LOW COST PHASED ARRAY ANTENNA TRANSCEIVER FOR WPAN APPLICATIONS

Transcription

1 LOW COST PHASED ARRAY ANTENNA TRANSCEIVER FOR WPAN APPLICATIONS Introduction WPAN (Wireless Personal Area Network) transceivers are being designed to operate in the 60 GHz frequency band and will mainly be used for home environment radio links. So far, three basic technologies have been developed for implementing these WPAN devices: 1. Transceivers with a fixed antenna beam and wide aperture: have limited RF performance and no user-tracking ability. 2. MIMO (Multi Input Multi Output): have potential user-tracking ability, but also have marginal RF performance due to high losses that affect waves at 60 GHz reflected by the walls which cancels the potential advantage. 3. Digital Active Phased Array Antenna systems (APAA): have user-tracking ability and good RF power density. In principle digital APAA can handle both compressed as well as uncompressed signals. Digital beam forming is performed by dividing the baseband signal power in as many parts as there are antenna array elements. Then, the bit stream corresponding to each antenna element is digitally phased accordingly with the requested phase value. Now the phased bit streams are used for modulating the RF carrier in one or more steps. At last the modulated carriers are radiated by the antenna array. The baseband processor is complex and expensive; the related conventional RF subsystem is complex and expensive as well. The digital APAA system becomes even more complex when the bit stream is not directly available: this happens when the signal is still compressed. In this case, the baseband processor must first perform a decompression function in order to make available the bit stream. This additional function can significantly increase the cost of the digital APAA. Moreover, if multiple radiated channels are required, the above process and its complications will be multiplied by the number of contemporary channels that are to be handled. We could conclude that ANALOG APAA should be the best technical solution. In fact, analog APAA can handle compressed and uncompressed signals because the signal 1

2 phasing operations are independent of the data stream, don't involve complex and expensive digital processors and can handle several contemporary channels. No analog APAA commercial transceiver at 60 GHz is available yet. In fact, currently, analog APAA conventional technology (based on conventional phase shifters) has prohibitive production costs, even for military applications. General description With this paper, Beam is presenting a new kind of ANALOG APAA transceiver at 60 GHz, characterized by low production cost and high RF performance. The system has several contemporary advantages and some major unique features that no other technology can deliver: 1. Baseband processor independence: signal phasing is performed at the RF level. 2. Simple RF circuitry. 3. Low production cost. 4. Ability to handle compressed as well as uncompressed signals without extra cost. 5. Full transparency to the modulation method. 6. High power density at the receiver antenna, which allows a Bit Error Rate, (BER), widely exceeding any practical need. 7. Automatic reciprocal detection of the network elements. 8. Insensitivity to external parasitic signals at the same frequency. 9. Wall penetration ability. 10. Ability to radiate several independent beams devoted to different users, still securing an excellent BER for each one of the links. 11. Occasional obstacles in the line of sight are avoided using reflected signal bounces from walls. TX and RX antennas are embedded on one external side of an LTCC, (Low Temperature Co-fired Ceramic), multilayer substrate. On the opposite external side, the MMIC (Monolithic Microwave Integrated Circuit) circuits of the transceiver are flip-chip mounted. Strip line microwave connections, passive RF components and bias lines are contained within the inner substrates of the LTCC multilayer structure. The system can have azimuth and elevation beam steering. Though, for WPAN applications, azimuth steering has been considered sufficient. Antenna Several solutions are under development (all of them in the micro strip field) with the aim of maximizing the gain, the bandwidth and the steering angle and minimizing the production cost. In the following, is described the layout and the simulated performance of one example (not the final, not necessarily the best): 2

3 Thanks to the high directivity of the antenna, any possible external parasitic signal at 60 GHz that may possibly enter the receiver is reduced to random noise by the phased array behavior. This circumstance doesn't cause significant interferrence to the service, as long as the system relies on an excess of S/N (Signal to Noise) ratio at the receiver. The same TX, (Transmitter), and RX antennas support and steer multiple independent beams, for radiating / receiving independent signals. TX / RX / Distributed Local Oscillator The TX and RX front end are a conventional MMIC subsystem, as far as circuits at 60 GHz can be considered conventional. On the contrary, the MMIC Local Oscillator, (patent pending), is a new and unconventional structure of conventional circuits used in unconventional way. The following is the block diagram of the Distributed Local Oscillator: 3

4 A reference signal is generated at 15 GHz which injects and locks the push-push oscillator 15 ~ 30 GHz. After splitting, the signals inject and lock the set of push-push oscillators at 30 ~ 60m GHz. This way at the output ports of the buffer amplifiers are delivered coherent signals for UP / DOWN conversion. The phase of each individual signal is changed, tuning the BRF (Band Rejection Filter) of each one of the PSIPPO (Phase Shifted Injected Push-Push Oscillator) at 30 ~ 60 GHz. The analog beam orientation of the system is operated at the level of the LO, avoiding conventional phase shifters. The MMIC technology utilizes bipolar transistors on SiGe, (Silicon Germanium). Overall System The following figure describes the block diagram of the system: As can be seen, the LO (local oscillator) signals are set in such a way that the antenna beam orientation is the same for TX and RX. This is necessary for the automatic selfalignment of the network elements. DOWN conversion The down conversion of the RF signal received by the RX is direct-coherent. Direct down conversion, (coherent as well as non-coherent), has the major advantage of using simple hardware. Potential disadvantages of simple direct down conversion are: 4

5 1. A possible shift from zero of the expected "zero IF", (IF: Frequency). This is due to a possible frequency offset between the received carrier and LO signal. 2. Possible dc offset. This can be generated by the self converted LO signal leaked through the mixer RF port and reflected back to the mixer by the LNA, (Low Noise Amplifier), output port. In the presented system, the "non-zero IF" is eliminated by the coherent down conversion, while the dc offset is minimized by using a balanced configuration for the last LNA stage. BER "MONTE CARLO" simulation Results of the Monte Carlo statistical simulation are depicted in the following figures. The calculation has been performed for QPSK, (Quadrature Phase Shift Keying), modulation. As can be seen, for Eb/No=29 db the BER is lower than any practical need. This excess of BER will be used for making possible 60 GHz reflected and through wall transmission. Conclusion This new Analog APAA transceiver has excellent RF performance, fast locking ability, low production cost, simple MW hardware. It has no complex and expensive baseband digital processor. It is transparent to any kind of modulation method and is signal protocol independent. It can process uncompressed and compressed signals as well, at no additional cost. For all these characteristics the new technology can be used also for highly competitive last-mile transceiver links, not only for home applications. 5