Concepts and features
R&S
®
ZNB/ZNBT
93
User Manual 1173.9163.02 ─ 62
Ch1 HW Sweep
Ch2 HW Sweep
Ch3 HW Sweep
…
Ch1 Sweep End
Ch2 Sweep End
Ch3 Sweep End
Ch1 Calc
Ch1 Results
Ch2 Calc
Ch2 Results
Ch3 Calc
Ch3 Results
Figure 4-1: Legacy sweep implementation (no pipelining)
For each sweep, trace data processing is triggered when the first data of the last partial
measurement arrives.
With pipelining, the next sweep can already start when the raw measurement data of
the previous sweep has been distributed completely. The calculation of the previous
sweep is processed in parallel to the next one, which effectively eliminates hardware
idle time during sweeps.
Ch1 HW Sweep
Ch2 HW Sweep
Ch3 HW Sweep
…
Ch1 Sweep End Ch2 Sweep End Ch3 Sweep End
Ch1 Calc
Ch1 Results
Ch2 Calc
Ch2 Results
Ch3 Calc
Ch3 Results
Figure 4-2: Pipelining
In the current implementation ("setup-based pipelining"), calculations are synchronized
to the end of the sweep sequence: when all channels of the active setup are swept, the
firmware waits until all data has been calculated before starting the next sweep
sequence.
With active pipelining, “the measurement result” (e.g. a pass/fail) will be available when
the next measurement is already running. Depending on the measurement scenario,
also a pipelined DUT handling can be necessary. For example, a device handler has to
provide a parking position for a DUT until the decision to skip or forward it is made.
The described optimization is only available for the alternated driving mode.
Basic concepts
