In a previous life.........
"Fred is pulling legs now I prosume, not"
I presume you meant "presume." I never kid about very good engineer or test equipment. The CSA-8000 is one of two or three pieces of test equipment that nearly made me cry to lose the use of, on leaving Alcatel.
The other two on the list were a Hewlett Packard 4278A Capacitance Meter and a Hewlett Packard 4195A Network/Spectrum Analyzer. I actually used to do real engineering and will very likely never get to do it that level again. Where do you think I learned anything about engineering, reading about it on the web? 
"Fred is pulling legs now I prosume, not"
I presume you meant "presume." I never kid about very good engineer or test equipment. The CSA-8000 is one of two or three pieces of test equipment that nearly made me cry to lose the use of, on leaving Alcatel.
The other two on the list were a Hewlett Packard 4278A Capacitance Meter and a Hewlett Packard 4195A Network/Spectrum Analyzer. I actually used to do real engineering and will very likely never get to do it that level again. Where do you think I learned anything about engineering, reading about it on the web?
About numbers: I read somewhere that 500 ps creates noise below the noise level of 16 bit signal and 6 ps is approx. 130 dB under full throttle. 2 ps is lower than 130 dB => good hearing..... for a while.
...hm. presume but I don't prosume so often therefore I can't spell it.😉
Guido, may I ask how much jitter your clock produce?
...hm. presume but I don't prosume so often therefore I can't spell it.😉
Guido, may I ask how much jitter your clock produce?
peranders said:
< 3ps rms, 3-sigma
For the 8412/14 I have a solution, it is called PLL and can be seen on our DAC page.
I can hear differences between 10ps and 3ps (and I am not alone here.....)
Ciao
Although not having the same knowledge and experience as
Fred to judge, my impression is the same, Guido knows very
well what he is doing. I would suspect he has got himself some
new customers just by answering peoples questions, and
probably revealing more than he should about his designs, and
that without even trying to advertise his products.
Fred to judge, my impression is the same, Guido knows very
well what he is doing. I would suspect he has got himself some
new customers just by answering peoples questions, and
probably revealing more than he should about his designs, and
that without even trying to advertise his products.
3 ps is good, indeed. But still I wonder how you can hear noise caused by jitter which is 25-30 dB blow the noise threshold for normal CD quality.
Re: Re: Relax, guys...
I just checked ELFA and see a square canned oscillator. The price is 2 USD and I suspect (not checked the datasheet) that's a regular CMOS inverter, the same kind as can be seen as original in a CDP.
-----------------------------------------------------------------
XOs vary quite a bit. I have just measured the Tent XO, an Euroquartz 10 ppm to order and a TCXO I bought from Hong Kong. Accurracy is not an issue as they are all close and within 5 ppm.
The Tent XO is very symmetrical and measures 49.9% duty cycle, symmetrical 3 ns rise and fall times and good square wave shape - much better than the LC scope shot. The Euroquartz is close with 51.1% duty cycle and a slight bump in the square wave. The TCXO has a poor 59% duty cylce and rather assymmetrical.
The Euroquartz sound slightly shraper than the Tent but is close. I have not tried the TCXO yet.
My experinece is that it is not just the clock but the CDP chipset that makes the final difference. Put a good clock in a cheap CDP and you don't get the full benefit.
I just checked ELFA and see a square canned oscillator. The price is 2 USD and I suspect (not checked the datasheet) that's a regular CMOS inverter, the same kind as can be seen as original in a CDP.
-----------------------------------------------------------------
XOs vary quite a bit. I have just measured the Tent XO, an Euroquartz 10 ppm to order and a TCXO I bought from Hong Kong. Accurracy is not an issue as they are all close and within 5 ppm.
The Tent XO is very symmetrical and measures 49.9% duty cycle, symmetrical 3 ns rise and fall times and good square wave shape - much better than the LC scope shot. The Euroquartz is close with 51.1% duty cycle and a slight bump in the square wave. The TCXO has a poor 59% duty cylce and rather assymmetrical.
The Euroquartz sound slightly shraper than the Tent but is close. I have not tried the TCXO yet.
My experinece is that it is not just the clock but the CDP chipset that makes the final difference. Put a good clock in a cheap CDP and you don't get the full benefit.
Re: Re: Re: Relax, guys...
Do you have any jitter values. I couldn't find any at Euroquartz.fmak said:
peranders said:
The assumption is wrong: It is is not about noise, it is about distortion (NON harmonic that is) and -120 dB is a normal figure here......
suggestion:
Buy a decent VCXO. Build a good power supply and clock your CDP using the VCXO and apply "noise" signals to the control port to find out what is detectable.
Recalculate the noise towards jitter, and do some experiments with the "noise" (real noise, single sine waves etc) and you will be amazed..........
enjoy
Re: Re: Re: Relax, guys...
Fred,
Ofcours, ofcourse, but many people are amazed by the potential of some DAC chips when changing clocks......
Like more things in life, balance is key
fmak said:
Fred,
Ofcours, ofcourse, but many people are amazed by the potential of some DAC chips when changing clocks......
Like more things in life, balance is key
Just a personal opinion:
I have a heavily tweaked digital crossover in my main system, where I can choose with a press of a button between several
input configurations:
- optical/coax/aes-ebu parallel from the same source
clock sync possibilities:
- spdif derived sync (CS8414)
- on board oscillator with external link to the cdp
- PLL sync to external oscillator from the cdp
- ASRC to an internal clock
internal clock:
- the standard clock on the DSP board
- a Tent clock with a separate supply
- an intentionally bad clock with high/very high jitter
CD sourced clock:
- standard cd quartz oscillator.
- LC osc for variable pitch
Telling the truth, I can not hear any differece between the clock sources and cable types, except only when I switch to the very bad oscillator, in which case I can hear almost dropout like constant noise, but nothing like the lost focus, less bass, annoyinh highs etc.
I have a heavily tweaked digital crossover in my main system, where I can choose with a press of a button between several
input configurations:
- optical/coax/aes-ebu parallel from the same source
clock sync possibilities:
- spdif derived sync (CS8414)
- on board oscillator with external link to the cdp
- PLL sync to external oscillator from the cdp
- ASRC to an internal clock
internal clock:
- the standard clock on the DSP board
- a Tent clock with a separate supply
- an intentionally bad clock with high/very high jitter
CD sourced clock:
- standard cd quartz oscillator.
- LC osc for variable pitch
Telling the truth, I can not hear any differece between the clock sources and cable types, except only when I switch to the very bad oscillator, in which case I can hear almost dropout like constant noise, but nothing like the lost focus, less bass, annoyinh highs etc.
fcserei said:
You must have a great system then, we'll envy you.
I expect to see a few items on TRADING POST soon ? 😀
No, the system is fairly constans, where I'm playing is the speakers, I can not choose betveen the Goodmans Axiom 80, the Urei 813 and the Stax electrostat.
Forgot to mention can I also have a hardvired direct i2s connection from the cd player, which I rarely use because of the inconvenience.
Forgot to mention can I also have a hardvired direct i2s connection from the cd player, which I rarely use because of the inconvenience.
Re: Re: Re: Re: Re: Re: Relax, guys...
As the jitters are not correlated, they do not add up, especially as the spectra may differ (a lot)
---------------------------------------------------------------------------
It's not a question of addition; rather a question of what are contributing to jitter.
This forum seems to concentrate on the clock, giving the impression to some that other factors don't come into play. It is low jitter in the entiren chain that gives good sound. 🙂
As the jitters are not correlated, they do not add up, especially as the spectra may differ (a lot)
---------------------------------------------------------------------------
It's not a question of addition; rather a question of what are contributing to jitter.
This forum seems to concentrate on the clock, giving the impression to some that other factors don't come into play. It is low jitter in the entiren chain that gives good sound. 🙂
Re: Re: Re: Re: Re: Re: Re: Relax, guys...
Fred
I fully agree
In theory only the jitter at the conversion counts. In practice lots of jitter ripples through (that is why we added reclocking on all digital signals entering a PCM63).
Lowering the jitter in a balanced way, throughout the system is key. Power supply, EMC and layout are involved here. Anyhow you need some clean clocks to start somewhere 🙂
regards
fmak said:
Fred
I fully agree
In theory only the jitter at the conversion counts. In practice lots of jitter ripples through (that is why we added reclocking on all digital signals entering a PCM63).
Lowering the jitter in a balanced way, throughout the system is key. Power supply, EMC and layout are involved here. Anyhow you need some clean clocks to start somewhere 🙂
regards
- Status
- Not open for further replies.