Lets’ assume we have a few buckets A….E filled with zeroes except A, which has 
sampled a one from input.

in ->   A       B       C       D       E
1       1       0       0       0       0

BBDs have two copy phases, 1 and 2. In Phase 1, A is transferred to B and D to 
E:

in      A ->    B       C ->    D       E
0       1       1       0       0       0

Phase 2 does this, assuming A samples a 2 from input;

in ->   A       B ->    C       D -> E
2       2       1       1       0       0

Back to phase 1, we see:

in      A ->    B       C ->    D       E
3       2       2       1       1       0


Phase 1 & 2 are one clock period T or 1/f. But the signal traversed over 4 
stages, from A to D. So, it’s faster than light (no, not yet, but it’s 2 stages 
per clock cycle).
Nonetheless, the actual sampling period is T, so the nyquist frequency is still 
f/2.

Hope this helps a bit, it’s indeed a bit confusing at times. 

Steffan





> On 13. Feb 2024, at 11:41, Victor <[email protected]> wrote:
> 
> This Message Is From an External Sender
> This message came from outside your organization.
> Hi all,
> 
> a BBD circuit is said to have a delay time equivalent to N/(2f), where N is 
> the number of stages and
> f is the clock frequency (Hz). The 2 in this seems to indicate that each 
> stage is traversed in T/2 secs,
> the clock period being T = 1/f.
> 
> Ok, so this means the clock is running at twice the sampling rate of the 
> system. In this case, the
> discrete-time baseband would extend to f Hz (positive side only), should it 
> not?
> Why I see it quoted everywhere that the BBD bandwidth is f/2 Hz?  Where did I 
> go wrong?
> 
> 
> Victor
> 
> 

Reply via email to