SC5305A Operating & Programming Manual
Rev 2.1.0
26
Equation 1 with dependency on temperature, we add on the temperature dependent gain factor
and obtain the following:
where
is the temperature of the device and
is the fixed temperature at which calibration was
performed. The “Reading the Device Temperature” section provides information on how unsigned raw
temperature data is converted to Celsius, a floating point type. Taking the frequency dependence of the
measured parameters into consideration, Equation 2 may be rewritten as
Note that the IF attenuation values,
do not need to be frequency dependent as discussed
earlier. Using the IF attenuator calibration is as simple as substituting the intended value with the
calibrated value. From Table 8, one would use 29.854 dB for an intended 30 dB attenuation. The other
two variables,
and
, are also frequency independent as they are only referred to at
the center of the IF band, and their values are simply summed in the total gain equation. Only those
parameters that depend on frequency and/or temperature are treated below.
To obtain calibrated gain values from the parameters that are a function of frequency, interpolation is
required to provide the best estimated values. A natural cubic Spline interpolation is suggested
for
,
, and
. The important input parameters for a cubic spline
interpolation are the two arrays
and
, and an arbitrary point
. The output of the interpolation is
some interpolated value
based on the inputs.
is the set of independent values,
is the set of
dependent values, and
is an arbitrary independent value to obtain the interpolated value
. For
example, Table 10 lists the input and output parameters to obtain the gain
.
Table 10. Parameters for a spline interpolation.
Frequency (MHz);
3
5
19
…
…
950
1050 …
3875
3900
Measured Gain;
33.223
33.423
33.213
…
…
32.652
32.482 …
29.980
29.450
1000
32.453
From experience, having a large [X] and [Y] array of points does not necessarily provide the best
interpolated value due the nature of trying to fit a function over many points and over many octaves of
frequency. Better results are obtained from a set of localized calibrated points around the point of
interest. The function
sc5305a_CalcGain
uses six localized [X] points to compute the interpolated point.
Using localized points, the example on Table 10 is re-tabulated in Table 11. Similarly, frequency
dependent preamplifier gain and RF attenuation may be derived.
Equation 2
Equation 3