should be close to the DAC and fed directly,. servo and others may be buffered using HC gates

cheers
 
Coulomb said:
Hello Guido, sorry to post this to the thread, but have you been getting my email enquiries? I have sent several emails and none have been answered.

My apologies if the questions are so stupid you can not spare the time to answer them. :) :)

Regards

Anthony


Hi

I've had severe problems with mail last week, am working on backlog now

sorry gents, windos microsoft network cloient server and other things that I do not want to know of - but didn't work properly (from one day to the other)

Will step over to a Mac one day

regards
 
Guido Tent said:

If gates are not required, do not use them as they add jitter

XO should be close to the DAC and fed directly,. servo and others may be buffered using HC gates

cheers

Thanks Guido!

So - I will connect one output from the XO2 directly to the SAA7321 Filter/DAC (where the original crystal was situated-pins 24 and 25), with as short as possible wires (twisted).

Another output from the XO2 will be connected directly to the SAA7310 decoder (no gates) at pins 25 and 26. Is this correct?

What happens to the original traces between the decoder and filter/dac that connect the Xtal pins?
 
Fin said:


Thanks Guido!

So - I will connect one output from the XO2 directly to the SAA7321 Filter/DAC (where the original crystal was situated-pins 24 and 25), with as short as possible wires (twisted).

Another output from the XO2 will be connected directly to the SAA7310 decoder (no gates) at pins 25 and 26. Is this correct?

What happens to the original traces between the decoder and filter/dac that connect the Xtal pins?


Hi

Correct

original traces do not serve any function, as they are " bridged" by connecting both clock inputs to XO2

cheers
 
Guido Tent said:
original traces do not serve any function, as they are " bridged" by connecting both clock inputs to XO2

Hi Guido

I imagined that they would cause some degradation as they carry a "not so clean" clock signal from the filter to the decoder!

Is it "best practice" to cut them or leave them intact?

Is there an advantage to feeding the clock directly to other parts of the CDP? If so, where?

I will be an expert on your clock one day! :D
 
Fin said:


Hi Guido

I imagined that they would cause some degradation as they carry a "not so clean" clock signal from the filter to the decoder!

Is it "best practice" to cut them or leave them intact?

Is there an advantage to feeding the clock directly to other parts of the CDP? If so, where?

I will be an expert on your clock one day! :D


Fin,

Yes, that is a potential danger, though they are seperated by series resistors

The DAC clock is most important to be clocked directly, the others could be seperated by using buffers

cheers
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.