Hi,
I have a nos CS8412 + TDA1541A dac. I did a low jitter external clock module with 74HC4040 and TCXO (0.5PPM 6ps). Unfortunately external clock did not make any sound improvement at all
I have listened music with and without the external clock but, my dac is sounding same. A year ago, when i bought an assembled 4xTDA1543 dac kit from ebay, i had listened it's sound with and without it's external clock and the result was the same. I have chceked internet forums and saw that some people wrote in their threads that; if the internal clock is properly working in the dac circuit then there should not be an audible sound imrovement with external clock! I am totaly amazed. The jitter of the internal clock in CS8412 digital receiver chip is 200ps. My external clock has only 6ps jitter but, no improvement. Any comment
Note: External clock is working properly.
My dac schematic is this:
I have a nos CS8412 + TDA1541A dac. I did a low jitter external clock module with 74HC4040 and TCXO (0.5PPM 6ps). Unfortunately external clock did not make any sound improvement at all
I have listened music with and without the external clock but, my dac is sounding same. A year ago, when i bought an assembled 4xTDA1543 dac kit from ebay, i had listened it's sound with and without it's external clock and the result was the same. I have chceked internet forums and saw that some people wrote in their threads that; if the internal clock is properly working in the dac circuit then there should not be an audible sound imrovement with external clock! I am totaly amazed. The jitter of the internal clock in CS8412 digital receiver chip is 200ps. My external clock has only 6ps jitter but, no improvement. Any comment Note: External clock is working properly.
My dac schematic is this:
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
clock
I see. So, i can use my low jitter 4 legs crystal to modify my Marantz CD80 cd player's existing 2 legs crystal. I2S connection also will be a great improvement i believe. Because, in this case the new low jitter crystal circuit in CD80 will sncronise the transport and TDA1541A together. Any comment?
Thanks
Noyan
I see. So, i can use my low jitter 4 legs crystal to modify my Marantz CD80 cd player's existing 2 legs crystal. I2S connection also will be a great improvement i believe. Because, in this case the new low jitter crystal circuit in CD80 will sncronise the transport and TDA1541A together. Any comment?
Thanks
Noyan
Hi,
There are many issues here.
A clock module itself may have very low jitter (say 0.5pS RMS) but unless it is fed with an essentially completely noise-free powersupply it's jitter will be very high (dependent on the noise in the circuit).
In a commercial clock module I designed I used a very large number of components only for the Supply, to get low noise so jitter is low. And still, with all that effort the Powersupply is the limit.
The second issue is that the schematic you show actually does not work at all. What your schematic does is to place the CS8412 into slave mode. This means the data is clocked out of the receiver using the local clock. This SEEMS a good idea at first. But unless you use the same clock that you use to clock the data out of the CS8412 to drive the Transport source you have major problems.
Even only the very common +/-50ppm frequency of the crystal in the transport will lead to frequently dropped or repeated frames of data. This sounds MUCH WORSE than a little bit of jitter.
the clock may, but the DAC does not and NEVER CAN work properly until you get rid of this extremely stupid reclocking scheme. Of course, you will still have ton's of jitter as the CS8412 not only adds around 200pS, but it also does not suppress any incoming jitter, so at the CS8412 output you are very likely to have a 500 - 1000pS jitter with the best sources.
So I guess you up a certain creek without paddle.
The problem is that external DAC's generally do not resolve the jitter issue in any kosher way (especially DIY ones don't) and so DAC's just do not work properly and never will, until the clock issue is resolved.
The idea is that "The clock belongs next to the DAC". However this clock must also control the drive, or things go awry. So the answer is a CD-Player where DAC and Transport are separated by as little distance as possible and both run from the same precision clock.
The other solution is to fit very sophisticated clock management into the DAC, so that the local clock is both essentially infinity adjustable and at the same time offers as low jitter as a fixed clock. That way it can track the incomming clock frequency but can at the time "ride out" all the jitter.
Being involved with designing something like that, let me assure that to get this done at a level of performance that matches fixed oscillators is significantly non-trivial.
Ciao T
There are many issues here.
A clock module itself may have very low jitter (say 0.5pS RMS) but unless it is fed with an essentially completely noise-free powersupply it's jitter will be very high (dependent on the noise in the circuit).
In a commercial clock module I designed I used a very large number of components only for the Supply, to get low noise so jitter is low. And still, with all that effort the Powersupply is the limit.
The second issue is that the schematic you show actually does not work at all. What your schematic does is to place the CS8412 into slave mode. This means the data is clocked out of the receiver using the local clock. This SEEMS a good idea at first. But unless you use the same clock that you use to clock the data out of the CS8412 to drive the Transport source you have major problems.
Even only the very common +/-50ppm frequency of the crystal in the transport will lead to frequently dropped or repeated frames of data. This sounds MUCH WORSE than a little bit of jitter.
the clock may, but the DAC does not and NEVER CAN work properly until you get rid of this extremely stupid reclocking scheme. Of course, you will still have ton's of jitter as the CS8412 not only adds around 200pS, but it also does not suppress any incoming jitter, so at the CS8412 output you are very likely to have a 500 - 1000pS jitter with the best sources.
So I guess you up a certain creek without paddle.
The problem is that external DAC's generally do not resolve the jitter issue in any kosher way (especially DIY ones don't) and so DAC's just do not work properly and never will, until the clock issue is resolved.
The idea is that "The clock belongs next to the DAC". However this clock must also control the drive, or things go awry. So the answer is a CD-Player where DAC and Transport are separated by as little distance as possible and both run from the same precision clock.
The other solution is to fit very sophisticated clock management into the DAC, so that the local clock is both essentially infinity adjustable and at the same time offers as low jitter as a fixed clock. That way it can track the incomming clock frequency but can at the time "ride out" all the jitter.
Being involved with designing something like that, let me assure that to get this done at a level of performance that matches fixed oscillators is significantly non-trivial.
Ciao T
Clock
Hi ThorstenL,
I am newby for digital world Thank you for your explanations but i could not understand where is the problem in my dac schematic Another but similar story; About year ago i have bought a dac from ebay (CS8412+ 4xTDA1543 parallel with selectable clock option. I have attached the shematic paper to this message for your rewiew. As you can see, the shematic of it about the same with the other. As i stated in my first message in this topic, when i was playing the clock switch there was no audible difference for me. What is wrong in those similar schematics and how can i fix the problem please can you describe me as possible as simply? Note: I did my clock unit for testing and it use battery supply and LM7805 regulator. In my opinion it should be silent. Or, may be i am wrong.
Thanks a lot
Noyan
Hi ThorstenL,
I am newby for digital world
Thank you for your explanations but i could not understand where is the problem in my dac schematic Another but similar story; About year ago i have bought a dac from ebay (CS8412+ 4xTDA1543 parallel with selectable clock option. I have attached the shematic paper to this message for your rewiew. As you can see, the shematic of it about the same with the other. As i stated in my first message in this topic, when i was playing the clock switch there was no audible difference for me. What is wrong in those similar schematics and how can i fix the problem please can you describe me as possible as simply? Note: I did my clock unit for testing and it use battery supply and LM7805 regulator. In my opinion it should be silent. Or, may be i am wrong.Thanks a lot
Noyan
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
In my little world of experimentation I have found that using I2S instead of SPDIF has, for me, been the ultimate solution. My limited knowledge of these chips and exactly how they work leads me to believe everything that Thorsten has said. You cannot eliminate the jitter contribution from the 8412. It would ultimately be better off eliminating it from the signal chain, and I2S would do exactly that.
Hi,
Okay. I'll try again.
1) The data from your CDP or Transport comes at a rate determined by the crystal in that Machine. A difference in clock frequency (absolute) of +/-100ppm is quite normal, +/-500ppm not uncommon and +/-5,000ppm have been encountered on occasion.
2) The CS8412 will lock onto that frequency in the CDP using it's own PLL.
3) The CS8412 holds three samples of data internally.
4) Your "external clock" operates at it's own frequency, which may be a few 100ppm slower or faster.
5) Once the difference between the clocks has accumulated one sample worth the CS8412 will be forced to switch to the sample before or after (remember, it always has three if the output is handled by the CS8412) and the second time it will have to omit or repeat one sample.
Bang, we now have 22.7uS (that 22,700,000.00ps) peak jitter.
If we have say 250ppm difference between the clocks we have 22,700,000ps jitter with an appx. 10Hz frequency.
Clearer?
Don't really know where to start. If I limit myself to stating what is right I'm quicker, lets just say next to nothing so closely, it makes no difference.
You cannot, without a complete redesign.
As you already have a CD-80 you already have a TDA1541 based player that can be modified for Non-OS, given a better clock and a better analog stage (I recently published a little how-to at the diy magazine at enjoythemusic.com) for much better performance than you will ever get from such badly designed DAC's.
With a 7805? Have you got any idea what the self noise of such a Regulator is? Maybe you want to look at a somewhat better regulator. What I am using in my commercial "superclock" is quieter by a factor 1000 or so...
Ciao T
Thank you for your explanations but i could not understand where is the problem in my dac schematic
Okay. I'll try again.
1) The data from your CDP or Transport comes at a rate determined by the crystal in that Machine. A difference in clock frequency (absolute) of +/-100ppm is quite normal, +/-500ppm not uncommon and +/-5,000ppm have been encountered on occasion.
2) The CS8412 will lock onto that frequency in the CDP using it's own PLL.
3) The CS8412 holds three samples of data internally.
4) Your "external clock" operates at it's own frequency, which may be a few 100ppm slower or faster.
5) Once the difference between the clocks has accumulated one sample worth the CS8412 will be forced to switch to the sample before or after (remember, it always has three if the output is handled by the CS8412) and the second time it will have to omit or repeat one sample.
Bang, we now have 22.7uS (that 22,700,000.00ps) peak jitter.
If we have say 250ppm difference between the clocks we have 22,700,000ps jitter with an appx. 10Hz frequency.
Clearer?
Don't really know where to start. If I limit myself to stating what is right I'm quicker, lets just say next to nothing so closely, it makes no difference.
You cannot, without a complete redesign.
As you already have a CD-80 you already have a TDA1541 based player that can be modified for Non-OS, given a better clock and a better analog stage (I recently published a little how-to at the diy magazine at enjoythemusic.com) for much better performance than you will ever get from such badly designed DAC's.
With a 7805? Have you got any idea what the self noise of such a Regulator is? Maybe you want to look at a somewhat better regulator. What I am using in my commercial "superclock" is quieter by a factor 1000 or so...
Ciao T
Master clock mod for Marantz CD80
I have already one 0.5PPM TCXO 11.2896mhz. Can i cancel the existing 11.2896mhz crystal oscillator in my Marantz CD80 and use this new clock circuit instead?
Thanks a lot
I have already one 0.5PPM TCXO 11.2896mhz. Can i cancel the existing 11.2896mhz crystal oscillator in my Marantz CD80 and use this new clock circuit instead?
Thanks a lot
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Sure, that would be the only way to go. But SAA7220 will add it's own jitter to the mix, and that is way higher than you think. Also, 0.5PPM dosn't necessarily mean that it is a low jitter oscillator (especially on low-frequency).
You are better off replacing those 5534 OpAmps with something better (there are several single OpAmps that are WAY better).
Attachments
Last edited:
Hi,
Yup. Ideally you remove the SAA7220 and convert to Non-Os.
IIRC, it is with SAA7220 removed link:
pin1 - pin 18
pin2 - pin 16
pin3 - pin 15
Even better, place a 74AC174 reclocker (and I2S attenuators per ec-designs as well as DEM reclocking per ec-designs) on a little plugin PCB to replace the SAA7220.
In this case you can drive XSYS with the 11.2896MHz clock as well as the reclocker & DEM reclocking.
If you remove the SAA7220 you need to drive XSYS with the clock and add the soft mute (pin 23) in suitable polarity to the analog muting.
I agree, in the analog stage upgrade the Op-Amp's or just replace with a nice tube stage.
As said, you can have a look at what I did on a Marantz CD-80 at Enjoy the Music.com - High End Audiophile Audio Equipment and Music Reviews. The Absolute Sound, Superior Audio, hi-fi+ and Review Magazine in the DIY-Magazine.
Ciao T
Yup. Ideally you remove the SAA7220 and convert to Non-Os.
IIRC, it is with SAA7220 removed link:
pin1 - pin 18
pin2 - pin 16
pin3 - pin 15
Even better, place a 74AC174 reclocker (and I2S attenuators per ec-designs as well as DEM reclocking per ec-designs) on a little plugin PCB to replace the SAA7220.
In this case you can drive XSYS with the 11.2896MHz clock as well as the reclocker & DEM reclocking.
If you remove the SAA7220 you need to drive XSYS with the clock and add the soft mute (pin 23) in suitable polarity to the analog muting.
I agree, in the analog stage upgrade the Op-Amp's or just replace with a nice tube stage.
As said, you can have a look at what I did on a Marantz CD-80 at Enjoy the Music.com - High End Audiophile Audio Equipment and Music Reviews. The Absolute Sound, Superior Audio, hi-fi+ and Review Magazine in the DIY-Magazine.
Ciao T
Thank you for your help and schematic. I will do it tomorrow. I have already made the nos mod and opamp mod by replacing ne5534s with opa627. After clock mod i wonder how it will sound.
Thanks again!
Hi,
Works fine here?
Try this directly for the article:
Modifying A Marantz CD-80 CD \Player A blow by blow account of taking a venerable CD player and giving it a new lease of life. Article By Thorsten Loesch
Ciao T
Funny link does not work here....
Works fine here?
Try this directly for the article:
Modifying A Marantz CD-80 CD \Player A blow by blow account of taking a venerable CD player and giving it a new lease of life. Article By Thorsten Loesch
Ciao T
I dunno what it is but I cannot reach anything from enjoythemusic site....
I have enabled cookies for that site also but it doesn't work
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.