vivkr
|
Hi,
As is well-known, clock jitter directly limits the maximum SNR of a high-speed, high-resolution ADC. In most references, it is mentioned that the jitter is treated like white noise in arriving at this definition.
However, phase noise of any clock generator source is typically not white, but has a "skirt".
1. Is the calculation for peak SNR achievable still valid then? For instance, if I were to convert the close-in phase noise information to jitter, could I use this number to estimate the peak SNR?
2. How can I do the above conversion? Manufacturers typically give close-in phase noise in dBc/Hz at say a 1 kHz offset. How does one interpret this for finding rms jitteR?
3. Most importantly, how can I decide what sort and amount of phase noise is tolerable in my test signal used for high-speed ADC testing?
If you wish you recommend any specific signal sources for >14bit and 10Msps range ADC testing, then please feel free to do so as well.
Thanks Vivek
|