Please see the picture below. It shows a center frequency of 10 MHz and a marked signal about 3.1 kHz higher.
We don’t understand this discrepancy because the signal is taken directly from the 10 MHz output of a Quartzlock off-air frequency standard.
Can you help please? Are we doing something wrong?
This 'error' is due to two things:
- There are 631 trace points, which results in a step size of 1585 Hz with a span of 1 MHz.
This is the marker resolution here.
- At a RBW of 30 kHz there is a nearly totally flat range of 8 kHz in the middle of this filter.
Using the marker resolution above we get 7 trace points with the same level value.
The peak marker will set itself on one of these points, and this can be some kHz beside the real peak.
The solution to this problem is to use a smaller RBW (but this will increase the sweep time)
or the 'Freq count' marker function.
If the 'Freq count' marker function is activated, the R&S FSH briefly stops the sweep at the marker position
and measures the frequency using the frequency counter. The resolution of the frequency counter is 0.1 Hz,
which is considerably higher than the resolution obtained without the FREQUENCY COUNT function.
Even though the resolution is high, frequency counting is extremely fast due to a special algorithm
for the I/Q baseband signal (approx. 30 ms at a resolution of 1 Hz). Basically, the accuracy of the
frequency readout depends only on the accuracy of the internal reference frequency (TCXO).
