STEL-2000A Data Sheet STEL-2000A+45 (45 MHz) STEL-2000A+20 (20 MHz) Digital, Fast Acquisition Spread Spectrum Burst Processor R Powered by ICminer.com Electronic-Library Service CopyRight 2003 FEATURES/BENEFITS ....................................................................................................................................... BLOCK DIAGRAM.............................................................................................................................................. PACKAGE OUTLINE.......................................................................................................................................... PIN CONFIGURATION...................................................................................................................................... GENERAL DESCRIPTION ................................................................................................................................. FUNCTIONAL BLOCKS..................................................................................................................................... Transmit and Receive Clock Generator Blocks.................................................................................. Input and Output Processor Blocks .................................................................................................... Differential Encoder Block.................................................................................................................... Transmitter PN Code Generator Block............................................................................................... BPSK/QPSK Modulator Block............................................................................................................. Frequency Control Register and NCO Block ..................................................................................... Downconverter Block............................................................................................................................ Receiver PN Code Register and PN Matched Filter Blocks............................................................. Power Detector Block ............................................................................................................................ Symbol Tracking Processor Block........................................................................................................ Differential Demodulator Block .......................................................................................................... Frequency Discriminator and Loop Filter Block ............................................................................... INPUT SIGNALS.................................................................................................................................................. OUTPUT SIGNALS.............................................................................................................................................. TRANSMITTER AND RECEIVER TEST POINTS .......................................................................................... CONTROL REGISTERS ...................................................................................................................................... Downconverter Registers...................................................................................................................... PN Matched Filter Registers................................................................................................................. Power Estimator Registers.................................................................................................................... Acquisition and Tracking Processor Registers .................................................................................. Demodulator Registers.......................................................................................................................... Output Processor Control Registers.................................................................................................... Transmit Control Registers................................................................................................................... REGISTER SETTING SEQUENCE ........................................................................................................ DECIMAL, HEX AND BINARY ADDRESS EQUIVALENTS....................................................................... REGISTER SUMMARY ....................................................................................................................................... ELECTRICAL CHARACTERISTICS ................................................................................................................. ABSOLUTE MAXIMUM RATINGS.................................................................................................... RECOMMENDED OPERATING CONDITIONS ............................................................................. D.C. CHARACTERISTICS.................................................................................................................... TRANSMITTER INPUT/OUTPUT TIMING .................................................................................... RECEIVER INPUT/OUTPUT TIMING.............................................................................................. MICROPROCESSOR INTERFACE TIMING ..................................................................................... APPENDIX I: THEORY OF OPERATION ...................................................................................................... Digital Downconversion ....................................................................................................................... Using the STEL-2000A with a Single ADC in Direct I.F. Sampling Mode ............................... Using the STEL-2000A with Two ADCs in Quadrature Sampling Mode ................................ Differential Demodulation ................................................................................................................... BPSK Demodulation ......................................................................................................................... QPSK Demodulation ........................................................................................................................ Frequency Error Generation................................................................................................................. Using the Modulator in the STEL-2000A Transmitter...................................................................... APPENDIX II: TYPICAL APPLICATION ....................................................................................................... 3 3 4 4 5 7 7 7 7 7 8 8 8 9 10 11 11 12 14 18 20 21 21 22 23 24 27 29 31 33 34 35 36 36 36 36 37 38 39 40 40 40 43 44 44 44 45 46 48 –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 2 FEATURES BENEFITS ■ Complete direct sequence spread spectrum burst transceiver in a single CMOS I.C. ■ High performance and high reliability with reduced manufacturing costs ■ Programmable functionality supports many different operational modes ■ Ideal for a wide range of wireless applications including data acquisition systems, transaction systems and wireless Local Area Networks (WLANs) ■ Supports PN chip rate of over 11 Mchips/sec. in transmit and receive modes (STEL-2000A+45 only) ■ Supports data rates up to 2.048 Mbps in compliance with FCC regulations (STEL-2000A+45 only) ■ Acquires within one symbol duration using digital PN Matched Filter ■ Fast response and very low overhead when operating in burst modes ■ Two independent PN sequences, each up to 64 chips long, for distinct processing of the acquisition/preamble symbol and subsequent data symbols ■ Allows high processing gain to maximize the acquisition probability, then reduced code length for increased data rate ■ Power management features ■ Low power consumption ■ Optional spectral whitening code generation ■ Randomizes data to meet regulatory requirements ■ Full or half duplex operation ■ Permits dual frequency (Frequency Division Duplex) or single frequency (Time Division Duplex) operation ■ 100-Pin PQFP packaging ■ Small footprint, surface mount BLOCK DIAGRAM (Figure 1) TXBITPLS TXTRKPLS TXIN INPUT DATA PROCESSOR TX OVERLAY CODE GENERATOR DIFFERENTIAL ENCODER QPSK MODULATOR TXIFOUT 7-0 TXIOUT TXQOUT BIT CLOCK TXMCHP TXIFCLK MTXEN MNCOEN MRXEN RXMABRT MFLD CSEL WR RESET DATA 7-0 ADDR 6-0 OEN TX CLOCK GENERATOR SYMBOL CLOCK TXIFCLK SIN FREQUENCY CONTROL REGISTER NCO OUTPUT DATA PROCESSOR TXACTIVE RXACTIVE TXTEST RXTEST 7-0 FREQUENCY DISCRIMINATOR & LOOP FILTER DIFFERENTIAL DEMODULATOR SYMBOL TRACKING PROCESSOR RXIIN 7-0 MATCHED FILTER DOWN CONVERTER RXIFCLK CORRECTED BIT CLOCK 2x CHIP CLOCK CORRECTED SYMBOL CLOCK SYMBOL CLOCK RXQIN 7-0 CHIP CLOCK RXSYMPLS RX PN CODE REGISTERS POWER DETECTOR RXQOUT RXIOUT RXDRDY COS CONTROL & µP INTERFACE RX OVERLAY CODE GENERATOR RXOUT TXCHPPLS TXACQPLS TX PN CODE GENERATORS CHIP CLOCK RX CLOCK GENERATOR RXIFCLK RXMSMPL RXMDET A –––––––––––––––––––––––––––––––––––––––––––––––––––– 3 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A PACKAGE OUTLINE (Figure 2) 0.941" ± 0.010" 0.742" ±0.005" 80 81 51 50 Top View 0.487" ±0.003" 0.705" ± 0.010" 0.11" nom. 0.009" ± 0.005" Pin 1 Identifier 31 100 1 0.031" ± 0.005" 30 Detail of pins 0.0256" ±0.002" 0.014" ± 0.002" 0.122" max. Package style: 100-pin PQFP. Thermal characteristics: θ ja = 41° C/W (STEL-2000A+45), θ ja = 68° C/W (STEL-2000A+20) PIN CONFIGURATION 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 VDD RXQIN 0 RXQIN 1 RXQIN 2 RXQIN 3 RXQIN 4 RXQIN 5 RXQIN 6 RXQIN 7 MRXEN VDD RXIFCLK VSS TXIFCLK VSS ÐÐÐÐÐÐ RESET MTXEN TXIN TXMCHP DATA0 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 ÐÐÐ WR ÐÐÐÐÐ CSEL VSS VDD ADDR0 ADDR1 ADDR2 ADDR3 ADDR4 ADDR5 ADDR6 VSS VDD 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 RXTEST7 RXTEST6 RXTEST5 RXTEST4 RXTEST3 RXTEST2 RXTEST1 RXTEST0 ÐÐÐÐ OEN VSS VDD RXSYMPLS RXSPLPLS ÐÐÐÐÐÐÐÐÐ RXDRDY RXQOUT RXIOUT RXOUT I.C. TXTEST TXACQPLS 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 TXTRKPLS TXCHPPLS TXBITPLS VSS VDD TXIFOUT7 TXIFOUT6 TXIFOUT5 TXIFOUT4 TXIFOUT3 TXIFOUT2 TXIFOUT1 TXIFOUT0 VSS VDD TXQOUT TXIOUT TXACTIVE N.C. VSS 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 VDD N.C. RXACTIVE RXMSMPL MFLD MNCOEN RXMABRT RXMDET VSS VDD RXIIN 0 RXIIN 1 RXIIN 2 RXIIN 3 RXIIN 4 RXIIN 5 RXIIN 6 RXIIN 7 N.C. VSS Note: I.C. denotes Internal Connection. Do not use for vias. –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 4 GENERAL DESCRIPTION the PN code. Since PN modulation is symbol-synchronous in the STEL-2000A, the data rate is defined by the PN chip rate and length of the PN code; i.e., by the number of chips per symbol. When operating with BPSK modulation, the maximum data rate for a PN code of length N is 11.264/N Mbps (STEL- The STEL-2000A is a programmable single-chip spread spectrum transceiver. The device performs all the digital processing required to implement a fast acquisition direct sequence (i.e., pseudonoise- or PNmodulated) spread spectrum full- or half-duplex system using differentially encoded BPSK, QPSK, or π/4 QPSK. A block diagram of the STEL-2000A is shown in Figure 1, while the package style and pin configuration are shown in Figure 2. The STEL-2000A is available in two speed grades; the STEL-2000A+20 (20 MHz maximum clock frequency), and the STEL-2000A+45 (45.056 MHz maximum clock frequency). The 45 MHz version features a high thermal conductivity package for superior heat dissipation, allowing the device to operate continuously at this speed. 2000A+45 only, 5/N Mbps in the STEL2000A+20). When operating with QPSK modulation (or π/4 QPSK with an external modulator), two bits of data are transmitted per symbol, and the maximum data rate for a PN code of length N is 22.528/N Mbps (STEL-2000A+45 only, 10/N Mbps in the STEL2000A+20). Conversely, for a given data rate Rb, the length N of the PN code employed must be such that the product of N x Rb is less than 11.264 (for BPSK) or 22.528 (for QPSK) Mcps (STEL-2000A+45 only). The STEL-2000A integrates the capabilities of a digital downconverter, PN matched filter, and DPSK demodulator into a single receiver, where the receiver input is the analog-to-digital converted I.F. signal. STEL-2000A transmit functions include a differential BPSK/QPSK encoder, PN modulator (spreader), and BPSK/QPSK modulator, where the transmitter output is a sampled digitally modulated signal ready for external digital-to-analog conversion (or, if preferred, the spread baseband signal may be output to an external modulator). These transceiver functions have been designed and integrated for the transmission and reception of bursts of spread data. In particular, the PN Matched Filter has two distinct PN coefficient registers (rather than a single one) in order to speed and improve signal acquisition performance. The STEL-2000A is thus optimized to provide reliable, high-speed wireless data communications. The data rate (Rb) and the PN code length (N), however, cannot generally be arbitrarily chosen. United States FCC Part 15.247 regulations require a minimum processing gain of 10 dB for unlicensed operation in the Industrial, Scientific, and Medical (ISM) bands, implying that the value of N must be at least 10. To implement such a short code, a Barker code of length 11 would typically be used in order to obtain desirable auto- and cross-correlation properties. With the STEL-2000A, a PN code length of 11 implies that the maximum data rate supported by the STEL-2000A in compliance with FCC regulations is 2.048 Mbps using differential QPSK (STEL-2000A+45 only). The STEL2000A further includes transmit and receive code overlay generators to insure that signals spread with such a short PN code length possess the spectral properties required by FCC regulations. The STEL-2000A receiver circuitry employs an NCO and complex multiplier referenced to RXIFCLK to perform frequency downconversion, where the input I.F. sampling rate and the clock rate of RXIFCLK must be identical. In Òcomplex inputÓ or Quadrature Sampling Mode, external dual analog-to-digital converters (ADCs) sample quadrature I.F. signals so that the STEL-2000A can perform true full single sideband downconversion directly from I.F. to baseband. At PN chip rates less than one-eighth the value of RXIFCLK, downconversion may also be effected using a single ADC in Òreal inputÓ or Direct I.F. Sampling Mode, as discussed in Appendix I. The STEL-2000A operates with symbol-synchronous PN modulation in both transmit and receive modes. Symbol-synchronous PN modulation refers to operation where the PN code is aligned with the symbol transitions and repeats once per symbol. By synchronizing a full PN code cycle over a symbol duration, acquisition of the PN code at the receiver simultaneously provides symbol synchronization, thereby significantly improving overall acquisition time. The receiver clock rate (RXIFCLK frequency) must be at least four times the receive PN spreading rate and is limited to a maximum speed of 45.056 MHz The input I.F. frequency is not limited by the capabilities of the STEL-2000A. To avoid destructive aliasing, the NCO should not be programmed above 50% of the I.F. sampling rate (the frequency of RXIFCLK); moreover, the signal bandwidth, NCO frequency, and (STEL-2000A+45 only, 20 MHz in the STEL2000A+20). As a result, the maximum supported PN chip rate is 11.264 Mchips/second (5 Mcps in the STEL-2000A+20), where a ÒchipÓ is a single ÒbitÓ of –––––––––––––––––––––––––––––––––––––––––––––––––––– 5 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A clock pulse if the correlation peak obtained by the PN Matched Filter fails to exceed the programmed detect threshold at the expected time during a given symbol. During each burst, a missed detect counter tallies each such event to monitor performance and allow a burst to be aborted in the presence of abnormally high interference. A timing gate circuit further minimizes the probability of false correlation peak detection and consequent false symbol clock generation due to noise or interference. I.F. sampling rate are all interrelated, as discussed in AppendixÊI. Higher I.F. frequencies, however, can be supported by programming the NCO to operate on in-band aliases as generated by the sampling process. For example, a spread signal presented to the STEL2000AÕs receiver ADCs at an I.F. frequency of f I.F., where fRXIFCLK < fI.F. < 2 x fRXIFCLK, can generally, as allowed by the signalÕs bandwidth, be supported by programming the STEL-2000AÕs NCO to a frequency of (fI.F. - fRXIFCLK), as discussed in Appendix I of this product specification. The maximum I.F. frequency is then limited by the track-and-hold capabilities of the ADC(s) selected. Signals at I.F. frequencies up to about 100 MHz can be processed by currently available 8-bit ADCs, but the implementation cost as well as the performance can typically be improved by using an I.F. frequency of 30 MHz or lower. Downconversion to baseband is then accomplished digitally by the STEL-2000A, with a programmable loop filter provided to establish a frequency tracking loop To minimize power consumption, individual sections of the device can be turned off when not in use. For example, the receiver circuitry can be turned off during transmission and, conversely, the transmitter circuitry can be turned off during reception when the STEL-2000A is operating in a half-duplex/time division duplex (TDD) system. If the NCO is not being used as the BPSK/QPSK modulator (i.e., if an external modulator is being used), the NCO can also be turned off during transmission to conserve still more power. The STEL-2000A is designed to operate in either burst or continuous mode: in burst mode, built-in symbol counters allow bursts of up to 65,533 symbols to be automatically transmitted or received, while, in continuous mode, the data is simply treated as a burst of infinite length. The STEL-2000AÕs use of a digital PN Matched Filter for code detection and despreading permits signal and symbol timing acquisition in just one symbol. The fast acquisition properties of this design are exploited by preceding each data burst with a single Acquisition/Preamble symbol, allowing different PN codes (at the same PN chip rate) to independently spread the Acquisition/Preamble and data symbols. In this way, a long PN code with high processing gain can be used for the Acquisition/Preamble symbol to maximize the probability of burst detection, and a shorter PN code can be used thereafter to permit a higher data rate. The fast acquisition characteristics of the STEL-2000A make it ideal for use in applications where bursts are transmitted relatively infrequently. In such cases, the device can be controlled so that it is in full ÒsleepÓ mode with all receiver, transmitter, and NCO functions turned off over the majority of the burst cycle, thereby significantly reducing the aggregate power consumption. Since the multiply operations of the PN Matched Filter consume a major part of the overall power required during receiver operation, two independent power-saving techniques are also built into the PN Matched Filter to reduce consumption during operation by a significant factor for both short and long PN spreading codes. The above features make the STEL-2000A an extremely versatile and useful device for spread spectrum data communications. Operating at its highest rates, the STEL-2000A is suitable for use in wireless Local Area Network implementations, while its programmability allows it to be used in a variety of data acquisition, telemetry, and transaction system applications. To improve performance in the presence of high noise and interference levels, the STEL-2000A receiverÕs symbol timing recovery circuit incorporates a Òflywheel circuitÓ to maximize the probability of correct symbol timing. This circuit will insert a symbol –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 6 FUNCTIONAL BLOCKS Transmit and Receive Clock Generators encoding of the signal is fundamental to operation of the STEL-2000AÕs receiver: the STEL-2000AÕs DPSK Demodulator computes ÒdotÓ and ÒcrossÓ product functions of the current and previous symbolsÕ downconverted I and Q signal components in order to perform differential decoding as an intrinsic part of DPSK demodulation. Timing in the transmitter and receiver sections of the STEL-2000A is controlled by the Transmit and Receive Clock Generator Blocks. These blocks are programmable dividers providing signals at the chip and symbol rates (as well as at multiples and submultiples of these frequencies) as programmed through the STEL-2000AÕs control registers. If desired, the com plete independence of the transmitter and receiver sections allows the transmit and receive clocks to be mutually asynchronous. Additionally, the STEL-2000A allows external signals to be provided as references for the transmit (TXMCHP) and receive (RXMSMPL) chip rates. Given the transmit PN chip rate, the PN-synchronous transmit symbol rate is then derived from the programmed number of PN chips per transmit symbol. At the receiver, symbol synchronization and the receive symbol rate are determined from processing of the PN matched filter output, or, if desired, can be provided from the programmed number of PN chips per receive symbol or an external symbol sync symbol, RXMDET. Burst control is achieved by means of the transmit and receive Symbols per Burst counters. These programmable 16-bit counters allow the STEL-2000A to operate automatically in burst mode, stopping at the end of each burst without the need of any external counters. The differential encoding scheme depends on whether the modulation format is to be BPSK or QPSK. For DBPSK, the encoding algorithm is straightforward: output bit(k) equals input bit(k) ⊕ output bit(kÐ1), where ⊕ represents the logical XOR function. For DQPSK, however, the differential encoding algorithm, as shown in Table 1, is more complex since there are now sixteen possible new states depending on the four possible previous output states and four possible new input states. New Input Previously Encoded OUT(I, Q)kÐ1 IN(I, Q)k 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 1 0 0 1 0 1 1 1 1 0 0 0 1 1 1 1 1 0 0 0 0 1 1 0 1 0 0 0 0 1 1 1 Newly Encoded OUT(I, Q) k Table 1. QPSK Differential Encoder Sequence Input and Output Processors Transmitter PN Code Generation When the transmitter and receiver are operating in QPSK mode, the data to be transmitted and the received data are processed in pairs of bits (dibits), one bit for the in-phase (I) channel and one for the quadrature (Q) channel. Dibits are transmitted and received as single differentially encoded QPSK symbols. Single-bit I/O data is converted to and from this format by the Input and Output Processors, accepting TXIN as the serial data to be transmitted and producing RXOUT as the serial data output. If desired, the received data is also available at the RXIOUT and RXQOUT pins in (I and Q) dibit format prior to dibitto-serial conversion. While receive timing is derived by the STEL-2000A Symbol Tracking Processor, transmit timing is provided by the Input Processor. In BPSK mode, the Input Processor will generate the TXBITPLS signal once per symbol to request each bit of data, while in QPSK mode it will generate the TXBITPLS signal twice per symbol to request the two bits of data corresponding to each QPSK symbol. When the STEL-2000A is used for burst signal operation, each burst is preceded by an Acquisition/Preamble symbol to facilitate acquisition. This Acquisition/Preamble symbol is automatically generated by the STEL-2000AÕs transmitter before information data symbols are accepted for transmission. Two separate and independent PN codes may be employed: one for spreading the Acquisition/Preamble symbol, and one for the subsequent information data symbols. As a result, a much higher processing gain may be used for signal acquisition than for signal tracking in order to improve burst acquisition performance. The Transmitter Acquisition/Preamble and Transmitter Data Symbol PN code lengths are completely independent of each other and can be up to 64 chips long. Transmit PN codes are programmed in the STEL-2000A as binary code values. The number of Transmitter Chips per Acquisition/Preamble Symbol is set by the value stored in bits 5-0 of address 43H , and the Transmitter Acquisition/Preamble Symbol Code coefficient values are stored in addresses 44H to 4BH. The number of Transmitter Chips per Data Differential Encoder Data to be transmitted is differentially encoded before being spread by the transmit PN code. Differential –––––––––––––––––––––––––––––––––––––––––––––––––––– 7 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A output sampling rate, R X I F C L K . Note that the maximum frequency of TXIFCLK is specified at 20 MHz when the internal modulator is being used for both the 20 and 45 MHz versions of the STEL-2000A. For operation at higher frequencies than this an external BPSK or QPSK modulator should be used in conjunction with the TXIOUT and TXQOUT signals. Symbol is set by the data stored in address 42H, and the Transmitter Data Symbol Code coefficient values are stored in addresses 4CH to 53H. A rising edge of the M T X E N input or of bit 1 of address 37 H causes the STEL-2000A to begin the transmit sequence by transmitting a single symbol using the Acquisition/Preamble PN code. The completion of transmission of the Acquisition/Preamble symbol is indicated with TXACQPLS, while the ongoing transmission of data symbols is signaled with TXTRKPLS. Data bits to be transmitted after the Acquisition/Preamble symbol are requested with TXBITPLS, where a single pulse per symbol requests data in BPSK mode and two pulses per symbol request data in QPSK mode. The user data symbols are then PN modulated using the Transmitter Data Symbol PN code. When the STEL-2000A is set to transmit in BPSK mode (by setting bitÊ0 of address 40H high), identical signals are applied to both the I and Q channels of the modulator so that the modulated output signal occupies only the first and third quadrants of the signal space defined in Table 2. Note that the modulator itself cannot generate π/4 QPSK signals, but the STEL-2000A can receive such signals and can be used with an external modulator for their transmission. Frequency Control Register and NCO The STEL-2000A incorporates a Numerically Controlled Oscillator (NCO) to synthesize a local oscillator signal for both the transmitter's modulator and receiver's downconverter. The NCO is clocked by the master receiver clock signal, RXIFCLK, and generates quadrature outputs with 32-bit frequency resolution. The NCO frequency is controlled by the value stored in the 32-bit Frequency Control Register, occupying 4 bytes at addresses 03H to 06H. To avoid destructive in-band aliasing, the NCO should not be programmed to be greater than 50% of RXIFCLK. As desired by the user, the output of the STEL-2000A receiverÕs Loop Filter can then be added or subtracted to adjust the NCO's frequency control word and create a closed-loop frequency tracking loop. If the receiver is disabled, either manually or automatically at the end of a burst, the Loop Filter output correcting the NCOÕs Frequency Control Word is disabled. When operating the transmitter and receiver simultaneously, however, the receiverÕs frequency tracking loop affects the NCO signals to both the receive and transmit sides; this can either be used to advantage or must be compensated for in the system design. The PN spreading codes are XORed with the data bits (in BPSK mode) or bit pairs (in QPSK mode) to transmit one complete code sequence for every Acquisition/Preamble and data symbol at all times. The resulting spread I and Q channel signals are brought out as the TXIOUT and TXQOUT signals for use by an external modulator and are also fed into the STEL-2000A's internal modulator. In BPSK mode, only TXIOUT is used by the STEL-2000AÕs modulator. If an external QPSK modulator is used, the carrier should be modulated as shown in Table 2 to be compatible with the STEL-2000A receiver. I, Q Bits Signal Quadrant Quadrant diagram 0 0 First 1 0 Second 2nd. 1st. 1 1 Third 3rd. 4th. 0 1 Fourth Table 2. QPSK Differential Encoder Sequence BPSK/QPSK Modulator The STEL-2000A incorporates an on-chip BPSK/QPSK modulator which modulates the encoded and spread transmit signal with the sine and cosine outputs of the STEL-2000AÕs NCO to generate a digitized I.F. output signal, TXIFOUT7-0. Since the NCO operates at a rate defined by RXIFCLK, the BPSK/QPSK modulator output is also generated at this sampling rate, and, consequently, TXIFCLK must be held common with RXIFCLK to operate the STEL2000AÕs BPSK/QPSK Modulator. The digital modulator output signal can then be fed into an external 8-bit DAC (operating at RXIFCLK) to generate an analog I.F. transmit signal, where the chosen I.F. is the STEL-2000A's programmed NCO frequency or one of its aliases with respect to the Downconverter The STEL-2000A incorporates a Quadrature (Single Sideband) Downconverter which digitally downconverts the sampled and digitized receive I.F. signal to baseband. Use of the Loop Filter and the NCO's builtin frequency tracking loop permits the received signal to be accurately downconverted to baseband. The Downconverter includes a complex multiplier in which the 8-bit receiver input signal is multiplied by the sine and cosine signals generated by the NCO. In Quadrature Sampling Mode, two ADCs provide quadrature (complex) inputs IIN and QIN , while, in Direct I.F. Sampling Mode, a single ADC provides IIN as a real input. The input signals can be accepted in –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 8 twice the nominal PN chip rate for proper receiver operation and less than or equal to one-half the frequency of RXIFCLK. If twice the PN chip rate is a convenient integer sub-multiple of RXIFCLK, then an internal clock can be derived by frequency dividing RXIFCLK according to the divisor stored in bits 5-0 of address 02H; otherwise, an external baseband sampling clock provided by RXMSMPL must be used. either twoÕs complement or offset binary formats according to the setting of bit 3 of address 01H. In Direct I.F. Sampling Mode, the unused RXQIN Q channel input (QIN) should be held to ÒzeroÓ according to the ADC input format selected. The outputs of the DownconverterÕs complex multiplier are then: IOUT = IIN . cos(ωt) Ð QIN . sin(ωt) QOUT = IIN . sin(ωt) + QIN . cos(ωt) The I.F. sampling rate, the baseband sampling rate, and the input signal levels determine the magnitudes of the Integrate and Dump FiltersÕ accumulator outputs, and a programmable viewport is provided at the outputs of the Integrate and Dump Filters to select the appropriate output bits as the 3-bit inputs to the PN Matched Filter. The viewport circuitry here and elsewhere within the STEL-2000AÕs receiver is designed with saturation protection so that extreme values above or below the selected range are limited to the correct maximum or minimum value for the selected viewport range. The viewports for the I and Q channels of the Integrate and Dump Filters are controlled by the values stored in bits 7-4 of address 01H. where ω = 2πfNCO These outputs are fed into the I and Q channel Integrate and Dump Filters. The Integrate and Dump Filters allow the samples from the complex multiplier (at the I.F. sampling rate, the frequency of RXIFCLK) to be integrated over a number of sample periods. The dump rate of these filters (the baseband sampling rate) can be controlled either by an internally generated dump clock or by an external input signal (RXMSMPL) according to the setting of bit 0 of address 01H. Note that, while the receiver will extract exact PN and symbol timing information from the received signal, the baseband sampling rate must be AAAA AAAAAAAAAAAAAAAAAAAA AAAA AAAAAAAAAAAAAAAAAAAA AAAAA AAAA AAAAAAAAAAAAAAAAAAAA Σ AAAAAAAAAAAAAAAAAAAA AA AAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAA AAAA AAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAA AA Σ AA AAAAAAAAAAAAAAAAAAAA AAAA AAAAAAAAAAAAAAAAAAAA AAAA AAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAA AAAAA AAAA AAAAAAAAAAAAA AAAAAAAAAAAAA AAAAAAAAAAAAA AAAAAAAAAAAAA AAAAAAAAAAAAA AAAAAAAAAAAAA Receiver PN Code Register and PN Matched Filter 3 I IN TA P0 FEP BLOCK 2T T TA P1 2T TA P2 I DELA Y TA P64 REGISTER BLOCK 2T TA P62 TA P63 2T 2T I MULTIPLIER A RRA Y BLOCK MD COEFFICIENT MEMORY A ND REGISTER BLOCK MP MD MP I A DDER + + + + + + + Q A DDER SEL MP SEL MD SEL MP SEL MD SEL SEL MP A QU.-PRE./DA TA COEF. SEL. + + + + + + + MD Q MULTIPLIER A RRA Y BLOCK FEP BLOCK 3 Q IN T 2T 2T 2T Q DELA Y REGISTER BLOCK 2T POST PROCESSOR BLOCK A BS 2T 10 MA GNITUDE GENERA TOR A BS THRESHOLD REGISTERS 8 10 COMPA RA TOR V IEWPORT A ND SA T. (OUTPUT CONTROL) 8 I SUM MA G 8 Q SUM 10 DET Figure 3. Matched Filter Detail symbols for information transmittal. Complementing operation of the STEL-2000AÕs transmitter, two separate and independent PN codes may be employed in the receiverÕs PN Matched Filter, one for despreading As discussed for the STEL-2000A transmitter, the STEL-2000A is designed for burst signal operation in which each burst begins with a single Acquisition/Preamble symbol and is then followed by data –––––––––––––––––––––––––––––––––––––––––––––––––––– 9 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A stored PN code coefficients at the baseband sampling rate; i.e., twice per chip. The 3-bit signals from each tap in the PN Matched Filter are multiplied by the corresponding coefficient in two parallel tapped delay lines. Each delay line consists of 64 multipliers which multiply the delayed 3-bit signals by zero or ±1 according to the value of the tap coefficient. The products from the I and Q tapped delay lines are added together in the I and Q Adders to form the sums of the products, representing the complex crosscorrelation factor. The correlation I and Q outputs are thus: the Acquisition/Preamble symbol, and one for the information data symbols. The code lengths are completely independent of each other and can be each up to 64 chips long. A block diagram of the PN Matched Filter is shown in Figure 3. The STEL-2000A contains a fully programmable 64tap complex (i.e., I and Q channel) PN Matched Filter with coefficients which can be set to ±1 or zero according to the contents of either the Acquisition/Preamble or Data Symbol Code Coefficient Registers. By setting the coefficients of the end taps of the filter to zero, the effective length of the filter can be reduced for use with PN codes shorter than 64 bits. Power consumption may also be reduced by turning off those blocks of 7 taps for which all the coefficients are zero, using bits 6-0 of address 39H . Each ternary coefficient is stored as a 2-bit number, so that a PN code of length N is stored as N 2-bit non-zero PN coefficients. Note that, as a convention, throughout this document the first PN Matched Filter tap encountered by the signal as it enters the I and Q channel tapped delay lines is referred to as ÒTap 0.Ó Tap 63 is then the last tap of the PN Matched Filter. n = 63 Output(I, Q) = Σ Datan(I, Q) * Coefficientn(I, Q) n=0 These I and Q channel PN Matched Filter outputs are 10-bit signals, with I and Q channel programmable viewports provided to select the appropriate output bits as the 8-bit inputs to the Power Detector and DPSK Demodulator blocks. Both I and Q channel viewports are jointly controlled by the data stored in bits 1-0 of address 28H and are saturation protected. Two power saving methods are used in the PN Matched Filter of the STEL-2000A. As discussed previously, the first method allows power to be shut off in the unused taps of the PN Matched Filter when the filter length is configured to be less than 64 taps. The second method is a proprietary technique that (transparently to the user) shuts down the entire PN Matched Filter during portions of each symbol period. The start of each burst is expected to be a single symbol PN-spread by the Acquisition/Preamble code. The receiver section of the STEL-2000A is automatically configured into acquisition mode so that the Matched Filter Acquisition/Preamble Coefficients stored in addresses 07H to 16H are used to despread the received signal. Provided that this symbol is successfully detected, the receiver will automatically switch from acquisition mode, and the Matched Filter Data Symbol Coefficients stored in addresses 17 H to 26H will then be used to despread subsequent symbols. Power Detector The complex output of the PN Matched Filter is fed into a Power Detector which, for every cycle of the internal baseband sampling clock, computes the magnitude of the vector of the I and Q channel correlation sums, I2 (k)+Q 2 (k) , where the magnitude is approximated as To allow the system to sample the incoming signal asynchronously (at the I.F. sampling rate) with respect to the PN spreading rate, the PN Matched Filter is designed to operate with two signal samples (at the baseband sampling rate) per chip. A front end processor (FEP) operating on both the I and Q channels averages the incoming data over each chip period by adding each incoming baseband sample to the previous one: Max{Abs(I),Abs(Q)} + 1/2 Min{Abs(I), Abs(Q)}. This 10-bit value represents the power level of the correlated signal during each chip period and is used in the Symbol Tracking Processor. i.e., FEPOUT = FEPIN (1 + zÐ1) Symbol Tracking Processor After the addition, the output of the FEP is rounded to a 3-bit offset 2Õs complement word with an effective range of ±3.5. such that the rounding process does not introduce any bias to the data. The FEP can be disabled by setting bit 0 of address 27 H to a 1, but for normal operation the FEP should be enabled. The output of the Power Detector Block represents the signal power during each chip period. Ideally, this output will have a high peak value once per symbol (i.e., once per PN code cycle) when the code sequence of the received signal in the PN Matched Filter is the same as (and is aligned in time with) the reference PN code used in the PN Matched Filter. At that instant, the I and Q channel outputs of the PN Matched Filter The PN Matched Filter computes the cross-correlation between the I and Q channel signals and the locally –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 10 threshold. A maximum power selector circuit is incorporated in the STEL-2000A to choose the highest of any three consecutive power levels each time this occurs, thereby enhancing the probability that the optimum symbol timing will be chosen in such cases. If desired, this function can be disabled by setting bit 3 of address 30H high. are, theoretically, the optimally despread I and Q symbols. To detect this maximum correlation in each symbol period, the signal power value is compared against a 10-bit user-programmable threshold value. A symbol clock pulse is generated each time the power value exceeds the threshold value to indicate a symbol detect. Since the Acquisition/Preamble symbol and subsequent data symbols can have different PN codes with different peak correlation values (which depend on the PN code length and code properties), the STEL-2000A is equipped with two separate threshold registers to store the Acquisition/Preamble Threshold value (stored in addresses 29 H and 2AH) and the Data Symbol Threshold value (stored in addresses 2B H and 2CH). The device will automatically use the appropriate value depending on whether it is in acquisition mode or not. The STEL-2000A also includes a circuit to keep track of missed detects; i.e., those cases where no peak power level exceeds the set threshold. An excessively high rate of missed detects is an indication of poor signal quality and can be used to abort the reception of a burst of data. The number of symbols expected in each receive burst, up to a maximum of 65,533, is stored in addresses 2EH and 30 H. A counter is used to count the number of missed detects in each burst, and the system can be configured to automatically abort a burst and return to acquisition mode if this number exceeds the Missed Detects per Burst Threshold value stored in address 2FH. Under normal operating conditions, the STEL-2000A will automatically return to acquisition mode when the number of symbols processed in the burst is equal to the value of the data stored in address 2E H and 30H . To permit the processing of longer bursts or continuous data, this function can be disabled by setting bit 6 of address 30H high. Since spread spectrum receivers are frequently designed to operate under extremely adverse signalto-noise ratio conditions, the STEL-2000A is equipped with a Òflywheel circuitÓ to enhance the operation of the symbol tracking function by introducing memory to the PN Matched Filter operation. This circuit is designed to ignore false detects at inappropriate times in each symbol period and to insert a symbol clock pulse at the appropriate time if the symbol detection is missed. The flywheel circuit operates by its a priori knowledge of when the next detect pulse is expected. The expected pulse will occur one symbol period after the last correctly detected one, and a window of ±1 baseband sample time is therefore used to gate the detect pulse. Any detects generated outside this time window are ignored, while a symbol detect pulse will be inserted into the symbol clock stream if the power level does not exceed the threshold within the window, corresponding to a missed detect. An inserted symbol detect signal will be generated precisely one symbol after the last valid detect, the nominal symbol length being determined by the value of Rx Chips Per Data Symbol stored in address 2D H. Differential Demodulator Both DPSK demodulation and carrier discrimination are supported in the STEL-2000A receiver by the calculation of ÒdotÓ and ÒcrossÓ products using the despread I and Q channel information generated by the PN Matched Filter for the current and previous symbols. A block diagram of the DPSK DemodulatorÕs I and Q channel processing is shown in Figure 4. Let Ik and Q k represent the I and Q channel outputs, respectively, for the kth symbol. The dot and cross products can then be defined as: Dot(k) = Ik Ik-1 + Qk Qk-1; and, Cross(k) = Qk Ik-1 - Ik Qk-1. Examination of these products in the complex plane reveals that the dot and cross products are the real and imaginary results, respectively, of complex multiplication of the current and previous symbols. The dot product alone thus allows determination of the phase shift between successive BPSK symbols, while the dot and cross products together allow The cross-correlation characteristics of a noisy received signal with the noise-free local PN code used in the STEL-2000AÕs PN Matched Filter may result in ÒsmearingÓ of the peak power value over adjacent chip periods. Such smearing can result in two or three consecutive power values (typically, the on-time and one-sample early and late values) exceeding the –––––––––––––––––––––––––––––––––––––––––––––––––––– 11 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A I Ik 8 + Ik – 1 – SYMBOL ROTATOR 8 DOT(k) = Ik . Ik – 1 + Q k . Q k – 1 17 C ROSS(k) = Q k . Ik – 1 – Ik . Q k – 1 + Qk – 1 Q + Σ T DELAY 1 + jX X = 0, +1, -1 Ø = 0, +45°, -45° 17 Σ Qk Figure 4. Differential Demodulator Detail For QPSK signals, a fixed phase rotation of π/4 (45°) is introduced in the DPSK Demodulator to the previous symbol to simplify the decision algorithm. Rotation of the previous symbol is controlled by the settings of bits 0 and 1 of address 33H,allowing the previous symbol to be rotated by 0° or ±45°. As noted, for BPSK or π/4 QPSK signals, a rotation of 0° should be programmed, but, for QPSK signals, a Ð45° signal rotation must be programmed to optimize the constellation boundaries in the comparison process between successive symbols. Note also that introduction of a ±45° rotation introduces a scaling factor of 1 /√2 to the signal level in the system as discussed in Appendix I, where this factor should be taken into account when calculating optimum signal levels and viewport settings after the DPSK Demodulator. determination of the integer number of π/2 phase shifts between successive QPSK symbols. Differential encoding of the source data implies that an absolute phase reference is not required, and thus knowledge of the phase shift between successive symbols derived from the dot and cross products unambiguously permits correct demodulation. Implementation of this approach is simplified if the polarities (i.e., the signs) alone of the dot and cross products provide the information required to make the correct symbol decision. For BPSK and π/4 QPSK signals, no modifications are needed: in BPSK, the sign of the dot product fully captures the signal constellation, while, in π/4 QPSK, the signal constellation intrinsically includes the phase rotation needed to align the decision boundaries with the four possible combinations of the dot and cross product polarities. Frequency Discriminator and Loop Filter – DOT SIGN QPSK 17 BPSK/ BPSK QPSK SELEC T C ROSS K2 4 MUX SIGN Σ + VIEW PORT Σ 29 AFC VIEW PORT C ONTROL ÷4 K1 29 8 FRZ. LF A 32 REG. INV. LF Σ Σ 31 TO NC O FC W Figure 5. Frequency Discriminator and Loop Filter Detail –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 12 to select the desired output bits as the 8-bit input to the Loop Filter Block. The viewport is controlled by the value stored in bits 7-4 of address 33 H. The Frequency Discriminator uses the dot and cross products discussed above to generate the AFC signal for the frequency acquisition and tracking loop, as illustrated in Figure 5. The specific algorithm used depends on the signal modulation type and is controlled by the setting of bit 2 of address 33 H. When bit 2 is set low, the Frequency Discriminator circuit is in BPSK mode and the following algorithm is used to compute the Frequency Discriminator (FD) function: The Loop Filter is implemented with a direct gain (K1) path and an integrated or accumulated (K2) path to filter the Frequency Discriminator error signal and correct the frequency tracking of the Downconverter. The order of the Loop Filter transfer function can be set by enabling or disabling the K1 and K2 paths, and the coefficient values can be adjusted in powers of 2 from 20 to 221. The Loop Filter transfer function is: FD = Cross x Sign[Dot], where Sign[Dot] represents the polarity of the argument. When bit 2 is set high, the discriminator circuitry is in QPSK mode and the carrier discriminator function is instead calculated as: Transfer Fn. = K1 + 1/4 K2 z Ð1 1ÊÐÊzÐ1 The factor of 1/4 results from truncation of the 2 LSBs of the signal in the integrator path of the loop so that, when added to the signal in the direct path, the LSBs of the signals are aligned. The coefficients K1 and K2 are defined by the data stored in bits 4-0 of addresses 35H and 34 H, respectively. In addition, bit 5 of addresses 35 H and 34H control whether the K1 and K2 paths, respectively, are enabled. These parameters thus give the user full control of the Loop Filter characteristics. FD = (Cross x Sign[Dot]) Ð (Dot x Sign[Cross]). In both cases, the Frequency Discriminator function provides an error signal that reflects the change in phase between successive symbols. With the symbol period known, the error signal can equivalently be seen as a frequency error signal. As a practical matter, the computation of the Frequency Discriminator function results in a 17-bit signal, and a programmable saturation protected viewport is provided –––––––––––––––––––––––––––––––––––––––––––––––––––– 13 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A INPUT SIGNALS RXIIN7-0 (Pins 91-98) PN chip rate. Receiver In-Phase Input. RXIIN is an 8-bit input port for in-phase data from external A/D converters. Data may be received in either two's complement or offset binary format as selected by bit 3 of address 01 H. The sampling rate of the RXIIN signals (the I.F. sampling rate of the A/Ds) may be independent of the baseband sampling rate (the Downconverter integrate and dump rate) and the PN chip rate, but must be equal to RXIFCLK and at least two times greater than the baseband sampling rate. Since the baseband sampling rate must be set at twice the PN chip rate, the I.F. sampling rate must thus be at least four times the PN chip rate. Data on the pins is latched and processed by RXIFCLK. When bit 0 of address 01H is set high, a rising edge on RXMSMPL will initiate a baseband sampling clock pulse to the Integrate and Dump filters and subsequent circuitry (e.g., PN Matched Filter, DPSK Demodulator, Power Estimator, etc.). The rising edge of RXMSMPL is synchronized internally so that, on the second rising edge of RXIFCLK that follows the rising edge of RXMSMPL, a pulse is internally generated that clocks the circuitry that follows. On the third rising RXIFCLK edge, the contents of the Integrate and Dump Filters of the Downconverter are transferred to the PN Matched Filter. The extra one RXIFCLK delay before transfer of the contents of the filters enables the internally generated baseband sampling clock to be free of race conditions at the interface between the Downconverter and PN Matched Filter. RXQIN 7-0 (Pins 2-9) Receiver Quadrature-Phase Input. RXQIN is an 8bit input port for quadrature-phase data from external A/D converters. Data may be received in either two's complement or offset binary format as selected by bit 3 of address 01H. As with RXIIN, the sampling rate of the RXQIN signals may be independent of the baseband sampling and PN chip rates in the receiver, but must be at least two times greater than the baseband sample rate (or, equivalently, at least four times greater than the PN chip rate). Data on the pins is latched and processed by RXIFCLK. RXMDET (Pin 88) Receiver Manual Detect. RXMDET enables the user to externally generate symbol timing, bypassing and overriding the internal symbol power estimation and tracking circuitry. This function may be useful when the dynamic characteristics of the transmission environment require unusual adjustments to the symbol timing. When bit 0 of address 30H is set high (Manual Detect Enable) and when bit 0 of address 31H is set low, a rising edge of RXMDET will generate a symbol correlation detect pulse. The function can also be performed by means of bit 0 of address 31H. T h e RXMDET input and bit 0 of address 31 H are logically ORed together so that, when either one is held low, a rising edge on the other triggers the manual detect function. The rising edge of RXMDET is synchronized internally so that, on the second rising edge of the baseband sampling clock that follows the rising edge of RXMDET, the correlated outputs of the PN Matched Filter I and Q channels will be transferred to the DPSK demodulator. Note that if the STEL-2000A is to be used in Direct I.F. Sampling Mode, then the I.F. signal should be input to the RXIIN input port only. RXQIN must then be held to arithmetic zero according to the chosen ADC format as selected by bit 3 of address 01H. In other words, to support Direct I.F. Sampling, RXQIN must be tied to a value of 7FH or 80H if offset binary input format has been selected or to a value of 00H if twoÕs complement input format has been selected. RXMSMPL (Pin 84) Receiver Manual Sample Clock. RXMSMPL enables the user to externally generate (independent of the I.F. sampling clock, RXIFCLK) the baseband sampling clock used for all processing after the digital downconverter, including the dump rate of the Integrate and Dump filters. This feature is useful in cases where a specific baseband sample rate is required that may not be derived by the internal sample rate timing generator which generates clock signals at integer sub-multiples of RXIFCLK. The signal is internally synchronized to RXIFCLK to avoid intrinsic race or hazard timing conditions. There must be at least two cycles of RXIFCLK to every cycle of RXMSMPL, and RXMSMPL should be set to twice the nominal receive RXMABRT (Pin 87) Receiver Manual Abort. RXMABRT enables the user to manually force the STEL-2000A to cease reception of the current burst of data symbols and prepare for acquisition of a new burst. This function can be used to reset the receiver and prepare to receive a priority transmission signal under precise timing control, giving the user the ability to control the status of the receiver for reasons of priority, signal integrity, etc. When bit 0 of address 32H is set low, a rising edge on –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 14 this power down sequence. Upon reactivation (when either MNCOEN or bit 0 of address 37H return high), the NCO must be reloaded with frequency control information either by means of the MFLD input or by writing 01 H into address 00H. RXMABRT will execute the abort function. The function can also be performed under microprocessor control by means of bit 0 of address 32 H. T h e RXMABRT input and bit 0 of address 32 H are logically ORed together so that, when either one is held low, a rising edge on the other triggers the abort function. The second rising edge of the baseband sampling clock that follows a rising edge of RXMABRT will execute the abort and also clear the symbols-perburst, samples-per-symbol, and missed-detects-perburst counters. The counters will be reactivated on the detection of the next burst preamble or by a manual detect signal. MTXEN (Pin 17) Manual Transmitter Enable. A rising edge on MTXEN causes the transmit sequence to begin, where the STEL-2000A first transmits a single Acquisition/Preamble symbol followed by data symbols. MTXEN should be set low after the last symbol has been transmitted. When MTXEN is set low, power consumption of the transmitter circuit is minimized. MTXEN operates independently of MRXEN and MNCOEN, where these signals have similar control over the receive and NCO circuitry, respectively. RXIFCLK (Pin 12) Receiver I.F. Clock. RXIFCLK is the master clock of the NCO and all the receiver blocks. All clocks in the receiver section and the NCO, internal or external, are generated or synchronized internally to the rising edge of RXIFCLK. The frequency of RXIFCLK must be at least four times the PN chip rate of the received signal. When bit 0 of address 01 H is set low, the baseband sampling clock, required to be at twice the nominal PN chip rate, will be derived from RXIFCLK according to the setting of bits 5-0 of addressÊ02H. M T X E N performs the same function as bit 1 of address 37H. and these two signals are logically ORed together to form the overall control function. When bit 1 of address 37H is set low, MTXEN controls the activity of the transmitter circuitry, and, when MTXEN is set low, bit 1 of address 37 H controls the activity of the transmitter circuitry. A rising edge on either MTXEN or bit 1 (whichever is in control, as defined above) initiates a transmit sequence. A falling edge initiates a reset sequence on the following TXIFCLK cycle to disable all of the transmitter data path, although the user programmable control registers are not affected by the power down sequence. MNCOEN (Pin 86) Manual NCO Enable. MNCOEN allows the power consumed by the operation of the NCO circuitry to be minimized when the STEL-2000A is not receiving and not transmitting data. The NCO can also be disabled while the STEL-2000A is transmitting as long as the STEL-2000A's on-chip BPSK/QPSK modulator is not being used. With the instantaneous acquisition properties of the PN Matched Filter, it is often desirable to shut down the receiver circuitry to reduce power consumption, resuming reception periodically until an Acquisition/Preamble symbol is acquired. Setting MNCOEN low will do this by holding the NCO in a reset state. After MNCOEN has been set high again to re-activate the NCO it will be necessary to reload the frequency control word into the NCO. Note that MNCOEN operates independently of MTXEN and MRXEN, where those pins have similar control over the transmit and receive circuitry, respectively. MRXEN (Pin 10) Manual Receiver Enable. MRXEN allows power consumption of the STEL-2000A receiver circuitry to be minimized when the device is not receiving. With the instantaneous acquisition properties of the PN Matched Filter, it is often desirable to shut down the receiver circuitry to reduce power consumption, resuming reception periodically until an Acquisition/Preamble symbol is acquired. Setting MRXEN low reduces the power consumption substantially. When MRXEN is set high, the receiver will automatically power up in acquisition mode regardless of its prior state when it was powered down. M R X E N operates independently of MTXEN and MNCOEN, where these signals have similar control over the transmit and NCO circuitry, respectively. MNCOEN performs the same function as bit 0 of address 37H, and these two signals are logically ORed together to form the overall control function. When this bit is set low, MNCOEN controls the activity of the NCO circuitry; when MNCOEN is set low, bit 0 of address 37H controls the activity of the NCO circuitry. When either bit 0 or MNCOEN (whichever is in control, as defined above) goes low, a reset sequence occurs on the following RXIFCLK cycle to effectively disable all of the NCO circuitry, although the user programmable control registers are not affected by M R X E N performs the same function as bit 2 of address 37H, and these two signals are logically ORed together to form the overall control function. When bit 2 of address 37H is set low, MRXEN controls the activity of the receiver circuitry and, when MRXEN is set low, bit 2 of address 37H controls the activity of the receiver circuitry. When either MRXEN or bit 2 (whichever is in control, as defined above) goes low, a reset sequence begins on the following RXIFCLK –––––––––––––––––––––––––––––––––––––––––––––––––––– 15 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A cycle and continues through a total of six RXIFCLK cycles to virtually disable all of the receiver data paths, although the user programmable control registers are not affected by the power down sequence. Note that this reset sequence will also reset the contents of address 38H to 0. If the R X T E S T 7-0 bus is being used for reading any function except the Matched Filter I and Q inputs the value required must be re-written after re-enabling the receiver. When bit 2 of address 40H is set high, a rising edge on TXMCHP will generate the chip clock to the differential encoder and the following circuitry (Acquisition/Preamble and Data Symbol PN spreaders, etc.). The rising edge of TXMCHP is synchronized internally so that, on the third rising edge of TXIFCLK following the rising edge of TXMCHP, the PN code combined with the differen tially encoded signal will change, generating the next chip. TXIN (Pin 18) TXIFCLK (Pin 14) Transmit Input. TXIN supports input of the information data to be transmitted by the STEL-2000A. In BPSK mode, the transmitter requires one bit per symbol period; in QPSK mode, two bits are required per symbol period. Transmitter I.F. Clock. TXIFCLK is the master clock of the transmitter. All transmitter clocks, internal or external, are generated or synchronized internally to the rising edge of TXIFCLK. The rate of TXIFCLK must be at least twice the transmit PN chip rate. It may be convenient to use the same external signal for both TXIFCLK and RXIFCLK, in which case the frequency of TXIFCLK will be at least four times the PN chip rate as required for RXIFCLK. Moreover, if the STEL-2000A's on-chip BPSK/QPSK Modulator is to be used, TXIFCLK and RXIFCLK must be identical and their frequencies must not exceed 20 MHz To initiate and enable transmission of the data, the user must raise MTXEN high. Data for transmission is requested with TXBITPLS, where one or two pulses per symbol are generated depending on whether the device is in BPSK or QPSK mode as set by bit 0 of address 40H. To allow monitoring of the state of the transmitter, the STEL-2000A will pulse TXACQPLS after the initial Acquisition/Preamble symbol is transmitted; the transmission of each subsequent symbol is indicated by pulses of TXTRKPLS. MFLD (Pin 85) Manual Frequency Load. MFLD is used to load a frequency control value into the NCO. The NCO may be loaded in various ways, but MFLD provides a synchronized external method of updating the NCO, while the other methods involve setting bit 0 of address 00H or using the programmable loop filter timing circuitry. MFLD is internally synchronized to RXIFCLK to avoid internal race or hazard timing conditions. If programmed for BPSK mode, data is requested by the STEL-2000A by a rising edge of output signal TXBITPLS, where TXBITPLS is generated once per symbol, one chip period before the end of the current symbol. At the end of the symbol duration, the TXIN data is latched into the device. TXBITPLS falls low immediately following the rising edge of TXIFCLK, which latches the TXIN value, and is generated repeatedly at the symbol rate as long as the input signal MTXEN remains high. The M F L D input and bit 0 of address 00H are logically ORed together so that, when either one is held low, a rising edge on the other triggers the frequency load function manually. The rising edge of MFLD is synchronized internally so that, on the sixth following rising edge of RXIFCLK, the frequency control word is completely registered into the NCO accumulator. The frequency load command must not be repeated until the six RXIFCLK cycle delay is completed. In QPSK mode, data is requested by the STEL-2000A by a rising edge of output signal TXBITPLS, where this signal is generated twice per symbol, first one chip period before the middle of the symbol and then one chip period before the end of the symbol. TXBITPLS requests the data exactly one chip cycle before latching the T X I N data into the device. TXBITPLS falls low immediately following the rising edge of TXIFCLK, which latches the TXIN value. –––– WR (Pin 28) ÐÐÐ Write Bar. WR is used to latch user-configurable information into the control registers. It is important to note that the control registers are transparent ÐÐÐ latches while WR is set low. The information will be ÐÐÐ latched when WR returns high. D A T A7-0 and ÐÐÐ ADDR 6-0 should be stable while W R is set low in order to avoid undesirable effects. TXMCHP (Pin 19) Transmit Manual Chip Pulse. TXMCHP enables the user to provide the PN chip rate clock pulses from an external source. This feature is useful in cases where a specific chip rate is required that cannot be derived by the internal clock generator which generates clocks of integer sub-multiples of TXIFCLK. The signal is internally synchronized to TXIFCLK to avoid intrinsic race or hazard timing conditions. DATA7-0 (Pins 20-27) Data Bus. DATA7-0 is an 8-bit microprocessor interface bus that provides access to all internal control –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 16 ––––– OEN (Pin 49) register inputs for programming. DATA7-0 is used in ÐÐÐ conjunction with the ADDR6-0 and WR signals to set the values of the control registers. ÐÐÐÐÐ Output Enable Bar. OEN is provided to enable or ÐÐÐÐÐ disable the RXTEST7-0 output bus. When OEN is set high, the RXTEST 7-0 bus will have a high impedance, allowing it to be connected to other busses, such as ÐÐÐÐÐ DATA 7-0. When OEN is set low, the RXTEST 7-0 bus will be active, allowing the RXTEST function selected to be accessed. ADDR6-0 (Pins 32-38) Address Bus. ADDR 6-0 is a 7-bit address bus that selects the control register location into which the information provided on the DATA 7-0 bus will be ÐÐÐ written. ADDR 6-0 is used in conjunction with WR and D A T A 7-0 to write the information into the registers. ––––––– RESET (Pin 16) ÐÐÐÐÐÐ Reset Bar. RESET is the master reset of the STEL2000A, clearing the control registers as well as the contents within the receiver, transmitter, and NCO ÐÐÐÐÐÐ data paths when it is set low. Setting RESET high enables operation of the circuitry. –––––– CSEL (Pin 29) ÐÐÐÐÐ Chip Select Bar. CSEL is provided to enable or disable the microprocessor operation of the STEL-2000A. ÐÐÐÐÐ ÐÐÐ When CSEL is set high, the A D D R 6-0 and WR become disabled and have no effect on the device. ÐÐÐÐÐ When CSEL is set low, the device is in its normal ÐÐÐ mode of operation and ADDR6-0 and WR are active. –––––––––––––––––––––––––––––––––––––––––––––––––––– 17 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A OUTPUT SIGNALS TXIOUT (Pin 77) chip period before the middle of the symbol and then one chip period before the end of the symbol. TXBITPLS requests the data exactly one chip cycle period before the T X I N data is latched into the device. TXBITPLS falls low immediately following the rising edge of TXIFCLK, where TXIFCLK latches the TXIN value. Transmitter In-Phase Output. TXIOUT is the inphase output transmission signal that has been differentially encoded and PN spread. TXIOUT changes on the rising edge of TXIFCLK following the falling edge of TXCHPPLS. TXQOUT (Pin 76) Transmitter Quadrature-Phase Output. TXQOUT is the quadrature-phase output transmission signal that has been differentially encoded and PN spread. TXQOUT changes on the rising edge of TXIFCLK following the falling edge of TXCHPPLS. In both modes the data must be valid on the second rising edge of TXIFCLK after the rising edge of TXBITPLS. TXCHPPLS (Pin 62) TXIFOUT7-0 (Pins 66-73) Transmitter Chip Pulse. TXCHPPLS is an output signal used to support transmission timing for the device. TXCHPPLS pulses high for one TXIFCLK cycle at the PN chip rate defined by the user. The chip rate is set either by programming a value in bits 5-0 of address 41H or through use of the external TXMCHP signal. TXIFOUT7-0 is the modulated transmit output signal from the on-chip BPSK/QPSK modulator. The signal is composed of the sum of the modulated TXIOUT and TXQOUT signals, modulated by the NCO cosine and sine outputs, respectively. Since the modulator is driven by the STEL-2000AÕs NCO, TXIFOUT 7-0 changes on the rising edges of RXIFCLK, and operation of the BPSK/QPSK modulator requires that R X I F C L K and TXIFCLK be identical and their frequencies must not exceed 20 MHz TXIFOUT 7-0 may be in either two's complement or offset binary format according to the setting of bit 1 of address 40H. TXTRKPLS (Pin 61) Transmitter Data Track Pulse. TXTRKPLS is an output signal that allows monitoring of data symbol transmissions. A rising edge of output signal TXTRKPLS occurs one chip period before the end of the current data symbol transmission. TXTRKPLS then falls low immediately following the rising edge of TXIFCLK. TXACQPLS (Pin 60) Transmitter Acquisition Pulse. TXACQPLS is an output signal generated at the final chip of the Acquisition/Preamble symbol. The Acquisition/Preamble symbol is generated automatically by the STEL-2000A upon user command (either via bit 1 of address 37H or MTXEN input) and immediately precedes transmission of user data. TXACQPLS is then provided to the user to indicate when the final chip of the Acquisition/Preamble symbol is being transmitted. Transmitter Active. A high level on TXACTIVE indicates that the transmitter is sending data symbols. This signal will be set high at the end of the Acquisition/Preamble symbol, indicating the start of the first chip of the first data symbol at the TXIOUT and TXQOUT pins. It will be set low at the end of the last chip period of the last data symbol of the burst at the TXIOUT and TXQOUT pins. TXBITPLS (Pin 63) RXOUT (Pin 57) TXACTIVE (Pin 78) Transmitter Bit Pulse. TXBITPLS is an output signal used to support transmission timing of user data for either BPSK or QPSK modes, as programmed by bit 0 of 40H. Receiver Output. RXOUT is the output data of the receiver following downconversion, despreading and demodulation. In BPSK mode, one data bit is provided per symbol; in QPSK mode, two data bits are provided per symbol with a half-symbol separation between the bits. Note that the data will be invalid during the first symbol of each burst; i.e., in BPSK mode the first bit will be invalid and in QPSK mode the first two bits will be invalid. In BPSK mode, user-provided data is requested by the STEL-2000A by a rising edge of TXBITPLS once per symbol. TXBITPLS requests the data one chip period before the TXIN data is latched into the device, and TXBITPLS falls low immediately following the rising edge of TXIFCLK, where TXIFCLK latches the TXIN value. RXIOUT (Pin 56) Receiver I Channel Output. RXIOUT is the I channel output data before dibit-to-serial conversion. RXIOUT can be used in conjunction with the RXQOUT signal in applications where the QPSK out- In QPSK mode, user-provided data is requested by the STEL-2000A by a rising edge of output signal TXBITPLS which occurs twice per symbol, first one –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 18 put data is required as parallel bit pairs. Note that the first bit of RXIOUT in each burst will be invalid. selected according to the value in bits 3-0 of address 38H and the assignments shown in Table 3. When one of these 4-bit values is written into address 38 H, the corresponding function becomes available at the RXTEST 7-0 outputs. Note that the matched filter power output values, available when the value in bits 3-0 of address 38 H is set to either 1H or 8H , is an unsigned binary number, ranging from 0 to 255 (0H to FFH). All other signal values available at this port are in TwoÕs Complement, ranging from Ð128 to +127 (80H to 7FH). The RXTEST 7-0 bus is a 3-state bus and is controlled by the OEN input. Note that the validity of the RXTEST7-0 bus signals at RXIFCLK frequencies over 20 MHz is dependent on the function selected. Functions that change more rapidly than once per symbol may be indeterminate at clock frequencies higher than 20 MHz. Note that the reset sequence that occurs whenever the receiver is disabled will also reset the contents of address 38H to 0. If the RXTEST 7-0 bus is being used for reading any function except the Matched Filter I and Q inputs the value required must be re-written after disabling the receiver. RXQOUT (Pin 55) Receiver Q Channel Output. RXQOUT is the Q channel output data before dibit-to-serial conversion. RXQOUT can be used in conjunction with the RXIOUT signal in applications where the QPSK data is required as parallel bit pairs. Note that the first bit of RXQOUT in each burst will be invalid. ––––––––– RXDRDY (Pin 54) ÐÐÐÐÐÐÐÐÐ Receiver Data Ready Bar. RXDRDY is provided as ÐÐÐÐÐÐÐÐÐ a receiver timing signal. RXDRDY is normally set high and pulses low during the baseband sampling clock cycle when a new RXOUT signal is generated. RXSPLPLS (Pin 53) Receiver Sample Pulse. RXSPLPLS is an output timing signal that provides internal timing information to the user. RXSPLPLS is the internally generated baseband sampling clock, referenced either externally or internally according to the setting of bit 0 of address 01H. All receiver functions, excluding those in the Downconverter, trigger internally on the rising edge of RXSPLPLS. TXTEST (Pin 59) Transmitter test output. TXTEST provides access to 3 test points within the transmitter as shown in Figure 6. The pin output is selected according to the state of the two least significant bits of the address line, ADDR 1-0 and the assignments shown in Table 4. Note that this method of accessing the transmitter test points is completely different than the method by which the receiver test points are accessed. The state of the other address lines does not affect this function, and this function is always enabled. RXSYMPLS (Pin 52) Receiver Symbol Pulse. RXSYMPLS is an output signal that provides the user internal timing information relative to the detection/correlation of symbols. Symbol information from the PN Matched Filter, DPSK Demodulator, and Output Processor is transferred on the rising edge of RXSPLPLS preceding the falling edge of RXSYMPLS. RXACTIVE (Pin 83) Receiver Active. A high level on RXACTIVE indicates that the receiver has detected an Acquisition/Preamble symbol and is currently receiving data symbols. RXACTIVE will be set high one bit period ÐÐÐÐÐÐÐÐÐ before the first rising edge of RXDRDY , indicating that the first data bit is about to appear at the RXOUT, RXIOUT, and RXQOUT pins. RXACTIVE will be set low immediately following the last rising ÐÐÐÐÐÐÐÐÐ edge of RXDRDY , indicating that the last data bit of the burst has been output at the RXOUT, RXIOUT, and RXQOUT pins. ADDR1-0 TXTEST Description 0H ISYM Unspread I symbol 1H QSYM Unspread Q symbol 2H SCODE Spreading code Table 4. Transmitter test functions Note that the validity of the TXTEST bus signals at TXIFCLK frequencies over 20 MHz is dependent on the function selected. Functions that change more rapidly than once per symbol may be indeterminate at clock frequencies higher than 20 MHz. RXTEST7-0 (Pins 41-48) These pins provide access to 16 test points within the receiver as shown in Figure 6. The pin outputs are –––––––––––––––––––––––––––––––––––––––––––––––––––– 19 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Bits 3-0 RXTEST7-0 output of 38H Bit 7 Bit 6 Bit 5 0H Bit 4 Bit 3 Bit 2 MFQIN2-0 Matched Filter Q input Bit 1 MFIIN2-0 Matched Filter I input 1H ¥Pk-POWER9-2. MF peak magnitude output, unsigned binary (changes once per symbol) 2H COS7-0. Cosine output of NCO (changes every cycle of RXIFCLK) 3H SIN7-0. Sine output of NCO (changes every cycle of RXIFCLK) 4H DCIOUT16-9. Downconverter I channel output (changes at RXIFCLK rate) 5H DCQOUT16-9. Downconverter Q output (changes at RXIFCLK rate) 6H ISUM9-2. Matched Filter I output (changes twice per chip) 7H QSUM9-2. Matched Filter Q output (changes twice per chip) 8H POWER9-2. MF magnitude output, unsigned binary (changes twice per chip) 9H ISUM7-0. MF viewport I output (changes twice per chip) AH Bit 0 QSUM7-0. MF viewport Q output (changes twice per chip) BH Pk-ISUM 7-0. MF peak I channel output (changes once per symbol) CH Pk-QSUM7-0. MF peak Q channel output (changes once per symbol) DH DOT 16-9. Dot product (changes once per symbol) EH CROSS16-9. Cross product (changes once per symbol) FH TXFBK 7-0. Loopback test output Table 3. Receiver test functions SERIAL TO DIBIT CONVERTER TX OVERLAY GENERATOR TXMCHP TX PN CODE GENERATORS RXTEST(1) RXTEST(8) RXTEST(9) RXTEST(F) COS RXTEST(4) RXTEST(2) RXTEST(0) (BITS 2-0) POWER DETECTOR VIEWPORT PEAK HOLD TXIOUT TXQOUT SIN NCO VIEWPORT DIFFERENTIAL DEMODULATOR AND CARRIER DISCRIMINATOR MATCHED FILTER Figure 6: Transmitter and Receiver Test Points I &D FILTERS RXIIN 7-0 COMPLEX MULTIPLIER RXTEST(5) RXTEST(0) (BITS 5-3) RXTEST(7) RXTEST(6) RXTEST(A) RXTEST(C) RXTEST(E) RXIOUT RXTEST(B) DIBIT TO SERIAL CONVERTER RXTEST(D) RX OVERLAY GENERATOR RXQOUT TXIFOUT 7-0 TXTEST(1) TXTEST(2) TX CLOCK GENERATOR RXOUT QPSK MODULATOR DIFFERENTIAL ENCODER RXTEST(3) TXIFCLK A A A A A A AA A AA A A A A A A AAA A A TXTEST(0) TXBITPLS TXTRKPLS TXIN RXQIN 7-0 A –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 20 CONTROL REGISTERS Downconverter Registers Address 00H: Bit 0 -- Frequency Control Word Load Bit 2 -- NCO Accumulator Carry In This bit is used to load a frequency control value into the NCO, thereby changing its output frequency. The signal is internally synchronized to RXIFCLK to avoid intrinsic race or hazard timing conditions. This bit is primarily used as an internal test function and should be set low for normal operation. When this bit is set high, 1 LSB is added to the NCO accumulator each clock cycle. When it is set low, the NCO accumulator is not affected. The loading of the NCO may be performed by various means. Setting this bit provides a synchronized internal means to control update of the NCO. Alternatively, the MFLD pin or the STEL2000AÕs programmable loop filter timing circuitry may be used. Bit 3 -- Two's Complement Input The RXIIN 7-0 and RXQIN7-0 input signals can be in either two's complement or offset binary formats. Since all internal processing in the device operates with two's complement format signals, it is necessary to convert the RXIIN7-0 and RXQIN7-0 inputs in offset binary format to twoÕs complement format by inverting the MSBs. The M F L D input and bit 0 of address 00H are logically ORed together so that, when either one is held low, a rising edge on the other triggers the frequency load function manually. The rising edge of this bit is synchronized internally so that, on the following sixth rising edge of RXIFCLK, the frequency control word is completely registered into the NCO accumulator. The frequency load command must not be repeated until after a delay of six RXIFCLK cycles. When this bit is set high, the device expects two's complement format inputs on RXIIN7-0 and RXQIN70 . When it is set low, the device expects offset binary format on RXIIN7-0 and RXQIN7-0. In two's complement format, the 8-bit input values range from Ð128 to +127 (80H to 7FH); in offset binary format, the values range from 0 to +255 (00 H to FFH). Address 01H: Bit 0 -- Manual Sample Clock Enable Bits 7-4 -- Integrate and Dump Filter Viewport Control This bit selects the source of the internal baseband sampling clock, which should be at twice the nominal PN chip rate. The clock reference may be either supplied externally by RXMSMPL or generated internally from RXIFCLK. The STEL-2000A incorporates viewport (data selector) circuitry to select any three consecutive bits from the 14-bit output of the Integrate and Dump (I & D) Filters in the Downconverter block as the 3-bit inputs to the dual-channel PN Matched Filter. The signal levels of the Integrate and Dump Filter I and Q outputs reflect the input signal levels and the number of samples integrated before the filter contents are Òdumped,Ó where the number of samples is determined by the baseband sampling rate (nominally, twice the PN chip rate) and the I.F. sampling rate (RXIFCLK). Setting the viewport thus effectively normalizes the I & D Filter outputs before further processing. The unsigned value, n, of bits 7-4 of address 01 H determines the 3-bit inputs to the PN Matched Filter as the 14-bit IÊ&ÊD Filter outputs divided by 2n . Equivalently, bits 7-4 control the viewport of the Integrate and Dump Filter outputs as shown in Table 5. Note that viewport control affects both I and Q channels of the Integrate and Dump Filters. When this bit is set high, the baseband sampling rate of the receiver is controlled by the external RXMSMPL signal. When it is set low, the sampling clock is generated internally (at a rate determined by the Sample Rate Control counter and set by bits 5-0 of address 02H) and the RXMSMPL input is ignored. Bit 1 -- Invert Loop Filter Value This bit allows the sign of the output signal from the loop filter to be inverted, thereby negating the value of the signal. The capability to invert the loop filter value permits the carrier frequency error component generated in the demodulator to be either added to or subtracted from the Frequency Control Word of the NCO. The correct setting will depend on several factors, including whether high-side or low-side downconversion is used. Bits 7-4 When this bit is set low, the loop filter output is negated before being summed with the Frequency Control Word of the NCO and is thus subtracted from the FCW; when this bit is set high, the loop filter output is not negated and is added to the FCW. I & D bits output to Matched Filter 0H 2-0 1H 3-1 –––––––––––––––––––––––––––––––––––––––––––––––––––– 21 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A 2H 4-2 to the following formula: 3H 5-3 ..... .... fRXIFCLK x FCW fNCO = ÐÐÐÐÐÐÐÐÐÐÐÐÐÐ 2 32 ..... .... AH 12-10 BH 13-11 In order to avoid in-band aliasing, fNCO must not exceed 50% of fRXIFCLK; normally, the FCW should be set so that fNCO does not exceed ~35% of fRXIFCLK. While this limitation may seem to restrict use of the NCO, higher I.F. transmit or receive frequencies can generally be achieved by using aliases resulting from digital sampling. The signal bandwidth with respect to fRXIFCLK, the modulation type, and the use of Direct I.F. or Quadrature Sampling Mode also restrict the choice of NCO frequency, as discussed in Appendix I. Table 5. Integrate & Dump Filter Viewport Control Saturation protection is implemented for those cases when the Integrate and Dump Filter output signal level overflows the scaled range selected for the PN Matched Filter. When the scaled value range is exceeded, the saturation protection limits the output word to the maximum or minimum value of the range according to whether the positive or negative boundary was exceeded. Matched Filter Registers Despreading of the received signal is accomplished in the STEL-2000A with a dual (I and Q channel) PN Matched Filter. Furthermore, the STEL-2000A is designed for burst signal operation, where each data burst begins with an Acquisition/Preamble symbol and is then followed by the actual information data symbols. Two separate and independent PN codes can be employed, one for the Acquisition/Preamble symbol, the other for the information symbols. Accordingly, the PN Matched Filter is supported by two PN code registers to independently allow the programming of two distinct codes up to 64 chips in length. The PN codes are represented as a sequence of ternary-valued tap coefficients, each requiring 2 bits of storage according to the mapping shown in Table 7. Address 02H: Bits 5-0 -- Receiver Baseband Sampling (Dump) Rate Control The baseband sampling rate should be set to twice the nominal PN chip rate of the received signal and must be less than or equal to half the rate of RXIFCLK. When bit 0 of address 01H is set low, the baseband sampling clock for the Integrate and Dump Filter and all subsequent receiver circuitry is referenced to RXIFCLK and generated internally. The receiver baseband sampling rate is then set to the frequency of RXIFCLK/(n+1), where n is the value stored in bits 50 and must range from 1 to 63. This feature is useful in cases where a specific sample rate is required that is an integer sub-multiple of fRXIFCLK. In cases where a sample rate is required that is not an integer submultiple of fRXIFCLK, an external baseband sampling rate can be provided by the RXMSMPL input. Tap Bits 1,0 X 0 Addresses 03 H through 06H: NCO Frequency Control Word The STEL-2000AÕs internal NCO is driven by a frequency control word that is the sum of the frequency discriminator error value (generated in the demodulator) and the 32-bit frequency control word (FCW) stored in this location. The four 8-bit registers at addresses 03H to 06H are used to store the 32-bit frequency control word as shown in Table 6. The LSB of each byte is stored in bit 0 of each register. ADDR6 ADDR5 ADDR4 ADDR3 Bits 31-24 Bits 23-16 Bits 15-8 Bits 7-0 Tap Coeff. 0 0 1 +1 1 1 Ð1 Table 7. PN Matched Filter Tap Values As a convention, Tap 0 is the first tap as the received signal enters the PN Matched Filter, and Tap 63 is the last. All active taps of the PN Matched Filter, from Tap 0 up to Tap (N-1), where N is the length of the PN code, should be programmed with tap coefficient values of +1 or -1 according to the PN code sequence. Setting the end coefficients of the PN Matched Filter registers to zero values permits the effective length of the filter to be made shorter than 64 taps. Table 6. NCO FCW Storage Addresses 07 H through 16H: Matched Filter Acquisition/Preamble Symbol Coefficients The NCO frequency is then set by the FCW according Addresses 07H to 16H contain the 64 2-bit Acquisi- –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 22 tion/Preamble PN code coefficient values. The 128 bits of information are stored in 16 8-bit registers at addresses 07H to 16H as shown in Table 8. É Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 63 Coeff. 62 Coeff. 61 Coeff. 60 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 59 Coeff. 58 Coeff. 57 Coeff. 56 É É É É É É É É Bits 7,6 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 7 Coeff. 6 Coeff. 5 Coeff. 4 Bits 7,6 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 3 Coeff. 2 Coeff. 1 Coeff. 0 É Bits 7,6 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 7 Coeff. 6 Coeff. 5 Coeff. 4 Bits 7,6 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 3 Coeff. 2 Coeff. 1 Coeff. 0 Address 17H Address 15H Bits 7,6 É Address 18H Address 16H Bits 7,6 É Table 9. Data Symbol Coefficient Storage Address 27H: Bit 0 -- Front End Processor Disable The Front End Processor (FEP) averages the two baseband samples per chip by adding consecutive pairs of samples. The function may be disabled for test purposes by using this bit: when set low, the FEP is enabled and in its normal mode of operation; when set high, the FEP is disabled. Address 08H Address 07H Power Estimator Registers Address 28H: Bits 1-0 -- Matched Filter Viewport Control The STEL-2000A incorporates viewport (data selector) circuitry to select any eight consecutive bits from the 10-bit outputs of the PN Matched Filter as the 8-bit inputs to the Power Estimator and DPSK Demodulator blocks. The Symbol Tracking Processor, however, operates on the full 10-bit PN Matched Filter outputs before the viewport is applied. The signal levels of the PN Matched Filter output reflect the number of chips per symbol and the signal-to-noise ratio of the signal. Setting the viewport thus effectively normalizes the PN Matched Filter outputs prior to further processing. The unsigned value, n, of bitsÊ1-0 of address 28H determines the 8-bit input to the Power Estimator and DPSK Demodulator blocks as the 10-bit PN Matched Filter output divided by 2n . Equivalently, bits 1-0 control the viewport of the PN Matched Filter output as shown in Table 10. Note that viewport control affects both I and Q channels of the PN Matched Filter output. Table 8. Acquisition/Preamble Coefficient Storage Addresses 17 H through 26H: Matched Filter Data Symbol Coefficients Addresses 17H to 26H contain the 64 2-bit Data Symbol PN code coefficient values. The 128 bits of information are stored in 16 8-bit registers at addresses 17H to 26H as shown in Table 9. The contents of addresses 17H to 26H are independent of and not affected by the contents of addresses 07H to 16H. Address 26H Bits 7,6 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 63 Coeff. 62 Coeff. 61 Coeff. 60 Bits 1-0 Address 25H Bits 7,6 Bits 5,4 Bits 3,2 Bits 1,0 Coeff. 59 Coeff. 58 Coeff. 57 Coeff. 56 É É É ISUM, QSUM 0 0 Bits 7-0 0 1 Bits 8-1 1 X Bits 9-2 Table 10. Matched Filter Viewport Control Saturation protection is implemented for those cases when the PN Matched Filter output signal level over- É –––––––––––––––––––––––––––––––––––––––––––––––––––– 23 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Filter correlated output values are then used to estimate the signal power according to the following approximation: flows the scaled range selected for the Power Estimator and DPSK Demodulator. When the scaled value range is exceeded, the saturation protection limits the output word to the maximum or minimum value of the range according to whether the positive or negative boundary was exceeded. Max{Abs(I),Abs(Q)}+1/2 Min{Abs(I), Abs(Q)}. The magnitude of the estimated power thus depends on several variables, including the setting of the Integrate and Dump Filter viewport, the PN code length and autocorrelation properties, and the magnitudes of the incoming RXIIN7-0 and RXQIN7-0 signals. The actual threshold values that should be programmed will therefore vary from application to application. Acquisition and Tracking Processor Registers The Acquisition and Tracking Processor Registers allow the user to configure how the PN Matched Filter outputs for the Acquisition/Preamble symbol and the data symbols that follow thereafter are treated in the Symbol Tracking Processor. Since operation of the STEL-2000A receiver presumes symbolsynchronous PN modulation, processing of the PN Matched Filter outputs can be used for symbol synchronization prior to DPSK demodulation. The Acquisition/Preamble symbol and the data symbols may have different PN spreading codes, however, and so the PN Matched Filter outputs may exhibit different signal levels due to the different code lengths and autocorrelation properties. The control registers in this block allow such differences to be treated, as well as permitting specification of the number of receive data symbols per burst and other parameters associated with burst data communications. Addresses 29 H and 2AH: Acquisition/Preamble Threshold Addresses 29H and 2AH contain the unsigned Acquisition/Preamble Threshold value, as shown in Table 11. This value is used for comparison with the estimated signal power from the PN Matched Filter to determine whether a successful correlation has been detected in acquisition mode. The Acquisition/Preamble Threshold value must be set by the user to satisfactorily detect the correlation peak in noise obtained when the received PN-spread Acquisition/Preamble is correlated against a local version of the Acquisition/Preamble PN code by the PN Matched Filter. Once the power estimation value exceeds the threshold value, a successful correlation is assumed to have been detected. Note that the Symbol Tracking Processor does not insert missed detect pulses when the device is in acquisition mode. The I and Q channel outputs of the PN Matched Filter are processed to estimate the correlation signal power at each baseband sampling instant. This estimated signal power is compared with the contents of the Acquisition/Preamble and Data Symbol Threshold registers, as appropriate, to determine whether ÒsuccessfulÓ correlation has been detected. Successful detection in acquisition mode immediately switches the receiver to despread and track the expected subsequent data symbols, while successful detection thereafter yields symbol synchronization. The threshold register values must be set by the user to satisfactorily detect the correlation peak in noise obtained when the received PN-spread signal is correlated against a local version of the PN code by the PN Matched Filter. Once the power estimation value exceeds the threshold register value, a successful correlation is assumed to have been detected. Further operations in the Symbol Tracking Processor then handle the possibility of multiple detects per symbol, missed detects, etc. ADDR 2A H ADDR 29H Bits 1-0 Bits 7-0 Acq. Thresh. Bits 9-8 Acq. Thresh. Bits 7-0 Table 11. Acquisition/Preamble Threshold Storage Addresses 2BH and 2CH: Data Symbol Threshold Addresses 2BH and 2C H contain the Data Symbol Threshold value, as shown in Table 12. This value is used for comparison with the estimated signal power from the PN Matched Filter to determine whether a successful correlation has been detected for each data symbol. The Data Symbol Threshold value must be set by the user to satisfactorily detect the correlation peak in noise obtained when the received PN-spread data symbol is correlated against a local version of the data symbol PN code by the PN Matched Filter. Once the power estimation value exceeds the threshold value, a successful correlation is assumed to have been detected. If bit 2 of address 30H is set low, then the Symbol Acquisition Processor will insert a detect pulse at the appropriate time if a successful correlation is not detected as expected a priori. The choice of the threshold values will be determined by several factors. Arithmetically, the digital baseband samples of the received signal are multiplied by the PN Matched Filter tap coefficients each baseband sample clock cycle and the results are summed to provide a correlation value. The I and Q PN Matched –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 24 ADDR 2CH ADDR 2BH Bits 1-0 Bits 7-0 Data Thresh. Bits 9-8 Data Thresh. Bits 7-0 and tallied for the current burst. When the accumulated number of missed detects is greater than the value stored in address 2F H, the device will terminate reception of the current burst and return to acquisition mode to await the next burst. Table 12. Data Symbol Threshold Storage The unsigned value in address 2F H must range from 1 to 255 (01H to FFH), where this value is the maximum number of missed detects per burst allowed before the burst terminates. This function can be disabled by setting bit 5 of address 30H high. Address 2DH: Bits 5-0 -- Rx Chips per Data Symbol The number of PN chips per data symbol in the receiver is controlled by address 2DH. The unsigned value must range from 1 to 63 (01H to 3F H), where the number of chips per data symbol will be this value plus 1. The a priori number of PN chips per data symbol, where this value must be equal to the number of non-zero coefficients stored in the Data Symbol Coefficient Registers (addresses 17H to 26 H) for the PN Matched Filter, is used to help control symbol timing in the receiver. Since acquisition is purely based on correlation of a single received Acquisition/Preamble symbol, the corresponding number of chips per Acquisition/Preamble symbol is not required and no similar register is provided for such use. Address 30H: Bit 0 -- Manual Detect Enable While the receiver is in acquisition mode, valid bursts may be ignored by setting this bit high. When it is set low (normal operation), the detection of a burst's Acquisition/Preamble symbol is enabled. Setting this bit high allows the user to force the device to ignore Acquisition/Preamble symbols that would normally be successfully acquired. This feature could be used, for example, in a system employing multiple receivers with identical PN codes in a Time Division Multiple Access scheme where time-synchronized device management could be supported through dynamic setting of this bit. Address 2EH: Receiver Data Symbols per Burst (bits 7-0) The data stored as two bytes in addresses 2EH (LS Byte) and 3A H (MS Byte) defines the number of data symbols per burst. This unsigned value must range from 3 to 65,535 (0003 H to FFFFH), and the number of data symbols per burst will be this value minus 2, giving a range of 1 to 65,533. Note that the range is slightly different from that supported by the STEL2000AÕs transmitter. Once the number of received Bit 1 -- Manual Punctual This bit enables the user to completely disable the internal tracking circuitry and force symbol information to be transferred to the demodulator punctually at the symbol rate determined by the number of chips per data symbol information programmed into address 2D H. This function overrides the symbol tracking algorithm, although the absence of a successful correlation will continue to be tallied as a missed detect and compared against the value stored in address 2F H to monitor signal quality unless disabled by bit 5 of address 30 H. When bitÊ1 is set low, the STEL-2000A will operate in its normal mode with symbol timing derived from the symbol tracking processor; when set high, symbol timing is derived from the a priori number of chips per data symbol stored in bits 5-0 of address 2D H. data symbols processed exceeds this number, the burst is assumed to have ended and the receiver immediately returns to acquisition mode, ready for the next burst. Address 2FH: Missed Detects per Burst Threshold To monitor the reception quality of the received burst data symbols, the STEL-2000A incorporates a feature within its tracking algorithm that tallies the number of received data symbols whose PN Matched Filter correlation output did not exceed the Data Symbol Threshold value. Bit 2 -- Force Continuous Acquisition This bit enables the user to force the receiver to remain in acquisition mode even after successful detection of the Acquisition/Preamble symbol. When so commanded, the receiver will continuously process only Acquisition/Preamble symbols and will not switch from acquisition mode. This function may be used under manual control to receive a series of repeated Acquisition/Preamble symbols in order to increase the confidence level of burst detection before beginning demodulation of the data symbol information. Whenever a "missed detect" occurs, the tracking algorithm will generate and insert a detect signal at the sample clock cycle corresponding to the expected correlation peak in order to maintain a continuous train of data symbols and symbol clocks. Simultaneously, a "missed detect" pulse will be generated internally –––––––––––––––––––––––––––––––––––––––––––––––––––– 25 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A not exceed the specified Data Symbol Threshold value. When the accumulated number of missed detects equals the Missed Detects per Burst Threshold value stored in address 2F H, the device will terminate the reception of the current burst with the next missed detect and return to acquisition mode to await the next burst. When this bit is set high, the device will be locked in acquisition mode and the Symbol Tracking Processor will not insert missed detect pulses; when set low, normal operation will be enabled whereby data symbols are automatically processed immediately following detection of an Acquisition/Preamble symbol. Bit 3 -- Bypass Max. Power Selector The STEL-2000AÕs receiver acquisition and tracking circuitry includes a function that continuously selects the highest estimated power level out of the three most recent consecutive estimated power levels from the PN Matched Filter. As the contents of the sliding 3-sample window change each cycle of the baseband sampling clock, a new determination of the highest power level is made from the current set of the three most recent power level values. The correlated I and Q channel values within the 3-sample window corresponding in time to the highest observed power level are then available to be processed in the demodulator. When bit 5 is set low, the "missed detect" function operates normally; when set high, this function is disabled, allowing the device to be operated until the end of the specified data burst even when the number of Òmissed detectsÓ exceeds the Missed Detects per Burst Threshold. Bit 6 -- Receiver Symbols Per Burst Off The data stored in addresses 2E H and 3AH defines the number of data symbols per burst that will be processed by the receiver. This unsigned value must range from 3 to 65,535 (0003H to FFFF H), and the number of data symbols per burst will be this value minus 2. Once the number of data symbols processed by the receiver exceeds this number, the burst is assumed to have ended and the receiver will immediately return to acquisition mode. This function assures that, within any 3-sample period, the I and Q channel values corresponding to the highest estimated power level will be selected over the two other pairs of correlated values even if the estimated power levels of the other pairs exceed the programmed threshold. The Maximum Power Selector is used in normal operation of the STEL2000A so that the tracking algorithm discriminates by estimated power levels rather than exact timing intervals, thereby allowing the receiver to adjust to dynamic changes of the symbol phase. In cases where specific correlation values are desired regardless of their associated power level, bit 3 of address 30H enables the 3-sample power discriminator to be bypassed, thereby making the outputs of the PN Matched Filter available directly to the demodulator. When bit 6 is set high, the function is disabled, providing an option to track data symbols under external control for bursts of more than 65,533 data symbols or indefinitely for continuous transmission; when set low, the function will operate normally as defined by the value stored in addresses 2E H and 3AH. When this bit is set high, the Maximum Power Selector is bypassed; when it is set low, the Selector is enabled, where this is the normal operating mode. Address 31H: Bit 0 -- Manual Detect Pulse This bit provides the user a means to externally generate symbol timing, bypassing and overriding the internal symbol power estimation and tracking circuitry. This function may be useful in applications where the dynamic characteristics of the transmission environment require unusual adjustments to the symbol timing. Bit 4 -- Half Symbol Pulse Off The STEL-2000A generates two bit clock pulses per symbol when operating in QPSK mode, one at the mid-point of each symbol and one at the end of each symbol. These clocks are used by the Output Processor to manage data flow. When this bit is set high, the mid-point pulse is suppressed; when it is set low, the device operates in its normal mode. This function is primarily used for test purposes and should not normally be used. When bit 0 of address 30H is set high (Manual Detect Enable) and when RXMDET is low, a rising edge on this bit will generate a detect pulse. The function can also be performed by means of the RXMDET input signal. Bit 0 of address 31H and the RXMDET input are logically ORed together so that, when either one is held low, a rising edge on the other triggers the manual detect function. The rising edge of this bit is synchronized internally so that on the second rising edge Bit 5 -- Missed Detects Per Burst Off To monitor the quality of the received burst data symbols, the Symbol Tracking Processor keeps track of the cumulative number of received data symbols per burst whose estimated correlation power level did –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 26 of the baseband sampling clock that follows, the rising edge of bit 0 will transfer the I and Q channel correlated output values of the PN Matched Filter to the DPSK Demodulator. The setting of this bit determines whether the Loop Filter's K2 accumulator is reset or not when the STEL2000A receiver function is turned off when the input signal MRXEN is set low. Address 32H: Bit 0 -- Receiver Manual Abort When bit 3 is set low, the Loop Filter's K2 accumulator will be reset to zero whenever MRXEN is set low to disable the receiver function. When bit 3 is set high, this function is disabled and the contents of the accumulator are not affected when MRXEN transitions from high to low. The optimum setting of this bit will depend on the stability of the oscillators used for carrier generation and frequency translation in the system and the length of the period between bursts. If the oscillators are stable and the period between bursts is not very long, the optimum setting of this bit will be low so that at the start of each burst the tracking loop will resume from its state at the end of the previous burst. If the oscillators are not stable or if the period between bursts is long with respect to the oscillatorsÕ stability, then the optimum setting may be high so that the tracking loop will restart from its initial state at the start of each burst. This bit enables the user to manually force the STEL2000A to cease reception of the present burst of data symbols and prepare for acquisition of a new burst. This function can be used to reset the receiver and prepare to receive a priority transmission signal under precise timing control, giving the user the ability to control the current state of the receiver as needed. When RXMABRT is set low, a rising edge on bit 0 of address 32H will execute the abort function. The function can also be performed by means of the RXMABRT input. The RXMABRT input and bit 0 of address 32H are logically ORed together so that, when either one is held low, a rising edge on the other triggers the abort function. The second rising edge of the internal baseband sampling clock that follows a rising edge of this bit will execute the abort and also clear the symbols-per-burst, samples-per-symbol, and missed-detects-per-burst counters. The counters will be reactivated on the detection of the next Acquisition/Preamble symbol or by a manual detect signal. Bits 7-4 -- AFC Viewport Control The STEL-2000A incorporates viewport (data selector) circuitry to select any eight consecutive bits from the 17-bit output of the Frequency Discriminator as the 8-bit input to the Loop Filter block to implement the STEL-2000AÕs AFC function. The unsigned value, n, of bitsÊ7-4 of address 33H determines the 8-bit input to the Loop Filter as the 17-bit Frequency Discriminator output divided by 2n. Equivalently, bits 7-4 control Demodulator Registers Address 33H: Bits 1-0 -- Signal Rotation Control These bits control the function of the Signal Rotation Block used in demodulation of the differentially encoded BPSK, QPSK, or π/4 QPSK signals. The previous symbol will be rotated in phase with respect to the current symbol as shown in Table 13, where Iout and Q out are the I and Q channel outputs of the Signal Rotation Block and Iin and Q in are the inputs. The normal settings are 0 X (no rotation) for BPSK and π/4 QPSK signals and 1 1 (Ð45° rotation) for conventional QPSK signals. Bits 1,0 Iout Qout Resulting Rotation 0X Iin Qin No rotation 10 IinÐ Qin Qin+ Iin +45° rotation 11 Iin+ Qin QinÐ Iin Ð45° rotation the viewport of the Frequency Discriminator output as shown in Table 14. Table 13. Signal Rotation Control Bit 2 -- Not used Bits 7-4 Discrim. bits output to Loop Filter 0H 7-0 1H 8-1 2H 9-2 3H 10-3 É É É É 8H 15-8 9H 16-9 AH - F H not used Bit 2 of address 33 H is not used and should always be set low. Table 14. AFC Viewport Control Bit 3 -- Loop Clear Disable Saturation protection is implemented for those cases when the Frequency Discriminator output signal level –––––––––––––––––––––––––––––––––––––––––––––––––––– 27 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A overflows the scaled range selected for the Loop Filter. When the scaled value range is exceeded, the saturation protection limits the output word to the maximum or minimum value of the range according to whether the positive or negative boundary was exceeded. Bits 4-0 Gain in K1 Path 00H 20 01H 21 ..... .... Address 34H: Bits 4-0 -- K2 Gain Value ..... .... 14H 2 20 15H 2 21 Bits 4-0 control the gain factor K2 within the Loop Filter. The gain factor multiplies the signal before the K2 accumulator by a value of 2n, where n is the 5-bit K2 Gain Value. The value must range from 0 to 21 (15H) as shown in Table 15. Bits 4-0 Gain in K2 Path 00H 20 01H 21 ..... .... ..... .... 14H 2 20 15H 2 21 Table 16. K1 Gain Values Bit 5 -- K1 On This bit enables or disables the K1 path of the Loop Filter. Setting this bit low disables the K1 path; setting this bit high enables the path and turns on K1. Bit 6 -- Freeze Loop This bit enables the Loop Filter to be held constant during symbol cycles, thereby fixing the output frequency of the NCO at the value established by the Loop Filter when bit 6 was set high. This function can be useful in cases where a carrier offset has been tracked by the Loop Filter and additional Doppler offsets are to be ignored. Table 15. K2 Gain Values Bit 5 -- K2 On This bit enables or disables the K2 path of the Loop Filter. Setting this bit low resets the K2 accumulator and keeps it reset; setting this bit high enables the path and turns on K2. When this bit is set high, it freezes the output of the Loop Filter; when it is set low, the Loop Filter is enabled and processes the frequency error information in the usual way. Bit 6 -- Carry In One Half Output Processor Control Registers Address 36H: Bit 0 -- Reverse I and Q 1/2 When this bit is set high, the value of of an LSB is added to the accumulator of the K2 path of the Loop Filter each symbol period. This function can be useful in cases where the scale and gain functions that precede the accumulator produce quantized values with significant error. In such cases, the processing of two's complement numbers by the accumulator will compound the error over time. Since truncation of two's complement numbers leads to a negative bias of 1/2 of an LSB when the error is random, adding 1 /2 of an LSB per symbol can compensate by averaging the error to zero. In QPSK mode, the order in which the received I and Q bit information is output may be reversed by setting this bit high. This function has the effect of interchanging I and Q channels. Normally, when this bit is set low, the I-channel bit will precede the Q-channel bit in each symbol period. When bit 0 is set high, the Q-channel bit will precede the I-channel bit each symbol period. Bit 1 -- BPSK Enable This bit configures the Output Processor to output either one bit per symbol (BPSK mode) or two bits per symbol (QPSK mode). In addition, it enables the user to output the I-channel information only or the Qchannel information only, depending on the value of bit 0. Table 17 shows the configuration of the output processor for all combinations of the values of bits 0 and 1. When bit 6 of address 34H is set high, a value of 1/2 will be added to the accumulator input each symbol cycle; when it is low, a zero will be added. Address 35H: Bits 4-0 -- K1 Gain Value Bits 4-0 control the gain factor K1 within the Loop Filter. The gain factor multiplies the signal by a value of 2 n , where n is the 5-bit K1 Gain Value. The value must range from 0 to 21 (15H), as shown in Table 16. Bit 1 Bit 0 0 0 Output Processor Mode QPSK mode with I-Channel Bit Preceding Q-Channel Bit –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 28 0 1 QPSK mode with Q-Channel Bit Preceding I-Channel Bit 1 0 BPSK mode with I-Channel Information Output 1 1 BPSK mode with Q-Channel Information Output the NCO circuitry and, when MNCOEN is set low, bit 0 controls the activity of the NCO circuitry. When either bit 0 or MNCOEN (whichever is in control, as defined above) goes low, a reset sequence occurs on the following RXIFCLK cycle to virtually disable all of the NCO circuitry, although the user programmable control registers are not affected by the power down sequence. Upon reactivation (when either MNCOEN or bit 0 of address 37H return high), the NCO must be reloaded with frequency control information either by means of the MFLD input or by writing 01 H into address 00H. Table 17. Output Processor Modes Bit 1 also sets the Frequency Discriminator into either BPSK or QPSK mode. The STEL-2000A receiver uses dot and cross product results generated within the DPSK Demodulator to develop the error signal used to form a closed-loop AFC for carrier frequency acquisition and tracking. Bit 1 -- Transmitter Enable A rising edge on this bit causes the transmit sequence to begin so that the STEL-2000A first transmits a single Acquisition/Preamble symbol followed by data symbols. Bit 1 of address 37H should be set low after the last symbol has been transmitted to minimize power consumption of the transmitter circuit. Bit 1 of address 37 H operates independently of bits 2 and 0, where those bits have similar control over the receive and NCO circuitry, respectively. When bit 1 is set high, the discriminator circuitry is in BPSK mode and the Frequency Discriminator function is calculated as: . Cross x Dot 16-0 MSB When bit 1 is set low, the discriminator circuitry is in QPSK mode and the Frequency Discriminator function is calculated as: (Cross16-0 x DotMSB ) Ð (Dot16-0 x CrossMSB). When input signal MTXEN is set low, bit 1 of address 37H controls the activity of the transmit circuitry and, when MTXEN is set low, bit 1 controls this function. When either bit 1 or MTXEN (whichever is in control, as defined above) goes low, a reset sequence occurs on the following TXIFCLK cycle to virtually disable all of the transmitter data path, although the user programmable control registers are not affected by the power down sequence. Bit 2 -- Invert Output This bit inverts the output bits of both the I and Q Channels. The inversion will occur at the output pins RXOUT, RXIOUT, and RXQOUT. When this bit is set low, the outputs are not inverted; when it is set high, the outputs are inverted. Address 37H: Bit 0 -- NCO Enable Bit 2 -- Receiver Enable The function of this bit is to allow the power consumed by the operation of the NCO circuitry to be minimized when the STEL-2000A is not receiving. The NCO can also be disabled while the STEL-2000A is transmitting provided that the STEL-2000AÕs onchip BPSK/QPSK modulator is not being used. With the instantaneous acquisition properties of the PN Matched Filter, it is often desirable to shut down the receiver circuitry to reduce power consumption, resuming reception periodically until an Acquisition/Preamble symbol is acquired. Setting bit 0 low will do this by holding the NCO in a reset state. After this bit has been set high again to re-activate the NCO it will be necessary to reload the frequency control word into the NCO. Note that this bit operates independently of bits 1 (Transmitter Enable) and 2 (Receiver Enable), which have similar control over the transmit and receive circuits, respectively. The function of this bit is to allow power consumed by the operation of the receiver circuitry to be minimized when the device is not receiving. With the instantaneous acquisition properties of the PN Matched Filter, it is often desirable to shut down the receiver circuitry to reduce power consumption, resuming reception periodically until an Acquisition/Preamble symbol is acquired. Setting bit 2 low reduces the power consumption substantially. When bit 2 is set high, the receiver will automatically power up in acquisition mode regardless of its prior state when it was powered down. Bit 2 of address 37 H operates independently of bits 1 and 0 of address 37H,where these signals have similar control over the transmit and NCO circuitry, respectively. Bit 2 of address 37H performs the same function as MRXEN, and these two signals are logically ORed together to form the overall control function. When bit 2 of address 37H is set low, MRXEN controls the activity of the receiver circuitry and, when MRXEN is set low, bit 2 of address 37H controls the activity of the receiver circuitry. When either bit 2 or MRXEN Bit 0 of address 37H performs the same function as MNCOEN, and these two signals are logically ORed together to form the overall control function. When this bit is set low, MNCOEN controls the activity of –––––––––––––––––––––––––––––––––––––––––––––––––––– 29 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Power control is not provided for Tap 0, the first tap of the PN Matched Filter, since Tap 0 is always used no matter what the PN code length. Setting a bit high in bits 6-0 of address 39H turns off the power to the corresponding block of taps of the PN Matched Filter. The power should only be turned off to those blocks of taps for which all the tap coefficients in that block have been set to zero (whichever is in control, as defined above) goes low, a reset sequence begins on the following RXIFCLK cycle and continues through a total of six RXIFCLK cycles to virtually disable all of the receiver data paths, although the user programmable control registers are not affected by the power down sequence. Note that this reset sequence will also reset the contents of address 38H to 0. If the R X T E S T 7-0 bus is being used for reading any function except the Matched Filter I and Q inputs the value required must be re-written after re-enabling the receiver. Address 3AH: Receiver Data Symbols per Burst (bits 15-8) The data stored as two bytes in addresses 2EH (LS byte) and 3AH (MS byte) defines the number of data symbols per burst. This unsigned value must range from 3 to 65,535 (0003 H to FFFFH), and the number of data symbols per burst will be this value minus 2, giving a range of 1 to 65,533. Note that the range is slightly different from that in the transmitter. Once the number of received data symbols processed exceeds this number, the burst is assumed to have ended and the STEL-2000A immediately returns to acquisition mode to await the next burst. Address 38H: Bits 3-0 -- RXTEST7-0 Function Select The data stored in bits 3-0 of address 38H selects the signal available at the RXTEST7-0 bus (pins 41-48). These pins provide access to 16 test points within the receiver according to the data stored in bits 3-0 of address 38H and the assignments shown in Table 3. Note that the validity of the RXTEST7-0 bus signals at RXIFCLK frequencies over 20 MHz is dependent on the function selected. Functions that change more rapidly than once per symbol may be indeterminate at clock frequencies higher than 20 MHz. Note that the reset sequence that occurs whenever the receiver is disabled will also reset the contents of address 38H to 0. If the RXTEST7-0 bus is being used for reading any function except the Matched Filter I and Q inputs the value required must be re-written after disabling the receiver. Address 3BH: Bit 0 -- Matched Filter Loopback Enable The STEL-2000A incorporates a loopback capability that feeds the encoded and spread transmit signals TXIOUT and TXQOUT directly into the PN Matched Filter inputs. This test mode allows the baseband portion of the system to be tested independently of the BPSK/QPSK Modulator and Downconverter. Address 39H: Bits 6-0 -- Matched Filter Power Saver Setting bit 0 of address 3BH high enables this loopback path; setting it low puts the device into its normal operating mode. The data stored in bits 6-0 of address 39H allows the unused sections of the PN Matched Filter to be turned off when the PN Matched Filter is configured to be less than 64 taps long for data symbols. All taps are always fully powered when the device is in acquisition mode. Bit 1 -- I.F. Loopback Enable The STEL-2000A incorporates a loopback capability that feeds the encoded, spread and modulated transmit signal TXIFOUT 7-0 directly into the receiver RXIIN7-0 input. This test mode allows the entire digital portion of the system to be tested. Since only the I channel is provided as an input, I.F. loopback requires that the PN chip rate and RXIFCLK rate be consistent with Direct I.F. Sampling Mode. The PN Matched Filter is split into seven 9-tap sections, and the power to each section is controlled by the settings of bits 6-0 of address 39H , as shown in Table 18. Bit in Addr. 39H MF Taps Controlled 0 1-9 1 10-18 Setting bit 1 of address 3BH high enables this loopback path; setting it low puts the device into its normal operating mode. 2 19-27 Bits 3-2 -- Receiver Overlay Select 3 28-36 4 37-45 5 46-54 6 55-63 The STEL-2000A incorporates programmable overlay code generators in both the transmitter and receiver. When enabled, the selected receiver overlay code is subtracted from the data symbols, one overlay bit per symbol in both BPSK and QPSK modes. No synchronization beyond the burst acquisition synchronization that is intrinsic to operation of the STEL-2000A is Table 18. Matched filter tap power control –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 30 required since the overlay code generators in both the transmitter and the receiver are automatically reset at the start of each burst. The addition of the overlay code randomizes the transmitted data sequence to guarantee that the spectrum of the transmitted signal will be adequately whitened and will not contain a small number of spectral lines even when the data itself is not random. configured in QPSK mode; when it is set high, the transmitter is configured in BPSK mode. Bit 1 -- Offset Binary Output The TXIFOUT7-0 output signals can be in either two's complement or offset binary formats. Since all internal processing in the device uses two's complement format signals, the MSB of the twoÕs complement modulated transmitter output must be inverted if the output is to be in offset binary format. Three transmit and receive overlay codes can be selected, where they are each maximal length sequences with lengths of 63, 511 and 1023 symbols. The receiver overlay codes are enabled and selected by the settings of bits 3-2 of address 3BH, as shown in Table 19. Bits 3-2 in Addr. 3BH When this bit is set high, the TXIFOUT7-0 output will be in offset binary format and, when it is set low, the signal will be in twoÕs complement format. In two's complement format, the 8-bit output values range from Ð128 to +127 (80H to 7FH); in offset binary format, the values range from 0 to +255 (00H to FFH). Bit 2 -- Manual Chip Clock Enable Overlay Code length and polynomial 0 Overlay code disabled 1 63: 1 + x2 + x 3 + x 5 + x 6 2 511: 1 + x2 + x 3 + x 5 + x 9 3 1023: 1 + x2 + x 3 + x 5 + x 10 This bit enables the PN chip rate to be controlled by either the internal chip rate clock generator or by the external input signal TXMCHP. The TXMCHP input allows the user to manually insert a single PN chip clock pulse or continuous stream of pulses. This feature is useful in cases where a specific chip rate is required that cannot be derived by the internal clock generator which generates clocks of integer sub-multiples of the frequency of TXIFCLK. The signal is internally synchronized to TXIFCLK to avoid race or hazard timing conditions. Table 19. Receiver Overlay Code Select Addresses 3CHto 3FH are unused. Transmit Control Registers Address 40H: Bit 0 -- Transmit BPSK When this bit is set high, TXMCHP will provide the PN chip rate clock; when it is set low, the clock will be provided by the internal chip rate clock generator controlled by bits 5-0 of address 41H. This bit configures the transmitter for either BPSK or QPSK mode transmission. and differential encoding. If programmed for BPSK mode, data is requested by the STEL-2000A by a rising edge of output signal TXBITPLS, where TXBITPLS is generated once per symbol, one chip period before the end of the current symbol. At the end of the symbol duration, the TXIN data is latched into the device. TXBITPLS falls low immediately following the rising edge of TXIFCLK, which latches the TXIN value, and is generated repeatedly at the symbol rate as long as the input signal MTXEN remains high. Bit 3 -- Invert Symbol This bit allows the user to invert the I and Q channel bits following differential encoding and before being spread by the PN code. This function has the same effect as inverting the PN code, which may be useful in some cases. When this bit is set high, the encoded I and Q channel bits will be inverted; when it is set low, the I and Q channel bits will not be inverted. Address 41H: Bits 5-0 -- TXIFCLK Cycles per Chip In QPSK mode, data is requested by the STEL-2000A by a rising edge of output signal TXBITPLS, where this signal is generated in this mode twice per symbol, first one chip period before the middle of the symbol and then one chip period before the end of the symbol. TXBITPLS requests the data exactly one chip cycle before latching the TXIN data into the device. TXBITPLS falls low immediately following the rising edge of TXIFCLK, which latches the TXIN value. Bits 5-0 set the transmitter baseband PN chip rate to the frequency of TXIFCLK/(n+1), where n is the value stored in bits 5-0. The value of the data stored in bits 5-0 must range from 1 to 63 (01H to 3F H). This feature is useful when the PN chip rate required is an integer sub-multiple of the frequency of TXIFCLK. In cases where a chip rate is required that is not an integer sub-multiple of the frequency of TXIFCLK, the rate may be controlled externally using TXMCHP. When bit 0 of address 40H is set low, the transmitter is Address 42H: –––––––––––––––––––––––––––––––––––––––––––––––––––– 31 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Bits 5-0 -- Tx Chips per Data Symbol that used for the STEL-2000AÕs PN Matched Filter: for a code of length N, code chip (N-1) will be the first chip transmitted and will first be processed by Tap 0 of the PN Matched Filter; the last chip per symbol to be transmitted, however, will be chip 0, and at that time chip (N-1) will be processed by TapÊ(N-1) and chip 0 by Tap 0 to achieve peak correlation. Operation with the subsequent data symbols is analogous. The number of chips per data symbol in the transmitter is stored in bits 5-0 of address 42H. The unsigned value must range from 1 to 63 (01 H to 3FH), and the number of chips per data symbol will be this value plus 1. This value controls data symbol timing in the transmitter. Address 43H: Bits 5-0 -- Tx Chips per Acquisition/Preamble Symbol Address 4CH through 53H: Data Symbol Code The number of chips per Acquisition/Preamble symbol in the transmitter is stored in bits 5-0 of address 43H. The unsigned value must range from 1 to 63 Addresses 4CH to 53H contain the binary Data Symbol PN code sequence values. The storage capacity, assignments, and operation are similar to that of the Acquisition/Preamble PN code sequence values. The configuration of the bits stored is shown in Table 21. (01H to 3F H), and the number of chips per data symbol will be this value plus 1. This value controls the Acquisition/Preamble symbol timing in the transmitter. Addr 53H, Bits 7-0 Addresses 44 H through 4BH: Transmitter Acquisition/Preamble Symbol Code Code bits 63-56 Each STEL-2000A burst transmission begins with an Acquisition/Preamble symbol and is then followed by the actual information data symbols. Two separate and independent PN codes can be employed, one for the Acquisition/Preamble symbol, the other for the information symbols. Accordingly, the STEL-2000A Transmit PN Code Generators, like the receiverÕs PN Matched Filter, support independent PN codes up to ............. ............. Addr 4DH, Bits7-0 Code bits 15-8 Addr 4C H, Bits 7-0 Code bits 7-0 64 chips in length for the two modes. Addresses 44H to 4BH contain the binary Transmitter Acquisition/Preamble Symbol PN code chip values, where the configuration of the stored bits is as shown in Table 20. Table 21. Data Symbol Codes Address 54H: Bits 1-0 -- Transmitter Overlay Select The STEL-2000A incorporates programmable overlay code generators in both the transmitter and receiver. When enabled, the selected transmitter overlay code is subtracted from the data symbols, one overlay bit per symbol in both BPSK and QPSK modes. No synchronization is required since the codes in both the transmitter and the receiver are automatically synchronized by resetting the code generators at the start of each burst. The addition of the overlay codes randomizes the transmitted data sequence to guarantee that the spectrum of the transmitted signal will be adequately whitened and will not contain a small number of spectral lines even when the data itself is Addr 4BH, Bits 7-0 Code bits 63-56 ............. ............. Addr 45H, Bits 7-0 Code bits 15-8 Addr 44H, Bits 7-0 Code bits 7-0 Table 20. Acquisition/Preamble Symbol Codes not random. Three transmit and receive overlay codes can be selected, where they are each maximal length sequences with lengths of 63, 511 and 1023 symbols. The transmitter overlay codes are enabled and selected by the settings of bits 1-0 of address 54 H, as shown in Table 22. The length, N, of the Acquisition/Preamble symbol code is set by the value of (N-1) stored in bits 5-0 of address 43H. An internal counter begins the transmission with the PN code chip corresponding to that value. The last chip transmitted per symbol is then code chip 0. Note that this convention agrees with –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 32 Bits 1-0 in Addr. 54H Transmitter Data Symbols per Burst (bits 15-0) Overlay Code length and polynomial 0 Overlay code disabled 1 63: 1 + x2 + x 3 + x 5 + x 6 2 511: 1 + x2 + x 3 + x 5 + x 9 3 1023: 1 + x2 + x 3 + x 5 + x 10 The data stored as two bytes in addresses 55H (LS byte) and 56 H (MS byte) defines the number of data symbols per burst for the transmitter. This unsigned value must range from 1 to 65,535 (0001 H to FFFFH) for burst mode operation, and the number of data symbols per burst will be this value plus 1. Note that the range is slightly different from that in the receiver. Once the number of transmitted data symbols exceeds this number, the burst is assumed to have ended and the transmitter is immediately turned off. If the data value is set to 0000H the symbols per burst counter is disabled, permitting the STEL-2000A to be used for continuous transmission of data. Table 22. Transmitter Overlay Code Select Bit 2 -- Not Used This bit should be set to zero at all times. Address 55H through 56H: –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– REGISTER SETTING SEQUENCE re-written after re-enabling the receiver. Setting the MRXEN input low will have the same effect as setting bit 2 low in address 37 H, requiring address 38H to be rewritten after it is set high again. Similarly, setting bit 0 in address 37 H low will hold the NCO in a reset state, and after this bit has been set high again it will be necessary to reload the NCO either by writing 01H into address 00H or with a rising edge on the MFLD input. Setting the MNCOEN input low will have the same effect as setting bit 0 low in address 37 H, requiring the NCO to reloaded after it has been re-enabled. The majority of the registers are completely independent and can be set or modified in any order. However, there are inter-relationships between the settings of bits 2 and 0 in address 37H with the data in address 38H, and the need to reload the NCO after enabling it, respectively. Consequently, any time the data in address 37H is modified it will be necessary to attend to both of these functions again. Setting bit 2 low in address 37H will cause the data in address 38H to be set to zero. Therefore, if the RXTEST 7-0 bus is being used for reading any function except the Matched Filter I and Q inputs the value required must be –––––––––––––––––––––––––––––––––––––––––––––––––––– 33 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A DECIMAL, HEX AND BINARY ADDRESS EQUIVALENTS Dec. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Hex. 00H 01H 02H 03H 04H 05H 06H 07H 08H 09H 0A H 0BH 0CH 0DH 0EH 0F H 10H 11H 12H 13H 14H 15H 16H 17H 18H 19H 1A H 1BH 1CH Binary 0000000 0000001 0000010 0000011 0000100 0000101 0000110 0000111 0001000 0001001 0001010 0001011 0001100 0001101 0001110 0001111 0010000 0010001 0010010 0010011 0010100 0010101 0010110 0010111 0011000 0011001 0011010 0011011 0011100 Dec. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 Hex. 1DH 1EH 1F H 20H 21H 22H 23H 24H 25H 26H 27H 28H 29H 2A H 2BH 2CH 2DH 2EH 2F H 30H 31H 32H 33H 34H 35H 36H 37H 38H 39H Binary 0011101 0011110 0011111 0100000 0100001 0100010 0100011 0100100 0100101 0100110 0100111 0101000 0101001 0101010 0101011 0101100 0101101 0101110 0101111 0110000 0110001 0110010 0110011 0110100 0110101 0110110 0110111 0111000 0111001 Dec. 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 Hex. 3A H 3BH 3CH 3DH 3EH 3F H 40H 41H 42H 43H 44H 45H 46H 47H 48H 49H 4A H 4BH 4CH 4DH 4EH 4F H 50H 51H 52H 53H 54H 55H 56H Binary 0111010 0111011 0111100 0111101 0111110 0111111 1000000 1000001 1000010 1000011 1000100 1000101 1000110 1000111 1001000 1001001 1001010 1001011 1001100 1001101 1001110 1001111 1010000 1010001 1010010 1010011 1010100 1010101 1010110 Table 23. Decimal, Hexadecimal and Binary Address Equivalents –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 34 REGISTER SUMMARY Address Contents Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 00H 01H Bit 0 NCO Load Integrate & Dump Filter Viewport Control 02H 2Õs C. Input NCO CÕIn Inv. LF RXMSMPL Receiver Baseband Sampling Rate Control 03-06H NCO Frequency Control Word (32 bits) 07-16H Matched Filter Acquisition/Preamble Symbol Coefficients 17-26H Matched Filter Data Symbol Coefficients 27H FEP Disable 28H MF Viewport Control 29-2AH Acquisition/Preamble Symbol Threshold, Bits 9-0 2B-2CH Data Symbol Threshold, Bits 9-0 2DH Receiver Chips per Data Symbol 2EH Receiver Data Symbols per Burst, Bits 7-0 2F H 30H Missed Detects per Burst Threshold Rx Symb./ Burst Off Missed Det. Per Bst. Off Half Symb. Pulse Off BÕpass Force Cont. Acquis. Max. Power Sel. Manual Punctual 31H Man. Det. Enable Man. Det. 32H Man. Abort 33H AFC Viewport Control 34H CarryIn1/2 K2 On K2 Gain Value 35H LF Freeze K1 On K1 Gain Value LF Clr. Dis Unused (0) Signal Rotation Contrl. 36H Inv. O/p BPSK En. Rev. I & Q 37H Rx. En. Tx En. NCO En. 38H RXTEST7-0 Read Address 39H Matched Filter Power Saver 3A H Receiver Data Symbols per Burst, Bits 15-8 3BH Receiver Overlay Sel. IF Lpbk En MF LpbkEn 3C-3FH 40H Inv. Symb. TXMCHP OÕBin. O/p 41H TXIFCLK Cycles per Chip 42H Tx Chips per Data Symbol 43H Tx Chips per Acquisition/Preamble Symbol 44-4BH Transmitter Acquisition/Preamble Symbol Code (64 bits) 4C-53H Transmitter Data Symbol Code (64 bits) 54H 55-56H Set to 0 Tx BPSK Transmitter Overlay Sel. Transmitter Data Symbols per Burst, Bits 15-0 –––––––––––––––––––––––––––––––––––––––––––––––––––– 35 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Table 24. Register Summary –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 36 ELECTRICAL CHARACTERISTICS ABSOLUTE MAXIMUM RATINGS Note: Stresses greater than those shown below may cause permanent damage to the device. Exposure of the device to these conditions for extended periods may also affect device reliability. All voltages are referenced to VSS. Symbol Parameter Range Units T stg Storage Temperature Ð55 to +150 °C VDDmax Supply voltage on VDD Ð0.3 to + 7 Volts VI(max) Input voltage Ð0.3 to V DD+0.3 Volts Ii DC input current ± 10 mA RECOMMENDED OPERATING CONDITIONS Symbol Parameter Range Units VDD Supply Voltage +5 ± 5% Volts Ta Operating Temperature (Ambient) 0 to +70 °C D.C. CHARACTERISTICS Operating Conditions: VDD = 5.0 V ±5%, VSS = 0 V, Ta = 0° to 70° C Symbol Parameter IDDQ Supply Current, Quiescent IDD IDD Min. Typ. Max. Units Conditions 5.0 mA Static, no clock Supply Current, Operational 380 ➀ mA fRXIFCLK = 45.056 MHz ➀ (Transmitter + Receiver) 170 mA fRXIFCLK = 20 MHz Supply Current, Operational 20 mA fTXIFCLK = 45.056 MHz ➀ (Transmitter only, 9 mA fTXIFCLK = 20 MHz Volts Logic '1' Volts Logic '0' All inputs, VIN = VDD ÐÐÐÐÐÐ TXIFCLK, RXIFCLK, RESET fRXIFCLK = 0 MHz) VIH(min) High Level Input Voltage 0.7 x VDD VIL(max) Low Level Input Voltage IIH(min) High Level Input Current 10 µA IIL(max) Low Level Input Current Ð10 µA 0.2 x VDD only, VIN = VSS IIL(max) Low Level Input Current Ð130 VOH(min) High Level Output Voltage VDD Ð 0.4 VOL(max) Low Level Output Voltage CIN Input Capacitance COUT Output Capacitance Ð15 µA All other inputs, VIN = VSS VDD Volts IO = Ð2.0 mA, All outputs 0.4 Volts IO = +2.0 mA, All outputs 2 pF All inputs 4 pF All outputs Ð45 0.2 ➀ The typical operational supply current under various operating conditions can be calculated from the equation: IDD = 5 x f RXIFCLK + 13 x fCHIP mA where fRXIFCLK is the frequency of RXIFCLK, and fCHIP is the chipping rate, both in MHz Power = VDD x IDD –––––––––––––––––––––––––––––––––––––––––––––––––––– 37 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A TRANSMITTER INPUT/OUTPUT TIMING t CH t CL TXIFCLK t CT t CT TXCHPPLS TXBITPLS, TXTRKPLS, TXACQPLS tSU t HD DON'T CARE TXIN VALID DON'T CARE t CT TXIOUT, TXQOUT TXIFOUT A.C. CHARACTERISTICS Operating Conditions: VDD = 5.0 V ±5%, VSS = 0 V, Ta = 0° to 70° C Symbol Parameter fTXIFCLK TXIFCLK Frequency (TXIFOUT not used) Min. Max. Units Conditions 45.056 MHz STEL-2000A+45 only 20.0 MHz STEL-2000A+20 only MHz Both speed grades fTXIFCLK TXIFCLK Frequency (TXIFOUT in use) 20 tCH TXIFCLK Pulse width, High 10 nsec. tCL TXIFCLK Pulse width, Low 10 nsec. tSU TXIN to TXIFCLK setup 3 nsec. tHD TXIN to TXIFCLK hold 5 nsec. tCT TXIFCLK to TXBITPLS , TXTRKPLS, TXACQPLS, TXIOUT or TXQOUT delay 35 nsec. Notes: 1. The number of TXIFCLK cycles per cycle of TXCHPPLS is determined by the data stored in bits 5-0 of address 41H. It is shown as 2 in the above timing diagram but can be set from 2 to 64. 2. The width of the TXBITPLS , TXTRKPLS and TXACQPLS signal pulses is equal to the period of TXCHPPLS; i.e., equal to the PN chip period. 3. In QPSK mode, the TXBITPLS signal pulses high twice during each symbol period, once during the center chip and once during the last chip. If the number of chips per symbol is even, the number of chip periods between the TXBITPLS pulse at the end of the previous symbol and the one in the center of the symbol will be one more than the number of chip periods between the TXBITPLS pulse in the center of the symbol and the one at the end. The falling edge of the second pulse corresponds to the end of the symbol period. 4. The TXTRKPLS signal pulses high once each symbol period, during the last chip period of that symbol. The falling edge corresponds to the end of the symbol period. –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 38 5. The TXACQPLS signal pulses high once each burst, transmission, during the last chip of the Acquisition/Preamble symbol. The falling edge corresponds to the end of this symbol period. RECEIVER INPUT/OUTPUT TIMING tCH tCL RXIFCLK t SU t HD RXIIN, RXQIN tCR tCR RXSPLPLS RXSYMPLS RXDRDY tCD RXOUT, RXIOUT, RXQOUT A.C. CHARACTERISTICS Operating Conditions: VDD = 5.0 V ±5%, VSS = 0 V, Ta = 0° to 70° C, Symbol Parameter fRXIFCLK RXIFCLK Frequency Min. Max. Units Conditions 45.056 MHz STEL-2000A+45 only 20.0 MHz STEL-2000A+20 only tCH RXIFCLK Pulse width, High 10 nsec. tCL RXIFCLK Pulse width, Low 10 nsec. tSU RXIIN or RXQIN to RXIFCLK setup 3 nsec. tHD RXIIN or RXQIN to RXIFCLK hold 7 nsec. tCR RXIFCLK to RXSPLPLS, RXSYMPLS, or ÐÐÐÐÐÐÐÐÐ RXDRDY delay 35 nsec. tCD RXIFCLK to RXOUT, RXIOUT, or RXQOUT delay 35 nsec. Notes: 1. The number of RXIFCLK cycles per cycle of RXSPLPLS is determined by the data stored in bits 5-0 of address 02H. It is shown as 2 in the above timing diagram but can be set from 2 to 64. ÐÐÐÐÐÐÐÐÐ 2. The rising edge of RXDRDY should be used to clock out the data (RXOUT , RXIOUT, or RXQOUT). –––––––––––––––––––––––––––––––––––––––––––––––––––– 39 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A RXTEST BUS AND MICROPROCESSOR INTERFACE TIMING CSEL ADDR 6-0 DON'T CARE VALID VALID DON'T CARE DATA 7-0 VALID VALID DON'T CARE DON'T CARE t HD tSU WRITE tW OEN t ZV RXTEST 7-0 Hi-Z tVZ Hi-Z VALID A.C. CHARACTERISTICS Operating Conditions: VDD = 5.0 V ±5%, VSS = 0 V, Ta = 0° to 70° C, Symbol tSU tHD tW tZV tVZ Parameter ÐÐÐÐÐ CSEL , ADDR, DATA to WRITE setup ÐÐÐÐÐ WRITE to CSEL , ADDR, DATA hold Min. WRITE pulse width ÐÐÐÐ OEN low to RXTEST valid ÐÐÐÐ OEN high to RXTEST high impedance Max. Units 5 nsec. 5 nsec. 5 nsec. 3 12 nsec. 2 8 nsec. Conditions –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 40 APPENDIX I: THEORY OF OPERATION Digital Downconversion In Direct I.F. Sampling Mode, the sampled signal is presented as input to the receiverÕs I channel input (RXIIN) and the Q channel input (RXQIN) is held to zero (where ÒzeroÓ is defined by the ADC input format ). As a result, only two of the four multipliers in the DownconverterÕs complex multiplier are used and the device does not make a true single-sideband downconversion from I.F. to baseband. In Quadrature Sampling Mode, by contrast, quadrature inputs to two ADCs provide I and Q inputs to the STEL2000A and the full complex multiplier is used. An illustration of the operation of Direct I.F. Sampling Mode is shown in the frequency domain in Figure 7, where the spectra have been drawn asymmetrically so that spectral inversions can be readily identified. The STEL-2000A receiverÕs downconverter circuitry allows use of two distinct modes, where the mode chosen will depend upon the application. In applications where the received PN chip rate is less than approximately 1 /8 of the I.F. sample clock (RXIFCLK) rate, the STEL-2000A can be used with a single A/D converter (ADC) and operate in Direct I.F. Sampling Mode. For higher chip rate applications, it is necessary to use the STEL-2000A in the full Quadrature Sampling Mode; i.e., using a quadrature signal source, two ADCs, and the on-chip NCO in its quadrature mode. Using the STEL-2000A with a Single ADC in Direct I.F. Sampling Mode Direct I.F. Sampling Mode allows one rather than two ADCs to be used, as will be explained below. If appropriate for the application, use of Direct I.F. Sampling Mode can reduce the system cost since quadrature downconversion or splitting and the second ADC used in Quadrature Sampling Mode are not required. The trade-off, however, is in the lower maximum PN chip rate that can be supported by the STEL-2000A in Direct I.F. Sampling Mode as compared to the maximum rate that can be supported by Quadrature Sampling Mode. The spectrum of a real input signal with center (I.F.) frequency of f1 and signal bandwidth B is shown in line 1 of Figure 7. The bandwidth B is the two-sided bandwidth, corresponding to a PN chip rate of 1/2ÊB Mcps. Note that throughout this discussion it is assumed that the signal bandwidth does not exceed 1/2Êf 1/ S A ; i.e., B < 2ÊfSA . Otherwise, the mixing and sampling processes to be described will result in destructive in-band aliasing. Also, clearly, the I.F. frequency must be able to support the signal bandwidth; i.e., 1/2ÊBÊ<Êf1. BANDWIDTH: B INPUT SPECTRUM 1 FREQ. –f1 f1 SPECTRUM OF SAMPLING PROCESS 2 –fSA FREQ. fSA 0 SPECTRUM AFTER A/D 3 –fSA 4 0 FREQ. –f1 0 –f1 0 f1 fSA QUADRATURE NCO SPECTRUM FREQ. –fSA 5 fSA SPECTRUM AFTER MIXER 2 f1 fSA –2 f 1 FREQ. –fSA 6 fSA 0 SPECTRUM AFTER IDEAL DIGITAL LOW PASS FILTER FREQ. –fSA 0 fSA Figure 7: Spectra of Signals in Direct I.F. Sampling Mode The input signal is sampled at the frequency f SA, the resulting spectrum is shown in line 3. As can be where the sampling spectrum is shown in line 2 and seen, the fundamental and harmonics of the sampling –––––––––––––––––––––––––––––––––––––––––––––––––––– 41 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A frequency result in images of the input signal spectrum at other frequencies, where here the images are centered about multiples of the sampling frequency. In other words, the spectrum of the sampled signal shown in line 3 contains aliases of the input signal at frequencies f 1 ± n fSA, where n can assume both positive and negative integer values. Since the sampling process is linear, no spectral inversion occurs; i.e., the original spectrum is translated along the frequency axis with no mirror reflections of the input spectrum created. true as long as the input spectrum is only defined for frequencies within a non-primary Nyquist region; i.e., defined only over frequencies f such that: (nÊ-Ê1Ê/2 )ÊfSAÊ<Ê|ÊfÊ|Ê<Ê(nÊ+Ê1Ê/2 )ÊÊfSA for positive integer values of n. Direct I.F. Sampling Mode with this type of signal is shown in Figure 8, where it can be seen that in line 3 the diagramÕs high frequency input has the same spectrum after sampling as does the low frequency input in Figure 7; consequently, all subsequent operations are identical to those in Figure 7. This result stems from the periodic nature of sampling: sampling an input frequency f1 is theoretically indistinguishable from sampling an input frequency (n fSA + f 1 ) for positive integer n and positive f1 < 1/Ê2 fSA . A slightly different result obtains, however, when sampling an input frequency (n fSA - f1 ), again for positive integer n and positive f1 < 1/Ê2 f SA. In this case, the positions of the spectrally inverted and spectrally correct aliases will be interchanged when compared with an input frequency of (n f SA + f1 ). As a consequence, the desired baseband signal after downconversion and filtering will also be spectrally inverted. This phenomenon is equivalent to high-side conversion; i.e., downconversion of a signal by means of a local oscillator at a frequency higher than the carrier frequency. If the modulation type is QPSK, demodulation of a spectrally inverted signal will result in the inversion of the Q channel data (which can be readily corrected); if the modulation type is BPSK, there is no effect on the demodulated data. The STEL-2000AÕs NCO provides a quadrature (sine and cosine) output that defines a complex signal. Line 4 shows its spectrum as an impulse at frequency -f1 , where the minus sign reflects the signalÕs use in downconversion and the absence of a positive impulse at frequency +f 1 results because the NCO output is truly complex. Aliases of this impulse are shown offset by integer multiples of fSA to reflect the sampled nature of the NCO output. When the input sampled signal of line 3 is then modulated with the complex signal of the STEL-2000AÕs quadrature NCO of line 4, the signal spectrum after mixing is as shown in line 5. The sections shown inside the shaded areas are the aliases of the baseband signal beyond the Nyquist frequency and are not of concern. The signals inside the primary baseband Nyquist region (|ÊfÊ|Ê<Ê1/2 fSA ) consist of the desired signal and a spectrally reversed or inverted image signal with center frequency separated from that of the desired signal by 2 f 1 , twice the I.F. frequency before sampling. This image signal can be removed by a subsequent ideal lowpass filter as shown in line 6. The above discussion has assumed ideal lowpass filtering to recover the desired signal at baseband, but, in the STEL-2000AÕs Downconverter, an ideal lowpass filter is not available. The quadrature Integrate and Dump filter of the Downconverter serves this purpose instead. The DownconverterÕs I n t e g r a t e a n d D u m p In Figure 7 the input signal is shown at a low I.F. frequency such that f1 < 1Ê/2 f SA ; i.e., the signal is only defined inside the primary Nyquist region. Provided, however, that B < 1Ê/2 fSA, that condition need not be BANDWIDTH: B INPUT SPECTRUM 1 FREQ. 0 –fSA – f1 fSA + f1 SPECTRUM OF SAMPLING PROCESS 2 FREQ. 0 –fSA fSA SPECTRUM AFTER A/D 3 –fSA FREQ. –f1 0 f1 fSA Figure 8: Direct I.F. Sampling Mode With I.F. Frequency (fSA + f 1 ) > Sampling Frequency fSA –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 42 Gain 0 dB –5 –10 –15 –20 –25 –30 –35 –400 1 2 Nf / fSA 3 4 5 Figure 9: G(ω) = sin(ωÕ)/ωÕ, where ωÕ = (2πNf)/fSA least ~13 dB. This sidelobe peak is a worst case, and much of the alias energy outside the desired band will be attenuated by more than 13 dB. Nonetheless, the presence of unattenuated energy from the unwanted alias degrades performance. It is for this reason that Direct I.F. Sampling Mode is only recommended for received PN chip rates less than 1Ê/8 Êf SA ; in other words, for BÊ<Ê1/Ê4 ÊfSA. The attenuation realized by the Integrate and Dump filter is then further determined by the choice of the I.F. frequency f1 and the I.F. sampling rate fSA. filter is a decimation filter, integrating input samples over a programmable number of sample periods, N, so that the output sampling rate is ( 1/N)th of the input sampling rate and the I.F. sampling rate fSA is decimated to the baseband sampling rate. Since the STEL2000AÕs PN Matched Filter requires two samples per chip, the baseband sampling rate must be at twice the PN chip rate and N must equal fSA /B. When the sampling rate is much greater than the signal bandwidth (or, equivalently, the chip rate), the Integrate and Dump filter is most effective in attenuating the unwanted aliased image. This performance can be seen from the transfer function G(ω) of a decimation filter, where: The choice of the I.F. frequency and sampling rate is crucial so that the unwanted alias of the signal in the baseband Nyquist region lies as far as possible from the desired signal to permit maximum attenuation. The optimum separation of the desired signal and the unwanted alias occurs when the alias is centered at the bounds of the baseband Nyquist region, |ÊfÊ|Ê=Ê1Ê/2ÊfSA as shown in Figure 10. In this case, the desired signal is equally spaced from the unwanted aliases in both the positive and negative frequency domains and f1 = 1Ê/4 ÊfSA. Consider, then, the worst case appropriate for Direct I.F. Sampling Mode. If BÊ<Ê1Ê/4 Êf SA as has been said to be appropriate for Direct I.F. Sampling Mode, then NÊ=Êf SA/BÊ=Ê4, the break frequency is 3Ê/8 ÊfSA or greater, and the attenuation provided by the Integrate and Dump filter is given by the curve of Figure 9 for values of (N f / fSA) greater than 3Ê/2 . Here, the attenuation is at least ~21 dB, offering much better attenuation of the unwanted alias than in the previous worst case example. Further analysis shows that if the input SNR is 15 dB, then the alias attenuated by 21 dB will reduce the SNR by approximately 1 dB. G(ω) = sin(ωÕ)/ωÕ and ωÕ = (2πNf)/fSA. Figure 9 shows a plot of the gain of this transfer function as a function of the normalized frequency (Nf/fSA ). To effect the desired low pass filter and eliminate the aliased image in the baseband Nyquist region appearing in line 5 of Figure 7, the attenuation must be suitably high for frequencies greater than, in the worst case, 1/2 ÊB. Given a defined signal bandwidth B, however, judicious choice of f 1 and fS A allows a higher break frequency to be chosen, as will be discussed. As an extreme worst case, if f1 = 1Ê/4ÊfSA and BÊ=Ê1/Ê2 ÊfSA, corresponding to the highest chip rate that can be handled for a given value of fSA, then the break frequency must be 1Ê/2 ÊB (equal to 1Ê/4ÊfSA ). In this example, then, N = f SA/B = 2 and the attenuation provided by the Integrate and Dump filter is given by the curve of Figure 9 for values of (N f / fSA) greater than 1Ê/2 . As can be seen, the attenuation will be at least equal to the peak of the corresponding lobe or at 5 SPECTRUM AFTER MIXER 2 f1 fSA – 2 f1 FREQ. –fSA fSA Figure 10: Optimum Condition for Bandpass Sampling –––––––––––––––––––––––––––––––––––––––––––––––––––– 43 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Quadrature Sampling Mode The optimum choice of I.F. frequency discussed above can be extended beyond the primary Nyquist region. Since an I.F frequency of nÊfSA + f 1 produces exactly the same result for any value of n, the general condition for optimum separation of the desired signal and the unwanted alias is: Quadrature Sampling Mode requires that quadrature I and Q channel I.F. inputs are sampled by two ADCs and input to the STEL-2000AÕs Downconverter. All four multipliers of the DownconverterÕs complex multiplier are then used to perform true single sideband downconversion to baseband. Quadrature inputs imply that the input signal is complex, and the input signal spectrum shown in line 1 of Figure 11 is thus only defined for positive frequencies. As a result, the image alias within the primary Nyquist region associated with Direct I.F. Sampling Mode does not appear and does not have to be attenuated by the Integrate and Dump filter. As in the prior discussion, this analysis holds as long as B < 1Ê/2 f SA, 1Ê/2 B < f1 , and the input spectrum is only defined for frequencies within a single Nyquist region; i.e., defined only over frequencies f such that: (nÊ-Ê1Ê/2 )ÊfSAÊ<Ê|ÊfÊ|Ê<Ê(nÊ+Ê1Ê/2 )ÊÊfSA for positive integer values of n. f1 = nÊf SAÊ+Ê1Ê/4 ÊfSA ; B< 1Ê/2 ÊfSA for positive integer n and positive B and f1. And, if care is taken to handle the effect of high side conversion, the following I.F. frequencies also fulfill the optimum condition: f1 = nÊf SAÊ-Ê1Ê/4ÊfSA ; B< 1Ê/2 ÊfSA for positive integer n and positive B and f1. Using the STEL-2000A with Two ADCs in BANDWIDTH: B INPUT SPECTRUM 1 FREQ. –f1 0 f1 SPECTRUM OF SAMPLING PROCESS 2 –fSA FREQ. fSA 0 SPECTRUM AFTER A/D 3 –fSA FREQ. –f1 0 f1 fSA QUADRATURE NCO SPECTRUM 4 FREQ. –f1 –fSA 5 0 fSA 0 fSA SPECTRUM AFTER MIXER FREQ. –fSA 6 SPECTRUM AFTER IDEAL DIGITAL LOW PASS FILTER FREQ. –fSA fSA 0 Figure 11: Spectra of Signals in Quadrature Sampling Mode –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 44 Differential Demodulation and that provided by the Downconverter. For DBPSK, the data modulation differences ∆¯mod can take only the values of 0° or 180°. Expressing the complex phase difference [¯(k)-¯(k-1)] in terms of these components, the decision can be seen to be based on: As noted in the preceding text, computation of the ÒdotÓ and ÒcrossÓ products is fundamental to operation of the DPSK Demodulator and Frequency Discriminator. Let Ik and Q k represent the I and Q channel inputs, respectively, for the k th symbol after downconversion and despreading. The dot and cross products can then be defined as: Dot(k) sout(k) = A(k).A(k-1).ej¯(k).e-j¯(k-1) = Ik Ik-1 + Qk Qk-1; and, = A(k).A(k-1).ej[∆¯mod(k) +∆¯rot(k) ] Cross(k) = Qk Ik-1 - Ik Qk-1. For DBPSK, only the real part of sout (k), Dot(k), is needed to determine the modulated phase transition: In the complex domain, these products can be seen to have been defined to form the complex conjugate product between two input samples, one symbol apart. Let the kth input sample, sin(k), be defined as: Dot(k) = A(k).A(k-1).cos(∆¯ mod(k)+∆¯rot(k)) = ±A(k).A(k-1).cos(∆¯ rot(k)) sin(k) ≡ I(k) + j Q(k), where the sign is determined by the transmitted data since cos[∆¯mod(k)] = ±1. As a result, where I(k) and Q(k) are the 8-bit peak power PN Matched Filter I and Q channel outputs directed to the DPSK Demodulator. In polar form, sin(k) may be conveniently defined as: Dot(k) ≈ ±A2(k) if the amplitude of the signal is constant for consecutive symbols and if the phase rotation ∆¯r o t(k) between symbols is small. The STEL-2000A DPSK Demodulator can thus use the sign of the dot product in order to make DBPSK symbol decisions without the introduction of any fixed phase rotation. sin(k) ≡ A(k).ej¯(k) with A(k) ≡ I2 (k)+Q 2 (k) QPSK Demodulation Q(k) ¯(k) ≡ arctan I(k) . ωfixed = arbitrary fixed phase rotation; For DQPSK modulation, the possible phase shifts between successive symbols due to the modulation are 0°, 90°, 180°, and 270°. Here, introduction of a phase shift (ωfixed) of ±45° to the previous symbol in the calculation of the dot and cross products is desired in order shift the possible phase differences to 45°, 135°, 225°, or 315° so that the DQPSK decision boundaries coincide with the signs of the dot and cross products. In the STEL-2000A DPSK demodulator, phase rotation is accomplished in the signal rotation block by the following transformation of the I and Q channel values:: Dot(k) = Re[sout(k)]; and, Irot(k) = [ I(k) - Q(k) ]/2 for 45° rotation Cross(k) = Im[sout(k)]. Irot(k) = [ I(k) + Q(k) ]/2 for -45° rotation The fixed phase rotation ωfixed has been introduced to later simplify the decision criteria, but the ability to express real and imaginary parts of the complex conjugate product between consecutive symbols with the dot and cross products is the key to their use in DPSK demodulation. Qrot(k) = [ I(k) + Q(k) ]/2 for 45° rotation Qrot(k) = -[ I(k) + Q(k) ]/2 for -45° rotation Simple substitution then shows that the complex conjugate product between consecutive symbols (with an arbitrary phase shift introduced to the previous symbol value) may be expressed as:: sout(k) = sin(k) . [s in(kÐ1) . ωfixed]* ≡ Dot(k) + j Cross(k) where The divide by 2 is part of the signal rotation function. This transformation is equivalent to multiplying by (1 ± j)/2 or (1/√2)ej¯(fixed) where ¯fixed is ±45°. In this case, sout(k) becomes: BPSK Demodulation In DPSK, the phase difference between successive samples is due to the data modulation phase differences, ∆¯ m o d , plus any induced phase rotation between symbols, ∆¯rot, resulting from, for example, a frequency offset between the received signalÕs I.F. sout(k) = A(k).A(k-1).ej¯(k).e-j¯(k-1).[ωfixed]* = A(k).A(k-1).ej[∆¯mod(k) +∆¯rot(k) ]. (1/√2)ej¯(fixed) –––––––––––––––––––––––––––––––––––––––––––––––––––– 45 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A so that Dot(k) The frequency discriminator function or error signal can similarly be generated based on the dot and cross products. The objective is an error signal reflecting the sine of the phase difference between the present and prior symbol after correcting for the estimated phase increments due to data modulation. In the STEL-2000A Frequency Discriminator, the frequency error is calculated through a decision-directed crossproduct algorithm and is then used with the Loop Filter to correct the NCO frequency. Assuming an input sin(k), where: ≈ (1/√2)A(k).A(k-1).cos(∆¯ mod(k) - ¯fixed ) Cross(k) ≈ (1/√2)A(k).A(k-1).sin(∆¯mod(k) - ¯fixed ) where the phase rotation ∆¯rot(k) due to the frequency offset between symbols has been assumed negligible. A summary of the Dot(k) and Cross(k) products for the possible values of ∆¯ mod(k) and ¯fixed is shown below, illustrating how the sign of the dot and cross products allow the symbol decision to be made: ¯fixed = -45° ∆¯mod(k) Dot(k) Cross(k sin(k) = I(k) + j Q(k), ¯fixed = +45° Dot(k) ) Cross(k the algorithm calculates the frequency discriminator function for DBPSK, sAFC/BPSK(k), as: ) sAFC/BPSK(k) = SIGN[Dot(k)] . Cross(k) 0° +A2 +A2 +A2 -A2 = SIGN[Dot(k)] . A(k).A(k-1).sin(¯(k)-¯(k-1)) 90° -A2 +A2 +A2 +A2 = SIGN[Dot(k)].A(k).A(k-1).sin(∆¯mod(k)+∆¯rot(k)) 180° -A2 -A2 -A2 +A2 ≈ SIGN[Dot(k)] . A 2 (k).cos[∆¯mod(k)].sin[∆¯rot(k)] 270° +A2 -A2 -A2 -A2 ≈ A2 (k).sin[∆¯rot(k)]. Table 25. Dot and Cross Product Summary The final result assumes that the amplitude of the signal is constant over consecutive symbols and shows that the discriminator function is directly related to the change in phase between successive symbols. Since the interval between successive symbols is fixed, the discriminator function can be interpreted as a frequency error signal. DQPSK Demodulation The STEL-2000A DPSK Demodulator decision logic is designed so that correct DQPSK decisions are made with a signal rotation of ¯ FixedÊ=Ê-45°. For π/4 QPSK modulation, however, the modulator itself inserts 45° between consecutive symbols, and the possible phase shifts between successive symbols due to modulation are 45°, 135°, 225°, and 315°. As a result, the DPSK Demodulator should be configured for π/4 QPSK with ¯FixedÊ=Ê0°. For DQPSK signals, the STEL-2000A computes the discriminator function sAFC/QPSK(k) as: sAFC/QPSK(k) = SIGN[Dot(k)].Cross(k).SIGN[Cross(k)].Dot(k) where the above expression can be reduced to the same as for DBPSK, Frequency Error Signal Generation sAFC/QPSK(k) ≈ A2(k).sin(∆¯rot(k)). –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 46 Using the Modulator in the STEL-2000A Transmitter The STEL-2000A incorporates a Direct Digital Synthesizer (DDS) based BPSK/QPSK modulator which can be used to generate the transmit output signal at some IF frequency, thereby eliminating the need for a separate modulator. Because it is a sampled data system, like the other parts of the system, care must be taken to ensure that the results of aliasing do not adversely affect the transmit signal. The sampling clock for the modulator is fRXIFCLK, since the NCO is driven by this clock, and for this reason both TXIFCLK and RXIFCLK must be common when the modulator is being used. When a DDS is used to generate an unmodulated signal the results are easy to predict because the stepped sinewave generated by the DDS has spectral lines at the frequencies n.fCLK ± fOUT for all integer values of n. However, when the DDS signal is modulated with a series of rectangular pulses each one of these spectral lines is replaced by a sinc function (i.e., sin(x)/x), the spectrum of the modulating signal itself, each with a very significant bandwidth. Any part of the spectrum of the baseband component (i.e. the one where n = 0 in the equation n.fCLK ± fOUT) above the Nyquist frequency (fCLK/2) results in its first alias (i.e. the one where n = 1 in the equation n.f CLK ± fOUT) intruding below the Nyquist frequency, and this may interfere with the desired signal. This is called aliasing distortion. A typical example is shown in Figure 12. The baseband component is shown as a the thick curve and the first alias is shown as the lighter curve; the higher order aliases are ignored for simplicity. Note that in this example the nulls between the sidelobes of the baseband signals coincide with those of the first alias; this is simply the result of the numbers chosen, and this will not occur in all cases. In a practical system the output signal would typically be filtered to eliminate the sidelobes, as well as the signals above the Nyquist frequency. As can be seen, the second and third sidelobes of the first alias fall right into this band of the spectrum, distorting the desired signal. The magnitude of this signal at the peak of the sidelobes is approximately Ð18 dBc. Clearly some improvement can be obtained by choosing a carrier frequency that increases the separation of the baseband main lobe from the first alias main lobe, as shown in Figure 13. 2 x CR 0 dB –13 –18 –21 –23 –25 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Nyquist Frequency Frequency/fTXIFCLK Figure 12: Spectrum of DDS modulated at 0.1 x f RXIFCLK when carrier frequency is set to 0.4 x fRXIFCLK 2 x CR 0 dB –13 –18 –21 –23 –25 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Nyquist Frequency Frequency/fTXIFCLK Figure 13: Spectrum of DDS modulated at 0.1 x f RXIFCLK when carrier frequency is set to 0.25 x fRXIFCLK –––––––––––––––––––––––––––––––––––––––––––––––––––– 47 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A overlaps the spectrum of the baseband main lobe, resulting in very sig nificant aliasing distortion which cannot be eliminated by filtering. This would severely effect the performance of the system and in general will be completely unacceptable. Reducing the carrier frequency to 25% of the clock dramatically reduces the distortion level again, as shown in Figure 15. In this case the carrier frequency has been reduced to 0.25 x fRXIFCLK and now the fourth and fifth sidelobes of the first alias fall in the same part of the spectrum as the baseband main lobe, reducing the distortion to approximately Ð23 dBc at the peak of the sidelobe. In both of the cases shown above, especially the second, the level of the distortion is low enough that the performance penalty would not be very great since the effective distortion is reduced by the processing gain in the despreader. However, in both of these cases the chipping rate was a very modest 10% of the frequency of the system clocks; if the chipping rate is increased to 40% the situation is very different, as shown in Figure 14. Although the distortion is still fairly severe, adequate performance may still be obtainable as a result of the processing gain, but the performance would be many dBs off the theoretical limit. The conclusion is that as the chipping rate of the system increases so does the effect of the aliasing distortion in the modulator, and the performance will degrade. A practical limit may be in the region of 15% of the clock frequency, but in reality every case is unique and should be evaluated before deciding to use the internal DDS modulator. Here, both the chipping rate and the carrier frequency have been set at 20% of the clock frequency and now the baseband main lobe straddles the Nyquist frequency. Consequently, the first alias of the main lobe 2 x CR 0 dB –13 –18 –21 –23 –25 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Nyquist Frequency Frequency/fTXIFCLK Figure 14: Spectrum of DDS modulated at 0.4 x f RXIFCLK when carrier frequency is set to 0.4 x fRXIFCLK 2 x CR 0 dB –13 –18 –21 –23 –25 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Nyquist Frequency Frequency/fTXIFCLK Figure 15: Spectrum of DDS modulated at 0.4 x f RXIFCLK when carrier frequency is set to 0.25 x fRXIFCLK –––––––––––––––––––––––––––––––––––––––––––––––––––– STEL-2000A Powered by ICminer.com Electronic-Library Service CopyRight 2003 48 APPENDIX II: TYPICAL APPLICATION SPREAD SPECTRUM BPSK/QPSK TRANSCEIVER BASEBAND MODULE 4x R X C LO C K R X IFC LK D A TA OUT R X IOUT RX C LO C K RXDRDY 8 R X IIN A /D LPF R X QOUT RF TR A N SC EIV ER STEL-2000A D A TA IN T X IN TX C LO C K T X B IT P LS 4x TX C H IP C LO C K LINEAR RF MODULE T X IFOUT 8 D /A LPF FIX ED O SC ILLA TO R (U p/D ow n C onv er t er ) T X IFC LK Figure 16: Spread Spectrum BPSK/QPSK Transceiver –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– –––––––––––––––––––––––––––––––––––––––––––––––––––– 49 Powered by ICminer.com Electronic-Library Service CopyRight 2003 STEL-2000A Information in this document is provided in connection with Intel® products. No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted by this document. Except as provided in Intels Terms and Conditions of Sale for such products, Intel assumes no liability whatsoever, and Intel disclaims any express or implied warranty, relating to sale and/or use of Intel® products including liability or warranties relating to fitness for a particular purpose, merchantability, or infringement of any patent, copyright or other intellectual property right. Intel products are not intended for use in medical, life saving, or life sustaining applications. Intel may make changes to specifications and product descriptions at any time, without notice. For Further Information Call or Write INTEL CORPORATION Cable Network Operation 350 E. Plumeria Drive, San Jose, CA 95134 Customer Service Telephone: (408) 545-9700 Technical Support Telephone: (408) 545-9799 FAX: (408) 545-9888 Copyright © Intel Corporation, December 15, 1999. All rights reserved Powered by ICminer.com Electronic-Library Service CopyRight 2003