Intéressante, l'anecdote sur la 575 meilleure que la 957 pour la moitié du prix !
Sinon, merci Phile d'avoir retrouvé les explications de John Swenson sur le sujet des horloges précisant :
"As jabbr mentioned what does seem to have good correlation with sound quality is close in phase noise, so the phase noise plot is the most important piece of information when looking at clock specs. There are many other clock parameters, and some companies get caught up in optimizing those parameters which do not have correlation with good sound. Thus you can have the situation of a company saying their clock is a "superclock" when that is referring to an optimized parameter which has nothing to do with good sound.
A very favorite one is temperature coefficient. A really good oscillator might have something like 1ppm (parts per million) which sounds really good, it is something the marketing department can really highlight, but unfortunately has nothing to do with the phase noise, which is what really matters. As a matter of fact As jabbr mentioned what does seem to have good correlation with sound quality is close in phase noise, so the phase noise plot is the most important piece of information when looking at clock specs. There are many other clock parameters, and some companies get caught up in optimizing those parameters which do not have correlation with good sound. Thus you can have the situation of a company saying their clock is a "superclock" when that is referring to an optimized parameter which has nothing to do with good sound.
A very favorite one is temperature coefficient. A really good oscillator might have something like 1ppm (parts per million) which sounds really good, it is something the marketing department can really highlight, but unfortunately has nothing to do with the phase noise, which is what really matters. As a matter of fact the circuitry which radically decreases the temperature coefficient in a TCXO (Temperature Compensated Crystal Oscillator) actually significantly increases the phase noise. So if you see someone using a TCXO in a DAC, you know they don't really know how to choose the right oscillator.
There is another class of oscillator called the OCXO which you might have heard of (Oven Compensated Crystal Oscillator) which heats a special crystal to a high temperature which produces an extremely low temperature coefficient, say 1PPB, yep 1 part per BILLION. But again that does not guarantee very low phase noise. It does turn out that the lowest phase noise oscillators are OCXOs, but not all OCXOs are really low phase noise. There are a lot of inexpensive OCXOs that are much worse than the Crystek 575. To get an OCXO that has lower phase noise than the 575 takes a lot of money"
Effectivement, la 575 est comparable à la 957 et coûte moins de 10€ ! Mais on ne la trouve pas à 25 MHz... alors qu'elle existe
https://www.digikey.fr/product-detail/fr...ND/5423299
Édit : si, on la trouve chez Mouser mais il y a trois mois d'attente
https://www.mouser.fr/ProductDetail/Crys...BicHc%2FWA
Sinon, merci Phile d'avoir retrouvé les explications de John Swenson sur le sujet des horloges précisant :
- C'est le bruit de phase qui compte
- Le contrôle de dérive liée à la température dans les TCXO fait du bruit
- Fabriquer une OCXO meilleure que la 575 coûte très cher
"As jabbr mentioned what does seem to have good correlation with sound quality is close in phase noise, so the phase noise plot is the most important piece of information when looking at clock specs. There are many other clock parameters, and some companies get caught up in optimizing those parameters which do not have correlation with good sound. Thus you can have the situation of a company saying their clock is a "superclock" when that is referring to an optimized parameter which has nothing to do with good sound.
A very favorite one is temperature coefficient. A really good oscillator might have something like 1ppm (parts per million) which sounds really good, it is something the marketing department can really highlight, but unfortunately has nothing to do with the phase noise, which is what really matters. As a matter of fact As jabbr mentioned what does seem to have good correlation with sound quality is close in phase noise, so the phase noise plot is the most important piece of information when looking at clock specs. There are many other clock parameters, and some companies get caught up in optimizing those parameters which do not have correlation with good sound. Thus you can have the situation of a company saying their clock is a "superclock" when that is referring to an optimized parameter which has nothing to do with good sound.
A very favorite one is temperature coefficient. A really good oscillator might have something like 1ppm (parts per million) which sounds really good, it is something the marketing department can really highlight, but unfortunately has nothing to do with the phase noise, which is what really matters. As a matter of fact the circuitry which radically decreases the temperature coefficient in a TCXO (Temperature Compensated Crystal Oscillator) actually significantly increases the phase noise. So if you see someone using a TCXO in a DAC, you know they don't really know how to choose the right oscillator.
There is another class of oscillator called the OCXO which you might have heard of (Oven Compensated Crystal Oscillator) which heats a special crystal to a high temperature which produces an extremely low temperature coefficient, say 1PPB, yep 1 part per BILLION. But again that does not guarantee very low phase noise. It does turn out that the lowest phase noise oscillators are OCXOs, but not all OCXOs are really low phase noise. There are a lot of inexpensive OCXOs that are much worse than the Crystek 575. To get an OCXO that has lower phase noise than the 575 takes a lot of money"
Effectivement, la 575 est comparable à la 957 et coûte moins de 10€ ! Mais on ne la trouve pas à 25 MHz... alors qu'elle existe
https://www.digikey.fr/product-detail/fr...ND/5423299
Édit : si, on la trouve chez Mouser mais il y a trois mois d'attente
https://www.mouser.fr/ProductDetail/Crys...BicHc%2FWA
Pluie du matin n'arrête pas le sous-marin