|
Global Positioning System (GPS) capability is one of the most popular new features for current and next-generation cell phones. Using the GPS feature and associated firmware, it is possible to determine one's current location, get directions from a current location to a new location, etc. In addition to burgeoning consumer demand, adoption of this feature is being driven by the recent US federal E911 mandate. However, integrating this feature into a cell phone is no small feat, given the relative proximity of the frequencies of cell phone signals to the GPS signals, the relatively large signal strengths of the cell phone (+30 to +33 dBm for GSM phones)[1] and the relatively low signals of the GPS (in the order of -150dBm)[2]. To incorporate this type of feature and maintain small device form factors, greater demands are being made on integrating circuitry, reducing component size, and mechanical and material engineering. However, designers must make strong efforts to maintain the performance of both the phone and the added feature.
Frequency Jammers
Cell phone users today can choose between several cell phone service providers who use a range of networks, such as GSM, EDGE and UMTS, to provide services. Table 1 illustrates the frequencies used in current cell phone networks to provide services. Transmit frequencies range from 380 MHz to 2570 MHz, and receive frequencies range from 390 MHz to 2690 MHz, while the GPS signal is at 1575.42 MHz. As seen in Table 1, there are currently gaps in the spectrum used by cell phones -- between 1465 MHz and 1710 MHz in the transmit band and between 1513 MHz and 1805 MHz in the receive band. The GPS signal falls into this gap; however, there is still a lot of noise in this area of the spectrum, and the noise bandwidth at the GPS receiver's input needs to be reduced to prevent the sensitivity of the receiver from being degraded.
To decrease the noise bandwidth, filtering is added to the input of the GPS receiver. Most GPS modules ready to be integrated into electronic devices contain a filter, such as a surface acoustic wave (SAW) filter, and an accompanying low noise amplifier within the module. Due to the physical proximity of the cell phone and GPS antennas, the noise level at the GPS input is much higher than it would be for a standalone GPS unit, so an additional filter may be required. The addition of the SAW filter decreases the effects of the noise on the receiver's sensitivity. The filter's insertion loss will increase the noise figure of the GPS receiver, but, if implemented properly, it should have the overall desired affect.
SAW filters are typically the filter of choice, as they are small, perform well and are readily available. As shown in Figure 1, typical filter dimensions are 1.2 X 1.4 X.46 mm[3].
1. SAW filter that has been placed on a PCB with 0402 resistors next to it.
SAW filters designed for GPS signals can be realized with insertion loss performance as low as 1.2 dB, thus minimizing the degradation to noise figure and receiver sensitivity. Other filters that show promise in the future are film bulk acoustic resonator (FBAR) and bulk acoustic wave (BAW) filters. FBAR quintplexers are currently available for the GSM network. The quintplexer has separate ports for the PCS receive band, the PCS transmit band, the cellular receive band, the cellular transmit band, the GPS and a common antenna[4]. With careful circuit board layout, this device could be used in a highly-compact design for a GSM phone with an integrated GPS module.
Layout considerations
Proper layout is critical for the successful design of a system integrating a GPS and a cell phone. The GPS receiver is highly sensitive and may need to decode very low power signals in the order of -150 dBm[2]. This sensitivity could make it susceptible to noise generated within the integrated system. Such noise should be minimized to the extent that it does not interfere with the GPS. Compartmentalizing sections with shields can help eliminate some noise. For this to be successful, it is important that the shields have intimate contact to ground. EM modeling should be performed with commercially available 2.5D and 3D simulators to verify that the ground ring to which the shield is connected is actually a ground for the desired frequencies. Careful simulation of the ground ring, choice of via size and via location ground layer choice, and interconnect to the main power source ground will lead to a robust shield scheme.
The choice of how to layout ground is also an important consideration. Often, top-side grounds are used along with one or multiple internal ground layers. Top-side grounds can be very helpful when trying to isolate signals that are routed on the board surface. The same thorough approach that was applied to the shield grounding ring would be helpful in ensuring the top-side ground is performing the desired task. Figure 2 shows a printed circuit board (PCB) stack-up with multiple internal grounds.
Typically, the ground on a circuit board is a fairly good current return path. To expect this return path to have no voltage potential differences over its area is a bit optimistic. Even on small PCBs, voltage potential differences can be observed many times. For this reason, PCBs with RF transceiver and digital processors are often fabricated with separate grounds for each. If only one type of circuit is on the PCB there is a possibility that one ground plane is enough
2. Generic PCB stack-up showing separate Digital and RF grounds.
Multiple internal grounds are typically used in two fashions. One implementation is to have more than one layer in the PCB stack-up, to pick up a common ground. The other is to have ground planes on one layer of the PCB stack-up separated from each other and joined only at the main power source ground terminal. If multiple layers are used as a common ground, there cannot be any voltage potential between the grounds or a noise path will be created.
|
