willyhoops;201812 Wrote: > I am not sure if SB3 jitter is below 20ns rms... but if i try this... > 20ns = 0.00000002 seconds then 1 / 0.00000002 = 50,000,000 then andyC's > statement "Standard tolerance on a quartz crystal is +/-50 parts per > million ... [and] you can't hear the difference" sounds spot on.
I still think you're unclear as to what jitter is. You've certainly misinterpreted the figures in my example above. A clock can have zero jitter and yet still have an inaccurate frequency. In my example, a crystal that's 50ppm too fast would result in a sample rate of 44.102205kHz. But, provided that each and every one of those samples was clocked into the DAC at precise intervals of 1/(44.102205k) seconds, there would be no jitter at all. Jitter is NOT the same thing as a frequency error. Try this: imagine a turntable playing a record. The motor is accurate and smooth, so the stylus tracks the groove at exactly the correct speed all the time. That's accurate frequency and no jitter. Now suppose the motor speeds up fractionally, so that instead of spinning at (say) 45rpm, it now spins smoothly at 45.00225 rpm. That's a frequency error of 50 parts per million, but still no jitter. You can't hear the difference. Now replace the nice, smooth belt drive with a rattly gearbox. The average speed remains the same, but there's torsional vibration as each tooth engages and disengages. So now, as the record turns, sometimes the relative speed between stylus and groove is slightly faster, and sometimes it's slightly slower. That's more like the effect of jitter. MSB confuse the issue slightly by stating that they use a crystal accurate to 2.5ppm. That's all well and good, but the absolute frequency accuracy really doesn't matter - they'll just have found that applications that require low jitter also tend to require an accurate frequency as well, so that's what crystal manufacturers make. Audio, however, is NOT one of those applications. They just mention the frequency spec because it looks good in marketing literature to people who don't really understand the issue. Knowing that their own clock is accurate also minimises the size of buffer they need - they only have to cope with 52.5ppm of mismatch, and not, say, 100ppm. SB3's jitter is way, way less than 20ns. Here's a plot of its digital output: http://www.cawte.nildram.co.uk/Jitter/sb3.jpg You can see that the transition points between 1 and 0 are very well defined in time - I actually struggled to see any jitter at all with my 'scope, and I suspect that what I did see was instability in the scope itself. -- AndyC_772 ------------------------------------------------------------------------ AndyC_772's Profile: http://forums.slimdevices.com/member.php?userid=10472 View this thread: http://forums.slimdevices.com/showthread.php?t=35261 _______________________________________________ audiophiles mailing list audiophiles@lists.slimdevices.com http://lists.slimdevices.com/lists/listinfo/audiophiles
