| kenev |
Has anyone replaced the onboard oscillator of the CD-PRO2 unit with an external low jitter clock?
I have searched the forum but didn' t manage to find something about this. There was a post somewhere regarding this issue, but things were not clear.
Reading the manual of the CD-PRO2 I find that there is a connector with the designation "16 MHz (optional)", but, as far as I understand this, ths is for providing the internal 16 MHz clock to an external unit.
I' m interested in building a KWAK-CLOCK and feeding my unit with it.
What I' m asking is where should I connect the clock (and from where should I disconnect the onboard clock). Has anybody some kind of experience with such a modification?
Evangelos |
|
|
| Guido Tent |
| quote: | Originally posted by kenev
Has anyone replaced the onboard oscillator of the CD-PRO2 unit with an external low jitter clock?
I have searched the forum but didn' t manage to find something about this. There was a post somewhere regarding this issue, but things were not clear.
Reading the manual of the CD-PRO2 I find that there is a connector with the designation "16 MHz (optional)", but, as far as I understand this, ths is for providing the internal 16 MHz clock to an external unit.
I' m interested in building a KWAK-CLOCK and feeding my unit with it.
What I' m asking is where should I connect the clock (and from where should I disconnect the onboard clock). Has anybody some kind of experience with such a modification?
Evangelos |
Hi
If you have a circuit diagram, I could be of help
Most logical you feed the new clock to the input of the original crystal oscillator (after removing the crystal ofcourse)
By the way, my most recent measurements on a Kwak clock showed horrible results.
best regards
Guido |
|
|
| Peter Daniel |
| Here's some info I found in CD-Pro documentation |
|
|
| Elso Kwak |
Hi Evangelos,
You almost answered your own question.
Just remove the original crystal and the two small caps as in the picture kindly provided by Peter.
You now have to connect the output of the KWAK-CLOCK to one of the pads of the original crystal connected to CRIN.
If you can not determine from the schematic wich pad is connected to CRIN just try both pads, one at a time, with a 1k resistor in series wth the clock-output. Keep the pad that works and remove the 1k resistor.
I hope to visit your beautiful country soon.....:) |
|
|
| Lupulroz |
I've made that modification. I don't remember what pin is the right one but this is critical. One of my SAA7324 died in this experiment :bawling: so take care, the chip is very sensitive.
My Philips dealer said it cannot provide this chip even if I wanted 1000 pcs. :eek:
I will look and tell you about the real clock input. |
|
|
| jean-paul |
| quote: | Originally posted by Guido Tent
Hi
If you have a circuit diagram, I could be of help
Most logical you feed the new clock to the input of the original crystal oscillator (after removing the crystal ofcourse)
By the way, my most recent measurements on a Kwak clock showed horrible results.
best regards
Guido |
Hi Guido, horrible in what sense ? Jitter, waveform ? I heard/read quite some good responses on Elso's clock. |
|
|
| Peter Daniel |
| So did you complete the modification or the loss of a chip made your unit useless? |
|
|
| Lupulroz |
| I completed the modification on other units. Of course the one with dead SAA it's useless until the chip is replaced. |
|
|
| Peter Daniel |
| What clock frequency did you use? From info I posted it seems like at least 3 different frequencies are possible. |
|
|
| Lupulroz |
| I used 8.4672 MHz external clock - ECL type with discrete buffer. |
|
|
| kenev |
Thanks all for your input.
I've made that modification. I don't remember what pin is the right one but this is critical. One of my SAA7324 died in this experiment so take care, the chip is very sensitive.
That's exactly what I' m afraid of about intervening in such a compact, delicate and with sensitive parts unit as the CD-PRO2.
Lupulroz,
Since, from what you posted, I understand that you have modified quite a large number of CD-PRO2 units, can you tell us, from your experience, some useful precaution and modification tips for disconnecting the internal crystal and attaching an external oscillator? I have seen that there are a lot of diyAudio forum members that have already built a CD-PRO2 based player and I guess that some of them would like to try the "clock upgrade". Any sharing of your experience would be useful.
By the way, my most recent measurements on a Kwak clock showed horrible results.
Guido,
Do I sense a strong controversy here? "Horrible" is such a grave word to describe. Could you elaborate on this?
Regarding crystal frequency, I quote from the SAA7324 data sheet:
Typical oscillation frequencies required are 8.4672, 16.9344 or 33.8688 MHz depending on the internal clock settings used and whether or not the clock multiplier is enabled.
How can someone check what is the frequency used in the unit he has in hand?
Best regards to all,
Evangelos |
|
|
| Elso Kwak |
Hi Evangelos,
The datasheet for the SAA7324 can be downloaded here:
http://www.semiconductors.philips.c...A7324H_M2B.html
It clearly shows that the CRIN is pin16.
As for your last question the frequency of the crystal is usually printed on it. I presume the crystal is located very near pin 16. The crystal looks like a flattened oval metal cylinder, couldn't miss it.
Hope this helps.:) |
|
|
| kenev |
Hi Elso,
Thanks for the input.
I had already the datasheet for the SAA7324.
My question was mainly of an informative and practical nature regarding this modification: if someone has already done this, I would like him to share his experience on such a delicate (IMO) "surgery". Also, his opinion on whether this is worth trying, relying on the benefits of this modification.
Anyway, thank you.;) |
|
|
| Guido Tent |
| quote: | Originally posted by jean-paul
Hi Guido, horrible in what sense ? Jitter, waveform ? I heard/read quite some good responses on Elso's clock. |
I measured the spectral response. It contains loads of RF (FM radio stations and others), but even worse, mixed products of harmonics of the mains at (only) -35 dB.
I did not measure jitter, but the spectral response needs to be cleaned up first.
I measured other clocks as well (LC audio, my own and the HP reference clock), as well as the set up, which was clean to -80dB.
I will perform jitter measurements once I have my HP53310A modulation domain analyser up and running.
If someone has a Trichord and an Audiocom available, I'd be willing to measure these as well, I am not wanting to pay for these expensive things
all the best, |
|
|
| Guido Tent |
| quote: | Originally posted by kenev
By the way, my most recent measurements on a Kwak clock showed horrible results.
Guido,
Do I sense a strong controversy here? "Horrible" is such a grave word to describe. Could you elaborate on this?
Evangelos [/B] |
Hi
The spectral content should be absolutely clean, at least down to -100dB from the carier, respecting some phase noise closeby.
So far, I can measure down to 80 dB, but I am setting up experiments to go deeper.
Jitter measurements (in the time domain) will follow
best regards
-
Guido |
|
|
| joeling39 |
Hi,
I have been toying with the CD-Pro 2 unit for a while. In stock form, it is not as good as my modified Theta Jade. However, I am really keen to try a new clock on the unit.
Problem with the above is that I can't figure out which is the crystal. After some enlightening posts here. My guess is that the crystal freq is 8.4672MHz. I have attached a picture of where I think the relevant elements of the orginal clock configuration is. Please take a look & comment if it is accurate.
Regards,
Joe Ling |
|
|
| joeling39 |
| Another picture |
|
|
| Elso Kwak |
Hi Joe,
This looks more like a SMT crystal-oscillator to me, just as in the Sony SCD-1 or 777ES. Very difficult to remove!
Best done VERY carefully with a hot air gun.:cool:
Were some posts on AA on this issue. |
|
|
| Peter Daniel |
| What about just cutting traces, insted of removing? |
|
|
| joeling39 |
Elso,
Thanks for the info. You clock resides in my Marantz CD63KI & is still functioning perfectly after all this while (maybe more than 1 year already).
Regards.
Joe Ling |
|
|
| Guido Tent |
| quote: | Originally posted by joeling39
Another picture |
Hi
From the picture I see that they did their utter best to place the crystal supporting capacitors as far away from the chip, well done boys, now the RF currents flow all over the board..........
Once you have measured and investigated these aspects, it is so easy........
regards, |
|
|
| Elso Kwak |
| quote: | Originally posted by Peter Daniel
What about just cutting traces, insted of removing? |
Great idea Peter!
Maybe it is a SMT crystal, I have never seen a thingie exactly like this before.:) |
|
|
| Peter Daniel |
| Here's the better look, for everybody interested. It is SMD and shouldn't be a problem desoldering it at the 4 points. |
|
|
| Guido Tent |
| quote: | Originally posted by Peter Daniel
Here's the better look, for everybody interested. It is SMD and shouldn't be a problem desoldering it at the 4 points. |
careful guys, this SMD can be glued as well |
|
|
| fmak |
| quote: | Originally posted by Elso Kwak
Hi Joe,
This looks more like a SMT crystal-oscillator to me, just as in the Sony SCD-1 or 777ES. Very difficult to remove!
Best done VERY carefully with a hot air gun.:cool:
Were some posts on AA on this issue. |
-------------------------------------------------------------------
The glue is evil. But the Sony SCD1/777 one is real easy. Solder wick the joints, and use a 35W iron on the body. Came off straight away and still worked perfectly!! |
|
|
| fmak |
I measured other clocks as well (LC audio, my own and the HP reference clock), as well as the set up, which was clean to -80dB.
----------------------------------------------------------
Guido
I am going to toy with external word clocks separate from the source clock. Any experience of whta is a good one?
My system with your XO3 sounds significantly different on source clock and internal 10ppm VXCO clock with twin PLL and I don't know why? The setup cocnsist of the XO3 outputting into a dCS 972 DSD/24-192 upsampler. Inserting another twin PLL dejitter device in between changes the sound again and perhaps not for the better!
The influence of good 75R BNC and AES/EBU cables into proper terminations actually controls the character of the sound and is just as significant!! A Belden 1505A 4 m 75R cable sounds worst in the sense that the sound stage flattens and looses image precision.
Comments. anyone???
:( :dodgy: |
|
|
| Guido Tent |
| quote: | Originally posted by fmak
Guido
I am going to toy with external word clocks separate from the source clock. Any experience of whta is a good one?
My system with your XO3 sounds significantly different on source clock and internal 10ppm VXCO clock with twin PLL and I don't know why? The setup cocnsist of the XO3 outputting into a dCS 972 DSD/24-192 upsampler. Inserting another twin PLL dejitter device in between changes the sound again and perhaps not for the better!
The influence of good 75R BNC and AES/EBU cables into proper terminations actually controls the character of the sound and is just as significant!! A Belden 1505A 4 m 75R cable sounds worst in the sense that the sound stage flattens and looses image precision.
Comments. anyone???
:( :dodgy: |
Fred,
Wordclock: I have been thinking about this too, but cannot come any further than dividing, and eventually reclocking (I like D flipflops very much)
Do I understand correctly that your main clock is not a free running one close to the DAC?
Careful with cascading PLL's, as their poles should be well seperated.......
regards |
|
|
| fmak |
Wordclock: I have been thinking about this too, but cannot come any further than dividing, and eventually reclocking (I like D flipflops very much)
-----------------------------------------
Commercial ones are v expensive. I was wondering if there are good diy designs.
------------------------------------------
Do I understand correctly that your main clock is not a free running one close to the DAC?
-------------------------------------------
My dCS 972 ouputs double speed twin links to the 954 DAC. It has its own VCXO which can be enabled, or can lock onto the incoming clock. Alternatively it can use word clock or GPS reference.
The dac has its own VCXO with twin PLL.
So, I have XO3, twin PLL in 972, feeding twin PLL in 954. The latter two are designed to work with each other.
The arrangement is still very sensitive to choice of cable. A RG62 75R cable I bought from Farnell for £6 actually sounds rather good! |
|
|
| Lupulroz |
Sorry for my late reply - I was offline for a while :cool:
Indeed the input clock pin is CRIN - 16.
Yes I modified more units type vae1250 which uses the same SAA7324. The simplest way is to desolder pins 15 and 16 from the pcb and feed clock into pin 16.
I will use CDPRO2 in my future cd-project and I'll be glad to share some results.
The overall sound was pretty good ;) I recommend this modification.
A picture with cd-player that used this approach can be found
here
The product is no more available :( |
|
|
| tone |
joeling39
Hi,
where did u bought the CDPRO2 from? locally or directly from daisy? and how much? i have lost ur email address..pls email me offline..thanks.
tone. |
|
|
| LC Audio |
I am sorry Elso and Guido
The reason why the SAA7324 fail with your clocks is this:
Your clocks give out a 5 V signal but the SAA7324 is a 3.3 V chip. (Absolute maximum rating for input signals are: 3.6 V)... (See datasheet from above link).
This and other 3.3V chips will be destroyed with the KWAK clock even if correctly mounted........
And there is really no way of attenuating the clock signal without degrading the jitter performance on this type of standard CMOS oscillator.
Good luck guys .. ;o) |
|
|
| tbla |
| quote: | | on this type of standard CMOS oscillator |
i believe we have a conflict of interest........
"when correctly built and mounted - the kwakclock sounds better than the lcaudio xo2 clock"......:devilr: |
|
|
| Elso Kwak |
Hi Lcaudio,
Don't be sorry, thanks for the hint.
May I suggest using the KWAK-CLOCK in this particular case with +/-3.3V supplies? This can simply be done by changing two resistors:
Make R5 8k66 and R8 12k. All other values remain the same. Best check the supply voltages before connecting the clock to this sensitive chip.
Except the one failure in this thread, possibly due to connecting a clock to pin 17 instead of pin 16, I have never heard of any failures.
:cool: |
|
|
| Elso Kwak |
| quote: | Originally posted by tbla
i believe we have a conflict of interest........
"when correctly built and mounted - the kwakclock sounds better than the lcaudio xo2 clock"......:devilr: |
Yes Troels,
The KWAK-CLOCK is NOT a standard CMOS oscillator!
:bigeyes: |
|
|
| LC Audio |
| Not a very surprising statement coming from you ;) |
|
|
| Elso Kwak |
| quote: | Originally posted by LC Audio
Not a very surprising statement coming from you ;) |
Hi Lcaudio,
It is a fact: there is no CMOS oscillator in my clock!!
:confused: :confused: :bigeyes:
999 posts, I can't believe it myself:clown: |
|
|
| LC Audio |
Sorry i was a little fast on the trigger. You are right!
I confused your clock for another design sent to me a few days ago, and that one was just a simple CMOS can on a PCB. I will check it's name and post here later.
I searched the web for your design, and it seems to be OK. |
|
|
| tbla |
| :drink: :drink: :drink: :drink: :drink: :drink: :drink: :drink: :drink: :drink: :drunk: :faint: |
|
|
| Guido Tent |
| quote: | Originally posted by LC Audio
I am sorry Elso and Guido
The reason why the SAA7324 fail with your clocks is this:
Your clocks give out a 5 V signal but the SAA7324 is a 3.3 V chip. (Absolute maximum rating for input signals are: 3.6 V)... (See datasheet from above link).
This and other 3.3V chips will be destroyed with the KWAK clock even if correctly mounted........
And there is really no way of attenuating the clock signal without degrading the jitter performance on this type of standard CMOS oscillator.
Good luck guys .. ;o) |
Oh, I never got into this part of the discussion. Ofcourse a 3V3 circuit should be fed with a 3V3 clock......
Destroyed: I doubt, at least, when series resistance is used (what i do) as the clock input very likely has ESD protection diodes, able to handle 10mA, if my memory serves me right (I think that chip was still made in Philips C075 process)
Anyhow, we should respect the voltages
thanks |
|
|
| joeling39 |
It lives !
Added Elso's clock to the CD-Pro. Now playing music through my backup DAC - Scott Nixon 1543 DAC
More info later (like how's the sound ? - can't tell at the moment).
Thank you Elso.
Regards,
Joe Ling
Malaysia |
|
|
| joeling39 |
| Hmm, can't seem to post pictures. How is it done ? |
|
|
| Elso Kwak |
| quote: | Originally posted by joeling39
Hmm, can't seem to post pictures. How is it done ? |
Hi Joe,
The problem with the pictures here is usually they are too big. If you download Irfan View from www.irfanview.com you can choose resize/resample. 800 bits is the maximum for this forum. Then just save the file as......And upload from your computer. Irfanview is free of charge!
Hope this helps.
Glad the clock is working. I was a bit anxious after Lcaudio's post.
:cool: |
|
|
| joeling39 |
| testing on pictures |
|
|
| joeling39 |
Let's try a bigger picture of the detail.
Note that the crystal has been destroyed while trying to pry it away from the PCB. I simpley soldered the 'clock out' from Elso's clock onto one "leg" of the crystal while the ground went to the ground of one of the capacitors on the PCB. |
|
|
| joeling39 |
Here's a general view of the whole setup.
CD Pro module on the left, Elso's clock in the middle & power supply for Elso's clock on the right.
Thanks Elso for the computer tip |
|
|
| BrianGT |
| quote: | Originally posted by joeling39
Here's a general view of the whole setup.
CD Pro module on the left, Elso's clock in the middle & power supply for Elso's clock on the right.
Thanks Elso for the computer tip |
Joe,
Can you post your schematic that you are using for your power supply for the cdpro II? I am in the process of planning out a cdpro II project, and I am not sure what to use for the power supply.
--
Brian |
|
|
| joeling39 |
Brain,
I bought my set from Robert of Ahfart audio in Singapore. He supplied the PCB & parts for the power supply as well. So, not sure whether I can give it to you. However, the schematic is nearly identical to these ones I found on the net : |
|
|
| joeling39 |
| for the 5V & 9V supply |
|
|
| tbla |
| how's the sound with the new clock....? |
|
|
| joeling39 |
Give me 1 week & I should be able to make a comparision with my modified Theta Pearl CDT.
Right now it is still sitting on the test bench. Have to haul it up to the listening room.
Regards,
Joe Ling
Malaysia |
|
|
| Josep |
| how's the sound with the new clock....? :smash: |
|
|
| joeling39 |
Hi,
Sorry for taking so long to post my first impression. I have done some brief listening maybe about 1.5 months back. All I can say is that based on the preliminary listening, the CD-Pro sounds more "smooth" than the Theta. That is good for some music like ballad but maybe not so good for harder stuff like rock. Too early to make a conclusion.
I still feel that there is potential for the CD Pro to replace the Theta. However, I am not thinking about that at the moment as I have managed to screw up my AN 1.1 DAC shortly after my last post (the ultimate mod - nasty hum whenever I switch on my DAC ). I am unable to listen to music on the main system ever since.
On the spur of the moment, decided to order an AN 21. balanced but still waiting after 1 month+. Sigh.
Once I get back on my feet, I will again investigate the CD-Pro.
Regards,
Joe Ling
:bawling: |
|
|
| joeling39 |
So sorry, appears that I missed something. Compared to the stock clock, it is definitely a discernable improvement. In that sense, it is something worth trying out as the outlay is marginal:)
Regards,
Joe Ling |
|
|
| Josep |
Is your external clock 8 Mhz as the original ?
Is possible to use a 16 Mhz clock without more changes ? (The SAA7324 supports 8, 16 and 32 Mhz but I dont know if there is necessary changes in the configuration). |
|
|
| joeling39 |
Dear Josep,
I used the 8mhz clock as per the original crystal. I am not technically competant enough to try the other clock frequencies.
Regards,
Joe Ling |
|
|
| devas |
| I'm about to put an external clock too, anyone with experience since then ? |
|
|
| devas |
Thanks Guido, actually I've just bought the kit from them...
What i'm looking for is feedeback and opinion about the changes it brings about , if it's worth doing it or not.
I already put a clock (not yours sorry!) in a Cd2.2 Consonance and it was a great improvement but also simpler to implement.
Fred |
|
|
|
