On 2006-01-10 12:59, ARCADIOS wrote:
what is the exact work that external wordclocks do?
Both AD and DA converters need a constant pulse for doing the 'measurement' of the input signal level, and also for reproducing the sampled values.
A 'regular' crystal clock circuit isn't stable enough if high quality is demanded, mostly due to temperature drift that changes the electrical specs of the circuit.
are they really necessary? in which cases mostly? how does it work.
for simplicity imagine a perfect sine wave that's scanned at infinetely precise intervals.
Each point of the curve corresponds precisely to it's position regarding level and time.
If the clock pulse drifts (unavoidable in real life, an effect called jitter) the 'perfect' position is missed.
The wave literally gets a small bumb, which is perceived as distortion. A waveshaping oscillator does this intentionally - the effect is quite impressive.
To reproduce the theortical resolution of a 20 bit value (or 24, sorry I don't remember) the clock deviation has to be less than 6 pico seconds, the 300th part of a nanosecond... no kidding
Anyway that's theory, but it nevertheless shows the dimension of precision required - and that's why these clocks cannot be cheapo.
An 18 bit converter will deliver superiror results than a 24 bit type if it's controlled by a more precise clock - imho that's one of the reasons for the great reputation of the old A16 (for example)
how the connections on the back of the stand alone wordclock are supposed to be connected?
standard is 75 Ohm coax cable, preferable by star distribution from a central point to each device
cheers, Tom