GNSS Antennas and Front Ends - A Software-Defined GPS and Galileo Receiver: A Single-Frequency Approach - page 59

Denote the quality factor of the filter by Q , the center frequency of the filter
(1575.42 MHz for GNSS L1) by f center , and the bandwidth of that filter by BW.
Then the quality of a filter is defined by

f center
BW .

Q = (4.5)

If we assume a 3 dB bandwidth and desire a filter to capture the main lobe of the
GPS spectrum (2.046 MHz wide), the Q factor for such a filter comes out to be
770—an extremely high value. To put things in perspective, fabrication of most
commercial filters (although it does depend on the technology) sets a minimal
bandwidth of 2% of the center frequency. This corresponds to a Q value of 50,
significantly less than the 770 computed above.
However, perform the same computation at the resulting 47.74 MHz IF from the
design in Figure 4.2. That Q factor is 47

046 or 23.33, which is a much more
realizable filter. Thus, the frequency translation to IF allows higher frequency
selectivity with less costly/complex components.
The second additional factor motivating frequency translation is feedback. The
amount of amplification in the RF chain is tremendous; over 100 dB of gain is
applied. If this is all attempted at a single frequency, then it is highly likely that
feedback will become an issue unless meticulous shielding and spatial separa-
tion across the RF chain is implemented. Otherwise, if the 100

. 74 / 2 .

dB of gain were
applied all within the 1575.42 MHz band even with quality RF cabling between
components, it is unlikely to prevent feedback within the amplification stages in
the RF chain. Utilizing multiple stages within a front-end design allows the gain
to be distributed across frequency. For example, in the single-stage downconver-
sion depicted in Figure 4.2 the gain within the RF chain is split between the RF
and IF paths. Thus the level of shielding and potential for feedback are reduced
as the output of the lower-frequency amplifiers cannot feedback into the input of
the higher-frequency amplifiers.

+

4.2.5 Analog-to-Digital Converter
The final component in the front-end path is the analog-to-digital converter. This
device is responsible for the conversion of the analog signal to digital samples.
There is a wide variety of ADCs available on the market, with a dizzying set of
parameters for each. Consider, for example, the Texas Instruments ADS830 ADC,
see focus.ti.com/lit/ds/symlink/ads830.pdf . Such an ADC has an overwhelming
number of parameters, the majority of which are not discussed here. An applica-
tion note can help users sort out the various parameters associated with ADCs;
see Anonymous (2000).
The key parameters to consider for this discussion are the number of bits ,the
maximum sampling frequency ,the analog input bandwidth ,andthe analog input
range .
The CDMA nature of the GNSS signal requires very little dynamic range from
the sampled signal. It has been shown that if single bit sampling is used, then