Stanford Research Systems SR780, Operating Manual And Programming Reference

Page: 286 / 497

4-104 Swept Sine Average Menu
SR780 Network Signal Analyzer
Integration Time
Set the Integration Time in increments of 3.906 ms [15.625 ms .. 1 ks].
At each frequency point, the inputs measure the signal at the source frequency. This is
done by multiplying the input data by the source sine (and cosine) wave and averaging
the results over an integration time. The actual integration time is always an exact number
of cycles at the source frequency. This rejects signals which are at a different frequency,
such as noise and harmonics. A long integration time results in a narrow detection
bandwidth at the source frequency. This improves signal to noise at the cost of longer
measurement times .
The Integration Time is converted to next largest number of exact cycles. The actual
number of integration cycles is the larger of the Integration Time (in cycles) and the
Integration Cycles. The integration time is always a minimum of 1 cycle or 15.625 ms.
To measure each point for the same amount of time (constant detection bandwidth), set
the Integration Cycles to 1 and the Integration Time to the desired time. The Integration
Time should be greater than 1 cycle of the lowest frequency in the sweep.
Changes made to the Integration Time during a sweep take effect immediately.
The estimated sweep time is displayed in the Horizontal Scale Bar. This time is simply
the sum of the Settle and Integrate times for all points in the sweep. Auto functions
(Source Auto Level, Auto Range, Auto Resolution) will change the actual sweep time.
Command: SITM (?) d {, x}
Integration Cycles
Set the number of Integration Cycles [1..32767].
At each frequency point, the inputs measure the signal at the source frequency. This is
done by multiplying the input data by the source sine (and cosine) wave and averaging
the results over an integration time. The actual integration time is always an exact number
of cycles at the source frequency. This rejects signals which are at a different frequency,
such as noise and harmonics. A long integration time results in a narrow detection
bandwidth at the source frequency. This improves signal to noise at the cost of longer
measurement times .
The integration time is an exact number of cycles. The actual number of integration
cycles is the larger of the Integration Time (in cycles) and the Integration Cycles. The
integration time is always a minimum of 1 cycle or 15.625 ms.
To measure each point for a time inversely proportional to the frequency, set the
Integration Time to 15.625 ms (minimum) and the Integration Cycles to the desired
number. Remember, the detection bandwidth increases with frequency in this case (the
cycles get shorter) which may result in increased detected noise at higher frequencies.
Changes made to the number of Integration Cycles during a sweep take effect
immediately.