Baron Services DSSR-250C Pulsar Digital Solid-State Radar System User Manual

Baron Services Inc Pulsar Digital Solid-State Radar System

Introduction and Specifications

Introduction and SpecificationsRVP8 User’s ManualMarch 20061–11. Introduction and SpecificationsThe RVP8 LineageSIGMET Inc. has a 20-year history of supplying innovative, high-quality signal processingproducts to the weather radar community. The history of SIGMET products reads like a historyof weather radar signal processing:Year Model UnitsSold Major Technical Milestones1981 FFT 10 First commercial FFT-based Doppler signal processor for weath-er radar applications. Featured Simultaneous Doppler and inten-sity processing.1985 RVP5 161 First single-board low-cost Doppler signal processor. First com-mercial application of dual PRF velocity unfolding algorithm.1986 PP02 12 First high-performance commercial pulse pair processor with18.75-m bin spacing and 1024 bins.1992 RVP6 150 First commercial floating-point DSP-chip based processor. Firstcommercial processor to implement selectable pulse pair, FFT orrandom phase 2nd trip echo filtering.1996 RVP7 >200 First commercial processor to implement fully digital IF process-ing for weather radar.2003 RVP8 First digital receiver/signal processor to be implemented using anopen hardware and software architecture on standard PC hard-ware under the Linux operating system. Public API’s are pro-vided so that customers may implement their own custom proc-essing algorithms.Much of the proven, tested, documented software from the highly-successful RVP7 (written inC) is ported directly to the new RVP8 architecture. This allows SIGMET to reducetime-to-market and produce a high-quality, reliable system from day one. However, the newRVP8 is not simply a re-hosting of the RVP7. The RVP8 provides new capabilities for weatherradar systems that, until now, were not available outside of the research community.Advanced Digital Transmitter OptionFor example, the RVP8 takes the next logical step after a digital receiver- a digitally synthesizedIF transmit waveform output that is mixed with the STALO to provide the RF waveform to thetransmitter amplifier (e.g., Klystron or TWT). The optional RVP8/Tx card opens the door foradvanced processing algorithms such as pulse compression, frequency agility and phase agilitythat were not possible before, or done in more costly ways.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–2Open Hardware and Software DesignCompared to previous processors that were built around proprietary DSP chips, perhaps the mostinnovative aspect of the RVP8 is that it is implemented on standard PC hardware and softwarethat can be purchased from a wide variety of sources. The Intel Pentium/PCI approach promisescontinued improvement in processor speed, bus bandwidth and the availability of low–costcompatible hardware and peripherals. The performance of an entry level RVP8 (currently dual2.4 GHz Pentium processors) is 6 times faster than the fastest RVP7 ever produced (with twoRVP7/AUX boards).Aside from the open hardware approach, the RVP8 has an open software approach as well. TheRVP8 runs in the context of the Linux operating system. The code is structured and public API’sare provided so that research customers can modify/replace existing SIGMET algorithms, orwrite their own software from scratch using the RVP8 software structure as a foundation onwhich to build.The advantage of the open hardware and software PCI approach is reduced cost and the abilityfor customers to maintain, upgrade and expand the processor in the future by purchasingstandard, low cost PC components from local sources.SoftPlane High–Speed I/O InterconnectThere are potentially many different I/O signals emanating from the backpanel of the RVP8.Most of these conform to well-known electrical and protocol standards (VGA, SCSI, 10–BaseT,RS-232 Serial, PS/2 Keyboard, etc.), and can be driven by standard commercial boards that areavailable from multiple vendors. However, there are other interface signals such as triggers andclocks that require careful timing. These precise signals cannot tolerate the PCI bus latency. Forsignals that have medium–speed requirements (~1 microsec latency) for which the PCI bus isinappropriate; and others that require a high–speed (~ 1 ns latency) connection that can only beachieved with a dedicated wire, the RVP8 Softplanet provides the solution.Physically, the Softplanet is a 16-wire digital “daisy-chain” bus that plugs into the tops of theRVP8/Rx, RVP8/Tx, and I/O boards. The wires connect to the FPGA chips on each card, and thefunction of each wire is assigned at run–time based on the connectivity needs of the overallsystem. The Softplanet allocates a dedicated wire to carry each high-speed signal; but groups ofmedium-speed signals are multiplexed onto single wires in order to conserve resources. Eventhough there are only 16 wires available, the Softplane is able to carry several high-speed signalsand hundreds of medium–speed signals, as long as the total bandwidth does not exceed about600MBits/sec.The Softplanet I/O is configured at run–time based on a file description rather than customwiring such as wirewrap. Neither the PCI backplane nor the physical Softplanet are customizedin any way. Since there is no custom wiring, a failed board can be replaced with a genericoff–the–shelf spare, and that spare will automatically resume whatever functions had beenassigned to the original board. Similarly, if the chassis itself were to fail, then simply pluggingthe boards into another generic chassis would restore complete operation. Cards and chassis canbe swapped between systems without needing to worry about custom wiring.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–3Standard LAN Interconnection for Data Transfer or Parallel ProcessingFor communication with the outside world, the RVP8 supports as standard a 10/100/1000 Base TEthernet. For most applications, the 100 BaseT Ethernet is used to transfer moment results (Z, T,V, W) to the applications host computer (e.g., a product generator). However, the gigabitEthernet is sufficiently fast to allow UDP broadcast of the I and Q values for the purpose ofarchiving and/or parallel processing. In other words, a completely separate signal processor caningest and process the I and Q values generated by the RVP8.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–41.1 System Configuration ConceptsThe hardware building blocks of an RVP8 system are actually quite few in number:SRVP8/IFDt IF Digitizer Unit- This is a separate sealed unit usually mounted in thereceiver cabinet. The primary input to the IFD is the received IF signal. In addition, theIFD has channels to sample the transmit pulse and to take in an external clock to phaselock the A/D conversion with the transmit pulse (not used for magnetron systems).SRVP8/Rxt Card- A PCI card mounted in the chassis. It connects to the IFD by aCAT-5E cable which can be up to 25m long. In addition, there are two BNC triggeroutputs and four RS-422 programmable I/O signals.SI/O-62t Card and Connector Panel- These handle all of the various I/O associatedwith a radar signal processor, such as triggers, antenna angles, polarization switchcontrols, pulse width control, etc. The Connector Panel is mounted on either the front orrear of the equipment rack and a cable (supplied) connects the panel to the I/O-62.SOptional RVP8/Txt card- This supplies two IF output signals with programmablefrequency, phase and amplitude modulation. In the simplest case it might merely supplythe COHO which is mixed with the STALO to generate the transmit RF for Klystron orTWT systems. More interesting applications include pulse compression and frequencyagility scanning. This card is not necessary for magnetron systems.SPC Chassis and Processor with various peripherals- a robust 4U rack mount unit witha dual-Xeon mother board, diagnostic front panel display, disk (mechanical or flash),CDRW, keyboard, mouse and optional monitor for local diagnostic work. Redundantpower supplies are used, and there are redundant fans as well.This modular hardware approach allows the various components to be mixed and matched tosupport applications ranging from a simple magnetron system to an advanced dual polarizationsystem with pulse compression. Typically SIGMET supplies turn-key systems, although someOEM customers who produce many systems purchase individual components and integrate themby themselves. This allows OEM customers to put their own custom “stamp” on the processorand even their own custom software if they so choose.For the turnkey systems provided by SIGMET, the basic chassis is a 6U rack mount unit asdescribed above. A 2U chassis can be provided for applications for which space is limited. Avery low cost approach is to use a desk side PC, but this is not recommended for applicationsthat require long periods of unattended operation.To illustrate various RVP8 configurations, some typical examples are shown below. For clarity,all the examples show the single–board computer approach. A mother board approach isequivalent.Example 1: Basic Magnetron SystemThe building blocks required to construct the basic system are:
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–5KeyboardMouseMonitor10/100 BaseT LAN InterfaceIFD Fiber DownlinkCOAX UplinkRS232C Antenna AnglesRVP8 Configuration Example: Basic Magnetron SystemTriggersIF SignalIF Magnetron BurstDAFCDigital STALOOptionalUtilities14-BitSBCRVP8/RxSIFD- IF Digitizer installed in the radar receiver cabinet. This can be located up to 100 metersfrom the RVP8 main chassis (fiber optic connection). The DAFC (Digital AFC) is an option tointerface to a digitally controlled STALO. Like the RVP7, the RVP8 provides full AFC withburst pulse auto-tracking.SRVP8/Rx- The digital receiver collects digitized samples from the IFD and does the processingto obtain I/Q. It also provides two trigger connections configurable for input or output.SSBC Card- Single Board Computer with dual SMP processors (PC) running Linux.The figure above shows a basic magnetron system constructed with an IFD, and two PCI cards.A standard RS-232 serial input (included with the SBC) is used for obtaining the antenna anglesand the output/input trigger is provided directly from the Rx card. This system has 5 times theprocessing power of the fastest version of the previous generation processor (RVP7/Main boardplus 2 RVP7/AUX boards) so that it is capable of performing DFT processing in 2048 rangebinswith advanced algorithms such as random phase 2nd trip echo filtering and recovery.Example 2: Klystron System with Digital TxIn this case, the IFD can receive a master clock from the radar system (e.g., the COHO). Thisensures that the entire system is phase locked. As compared to the previous example there aretwo additional cards shown in this example:
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–6RVP8 Configuration Example: High Performance KlystronKeyboardMouseMonitor10/100/1000 Base TIFDFiber DownlinkCOAX UplinkTriggersIF SignalReference ClockDigitally Synthesized COHOParallel or Synchro AZParallell or Synchro ELPulse width IF Tx WaveformUtilitiesIF Tx Waveform14-BitI/OĆ62SBCRVP8/TxRVP8/RxConnector PanelSRVP8/Tx- The digital transmitter card provides the digital Tx waveform. A second output can beused to provide a COHO in the event that the RVP8 is used to provide the system master clock.In any case, the IF transit waveform and the A/D sampling are phase locked.SSIGMET I/O-62 card for additional triggers, parallel, synchro or encoder AZ and EL angleinputs, pulse width control, spot blanking control output, etc. These signals are brought in via theconnector panel.The figure shows the SIGMET SoftPlanet which carries time-critical I/O such as clock andtrigger information which is not appropriate for the PCI bus. These signals are limited to thecards provided by SIGMET, i.e., the SoftPlanet is not connected to any of the standardcommercial cards.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–7IFDFiber DownlinkCOAX UplinkRVP8 Configuration Example: Dual Polarization Magnetron SystemKeyboardMouseMonitor10/100/1000 BaseT LANIFDHorizontal IF SignalDAFCDigital STALOOptionalVertical IF SignalUtilities14-Bit14-BitIF Magnetron BurstSynch ClockFiber DownlinkHorzVert I/OĆ62SBCRVP8/RxRVP8/RxCOAX UplinkTriggersParallel or Synchro AZParallell or Synchro ELPulse Width ControlConnector PanelPolarization ControlExample 3: Dual Polarization Magnetron SystemIn this system 2 IFD’s and two RVP8/Rx cards are used for the horizontal and vertical channelsof a dual-channel receiver. The legacy RVP7 technique of using a single IFD and two IFfrequencies for the horizontal and vertical channels (e.g., 24 and 30 MHz) is also supported bythe RVP8. In the case of either dual or single IFD’s, there is a synch clock provided by either theSTALO reference frequency (e.g., 10 MHz) or by the RVP8 itself.The RVP8 supports calculation of the complete covariance matrix for dual pol, including ZDR,PHIDP (KDP), RHOHV, LDR, etc. Which of these variables is available depends on whether thesystem is a single–channel switching system (alternate H and V), a STAR system (simultaneoustransmit and receive) or a dual channel switching system (co and cross receivers). Note that forthe special case of a single channel switching system, only one IFD is required.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–8COTS AccessoriesAside from the basic PCI cards required for the radar application, there are additional cards thatcan be installed to meet different customer requirements, e.g.,S10/100–BaseT Ethernet card for additional network I/O (e.g., a backup network).SRS-232/RS-422 serial cards for serial angles, remote TTY control, etc.SSound card to synthesize audio waveforms for wind profiler applications.SGPS card for time synch.SIEEE 488 GPIB card for control of test equipment.The bottom line is that the PCI open hardware approach provides unparalleled hardwareflexibility. In addition, the availability of compatible low-cost replacement or upgrade parts isassured for years into the future.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–91.1.1 IFD IF DigitizerThe IFD 14-bit IF digitizer is a totally sealed unit for optimum low-noiseperformance. The use of digital components within the IFD is minimizedand the unit is carefully grounded and shielded to make the cleanestpossible digital capture of the input IF signal. Because of this, the IFDachieves the theoretical minimum noise level for the A/D convertors.There are 4 inputs to the IFD:SIF video signal.SA secondary IF video signal, used for dual polarization or verywide dynamic range applications.SIF Burst Pulse for magnetron or IF COHO for Klystron.SOptional reference clock for system synchronization. For aKlystron system, the COHO can be input. Magnetron systems donot require this signal. This clock can even come from theRVP8/Tx card itself.All of these inputs are on SMA connectors. The IF signal input is madeimmediately after the STALO mixing/sideband filtering step of thereceiver where a traditional log receiver would normally be installed.The required signal level for both the IF signal and burst is +6.5 dBm forthe strongest expected input signal. A fixed attenuator or IF amplifiermay be used to adjust the signal level to be in this range.Digitizing is performed for both the IF signal and burst/COHO channelsat approximately 72 MHz to 14-bits. This provides 92 to 105 dB ofdynamic range (depending on pulse width) without using complex AGC,dual A/D ranging or down mixing to a lower IF frequency.All communication to the main RVP8 chassis goes over a special CAT5Etype cable. The major volume of data is the raw time series samples sentdown to the RVP8 Rx card. Coming back up is trigger timing and AFCinformation to the IFD.The RVP8 provides comprehensive AFC support for tuning the STALO of a magnetron system.Alternatively, the magnetron itself can be tuned by a motorized tuning circuit controlled by theRVP8. Both analog (+–10V) and digital tuning (with optional DAFC to 24 bits) are supported.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–101.1.2 Digital Receiver PCI Card (RVP8/Rx)The RVP8/Rx card receives the digitized IF samples from the IFD via the fiber opticlink. The advantage of this design is that the receiver electronics (LNA, RF mixer,IF preamp, and IFD) can be located as far as 100–meters away from the RVP8 mainchassis. This makes it possible to choose optimum locations for both the IFD and theRVP8, e.g., the IFD could be mounted on the antenna itself, and the processor boxin a nearby equipment room.The RVP8/Rx is 100% compatible with the 14-bit RVP7/IFD, but it also includeshooks for future IFD’s operating at higher sampling clock rates. Two additional BNCconnectors are included on the board’s faceplate. These can be used for trigger input,programmable trigger output, or a simple LOG analog ascope waveform.A remarkable amount of computing power is resident on the receiver board, in theform of an FIR filter array that can execute 6.9 billion multiply/accumulate cyclesper second. These chips serve as the first stage of processing of the raw IF data sam-ples. Their job is to perform the down–conversion, bandpass, and deconvolutionsteps that are required to produce (I,Q) time series. The time series data are then trans-ferred over the PCI bus to the SBC for final processing.The FIR filter array can buffer as much as 80 microsec of 36MHz IF samples, and then computea pair of 2880–point dot products on those data every 0.83 microsec. This could be used toproduce over-sampled (I,Q) time series having a range resolution of 125–meters and abandwidth as narrow as 30Khz. The same computation could also yield independent 125–metertime series data from an 80 microsec compressed pulse whose transmit bandwidth wasapproximately 1MHz.Finer range resolutions are also possible, down to a minimum of 25–meters. A special feature ofthe RVP8/Rx is that the bin spacing of the (I,Q) data can be set to any desired value between 25and 2000 meters. Range bins are placed accurately to within +2.2 meters of any selected grid,which does not have to be an integer multiple of the sampling clock. However, when an integermultiple (N x 8.333–meters) is selected, the error in bin placement effectively drops to zero.Dual polarization radars that are capable of simultaneous reception for both horizontal andvertical channels can be interfaced to the RVP8 using a separate RVP8/Rx and IFD for eachchannel. Note that the multiplexed dual IF approach used for the RVP7 with a single IFD canalso be used.One of the primary advantages of the digital receiver approach is that wide linear dynamic rangecan be achieved without the need for complex AGC circuits that require both phase andamplitude calibration.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–11Calibration Plot for RVP8/IFDThe figure above shows a calibration plot for a 14-bit IFD with the digital filter matched to a 2microsecond pulse. The performance in this case is >100 dB dynamic range.The RVP8 performs several real time signal corrections to the I/Q samples from the Rx,including:Amplitude Correction- A running average of the transmit pulse power in the magnetron burstchannel is computed in real-time by the RVP8/Rx. The individual received I/Q samples arecorrected for pulse–to–pulse deviations from this average. This can substantially improve the“phase stability” of a magnetron system to improve the clutter cancelation performance to nearKlystron levels.Phase Correction- The phase of the transmit waveform is measured for each pulse (either theburst pulse for magnetron systems or the Tx Waveform for coherent systems). The I/Q valuesare adjusted for the actual measured phase. The coherency achievable is better than 0.1 degreesby this technique.Large Signal Linearization- When an IF signal saturates, there is still considerable informationin the signal since only the peaks are clipped. The proprietary large signal linearizationalgorithm used in the RVP8 provides an extra 3 to 4 dB of dynamic range by accounting for theeffects of saturation.The RVP8/Rx card provides the same comprehensive configuration and test utilities as theRVP7, with the difference that no external host computer is required to run the utilities. Theseutilities can be run either locally or remotely, over the network! Some examples are shownbelow:
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–12Digital IF Band Pass Design ToolThe built–in filter design tool makes it easy foranyone to design the optimal IF filter to matcheach pulse width and application. Simply specifythe impulse response and pass band and the filterappears. The user interface makes it easy to wid-en/narrow the filter with simple keyboard com-mands. There is even a command to automatical-ly search for an optimal filter. This display can also show the actual spectrumof the transmit burst pulse for quality control andcomparison with the filter.Burst Pulse Alignment ToolThe quality assessment of the transmit burstpulse and its precise alignment at range zeroare easy to do, either manually using this tooland/or automatically using the burst pulseauto-track feature. This performs a 2D searchin both time and frequency space if a validburst pulse is not detected. The automatictracking makes the AFC robust to start–uptemperature changes and pulse width changesthat can effect the magnetron frequency.AFC alignment/check is now much easiersince it can be done manually from a centralmaintenance site or fully automatically.Received Signal Spectrum Analysis ToolThe RVP8 provides plots of the IF signal versusrange as well as spectrum analysis of the signalas shown in this example.In the past, these types of displays and tools re-quired that a highly-skilled engineer transportsome very expensive test equipment to the radarsite. Now, detailed analysis and configurationcan all be done from a central maintenance facil-ity via the network. For a multi-radar networkthis results in substantial savings in equipment,time and labor.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–131.1.3 Mother Board or Single-Board Computer (SBC)The dual-CPU Pentium mother board or single-board computer (SBC) acts as the hostto the Linux operating system and provides all of the compute resources for processingthe I/Q values that are generated by the RVP8/Rx card. Standard keyboard, mouse andmonitor connections are on the Rx backpanel, along with a 10/100/1000 BaseT Ether-net port. The system does not require that a keyboard, mouse or monitor be connectedwhich is typically the case at an unattended site. An SBC example is shown on the left.Motherboards and SBC’s are available from many vendors, at various speeds Typical-ly the SBC is equipped with 128 MB RAM. The RVP8 chassis has a front bay for eithera >20 GB hard disk or a Flash Disk. The Flash Disk approach is well suited to applica-tions where high–reliability is important. CDRW is also provided for software mainte-nance. Note that the latest versions of the RVP8 software and documentation can al-ways be down-loaded from SIGMET’s web site for FREE.The SBC also plays host for SIGMET’s RVP8 Utilities which provide test, configura-tion, control and monitoring software as well as built–in on-line documentation.1.1.4 Digital Transmitter PCI Card (RVP8/Tx) Many of the exciting new meteorological applications for the RVP8 are made possibleby its ability to function as a digital radar transmitter. The RVP8/Tx PCI card synthe-sizes an output waveform that is centered at at the radar’s intermediate frequency. Thissignal is filtered using analog components, then up–converted to RF, and finally am-plified for transmission. The actual transmitter can be a solid state or vacuum tube de-vice. The RVP8 can even correct for waveform distortion by adaptively “pre–distort-ing” the transmit waveform, based on the measured transmit burst sample.The Tx card has a BNC output for the IF Tx waveform. In addition, there is a secondoutput for an auxiliary signal or clock, or for a clock input. At the bottom of the cardis a 9–pin connector for arbitrary I/O (e.g., TTL, RS422, additional clock).The RVP8 digital transmitter finds a place within the overall radar system that exactlycomplements the digital receiver. The receiver samples an IF waveform that has beendown–converted from RF, and the transmitter synthesizes an IF waveform for up–con-version to RF. The beauty of this approach is that the RVP8 now has complete controlover both halves of the radar, making possible a whole new realm of matched Tx/Rxprocessing algorithms. Some examples are given below:SPhase Modulation- Some radar processing algorithms rely on modulating the phase of thetransmitter from pulse to pulse. This is traditionally done using an external IF phase modulatorthat is operated by digital control lines. While this usually works well, it requires additionalhardware and cabling within the radar cabinet, and the phase/amplitude characteristics may not
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–14be precise or repeatable. In contrast, the RVP8/Tx can perform precise phase modulation to anydesired angle, without requiring the use of external phase shifting hardware.SPulse Compression- There is increasing demand for siting radars in urban areas that alsohappen to have strict regulations on transmit emissions. Often the peak transmit power islimited in these areas; so the job for the weather radar is to somehow illuminate itstargets using longer pulses at lower power. The problem, of course, is that a simple longpulse lacks the ability (bandwidth) to discern targets in range. The remedy is to increasethe Tx bandwidth by modulating the overall pulse envelope, so that a reasonable rangeresolution is restored. The exceptional fidelity of the RVP8/Tx waveform can accomplishthis without introducing any of the spurious modulation components that often occurwhen external phase modulation hardware is used.SFrequency Agility- This has been well studied within the research community, but hasremained out of the reach of practical weather radars. The RVP8/Tx changes all of this,because frequency agility is as simple as changing the center frequency of thesynthesized IF waveform. Many new Range/Doppler unfolding algorithms becomepossible when multiple transmit frequencies can coexist. Frequency agility can also becombined with pulse compression to remedy the blind spot at close ranges while the longpulse is being transmitted.SCOHO synthesis- The RVP8/Tx output waveform can be programmed to be a simpleCW sine wave. It can be synthesized at any desired frequency and amplitude, and itsphase is locked to the other system clocks. If you need a dedicated oscillator at somerandom frequency in the IF band, this is a simple way to get it.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–151.1.5 I/O-62 PCI Card and I/O PanelThe SIGMET I/O-62 is a short format PCI card that provides extensive I/O capabilitiesfor the RVP8. A typical installation would have one I/O-62 and an RVP8 ConnectorPanel shown above. The Softplanet is used to interconnect the I/O 62 with other SIG-MET PCI cards. Note that the identical card is used in the SIGMET RCP8 radar/anten-na control processor which in general does not use the Softplanet connection. TheI/O-62 has a single 62-position, high-density “D” connector. This is attached to theRVP8 Connector Panel (typically mounted on the front or back of the rack which holdsthe RVP8). A standard 1:1 cable connects the remote panel to the I/O-62 card in theRCP8 chassis. The standard connector panel provided by SIGMET meets the needs ofmost radar sites.The best part is that the I/O-62 is configurable in software, i.e., there is no need to openthe chassis to configure jumpers or switches. This means that when a spare board isadded, there is no need to perform hardware configuration or custom wiring.The physical I/O lines are summarized in the system specifications section.ESD Protection FeaturesSince the I/O lines are connected to the radar system, there is a potential for lightning or otherESD type damage. This is addressed aggressively by the I/O-62 in two ways:SEvery wire is protected by a Tranzorbt diode which transitions from an open to a fullclamp between ±27 to ±35 VDC. Additionally, the Connector Panel uses Tranzorbtdiodes on every I/O line for double protection.SHigh-voltage tolerant front-end receivers/drivers are used. All components connected tothe external pins can tolerate up to ±40V. For example, the TTL and wide range inputsuse protectors that normally look like 100 Ohm resistors, but open at high voltage.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–16Run Time FPGA ConfigurationThe SIGMET I/O-62 card is built around a 100K–Gate FPGA which, in addition to driving theI/O signals on the 62-position connector, also coordinates the PCI and Softplanet traffic. Thesechips are SRAM–based, meaning that they are configured at run time. This allows the FPGAcode to be automatically upgraded during each RVP8 code release without needing to physicallyreprogram any parts.The board’s basic I/O services use up only 40% of the complete FPGA. The leftover spacemakes it possible to add smart processing right on the I/O-62 board to handle custom needs. Forexample the 16–bit floating–point (I,Q) data in the previous example could be reformatted into a32–bit fixed–point stream. Other examples include generating custom serial formats, datadebouncing, and signal transition detection. In general, I/O functions that would either betedious or inappropriate for the host computer SBC can likely be moved onto the I/O-62 carditself.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–171.2 Comparison of Analog vs Digital Radar Receivers1.2.1 What is a Digital IF Receiver?A digital IF receiver accepts the analog IF signal (typically 30 MHz), processes it and outputs astream of wide dynamic range digital “I” and “Q” values. These quantities are then processed toobtain the moment data (e.g., Z, V, W or polarization variables). Additionally, the digitalreceiver can accept the transmit pulse “burst sample” for the purpose of measuring thefrequency, phase and power of the transmit pulse. The functions that can be performed by thedigital receiver are:SIF band pass filteringS“I” and “Q” calculation over wide dynamic rangeSPhase measurement and correction of transmitted pulse for magnetron systems – fromburst sampleSAmplitude measurement and correction of transmitted pulse – from burst sampleSFrequency measurement for AFC output – from burst sampleThe digital approach replaces virtually all of the traditional IF receiver components with flexiblesoftware-controlled modules that can be easily adapted to function for a wide variety of radarsand operational requirements.The digital receiver approach made a very rapid entry into the weather radar market. Up untilthe about 1997 weather radars were not supplied with digital receivers. Today in 2003 nearly allnew weather radars and weather radar upgrades use the digital receiver approach. Much of thisrapid change is attributed to the previous generation RVP7 which is the most widely soldweather radar signal processor of all time.The number one advantage of a digital receiver is that it achieves a wide linear dynamic range(e.g., >95dB depending on pulse width) without having to use AGC circuits which are complexto build, calibrate and maintain. However, there are other advantages as well:SLower initial cost by eliminating virtually all IF receiver components.SLower life cycle cost do to reduced maintenance.SSelectable IF frequency.SSoftware controlled AFC with automatic alignment.SProgrammable band pass filterSDual or multiple IF multiplexingSImproved remote monitoring down to the IF level.The following sections compare the digital receiver approach to the analog receiver approach.This illustrates the advantages of the digital approach and what functions are performed by adigital receiver.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–181.2.2 Magnetron Receiver ExampleA typical analog receiver for a magnetron system is shown in the top portion of Figure 1–1. Thereceived RF signal from the LNA is first mixed with the STALO (RF–IF) and the resulting IFsignal is applied to one of several bandpass filters that match the width of the transmitted pulse.The filter selection is usually done with relays. The narrow band waveform is then split. Half isapplied to a LOG amplifier having a dynamic range of 80–100dB, from which a calibratedmeasurement of signal power can be obtained. The LOG amplifier is required because it isalmost impossible to build a linear amplifier with the required dynamic range. However, phasedistortion within the LOG amplifier renders it unsuitable for making Doppler measurements;hence, a separate linear channel is still required.The linear amplifier is fed from the other half of the bandpass filter split. It may be preceded bya gain control circuit (IAGC) which adjusts the instantaneous signal strength to fall within thelimited dynamic range of the linear amplifier. The amplitude and phase characteristics of theIAGC attenuator must be calibrated so that the “I” and “Q” samples may be corrected duringprocessing.The IF output from the linear amplifier is applied to a pair of mixers that produce “I” and “Q”.The mixer pair must have very symmetric phase and gain characteristics, and each must besupplied with an accurate 0-degree and 90-degree version of the Coherent Local Oscillator(COHO). The later is usually obtained by sampling a portion of the transmitted pulse, and thenphase locking an oscillator (COHO) that continues to “ring” afterward. Phase locked COHO’sof this sort can be very troublesome – they often fail to lock properly, drift with age, and fail tomaintain coherence over the full unambiguous range.The transmit burst that locks the COHO is also used by the Automatic Frequency Control (AFC)loop. The AFC relies on an FM discriminator and low pass filter to produce a correction voltagethat maintains a constant difference between the magnetron frequency and the reference STALOfrequency. The AFC circuit is often troublesome to set and maintain. Also, since it operatescontinuously, small phase errors are continually being introduced within each coherentprocessing interval.In contrast, the RVP8 digital receiver is shown in to lower portion of Figure 1–1. The only oldparts that still remain are the microwave STALO oscillator, and the mixer that produces thetransmit burst. The burst pulse and the analog IF waveform are cabled directly into the IFD onSMA coax cables. Likewise, the AFC control voltage is also a simple direct connection eitherwith analog tuning (+–10V from IFD) or digital control via the optional DAFC interface. Thesecables constitute the complete interface to the radar’s internal signals; no other connections arerequired within the receiver cabinet.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–19Figure 1–1: Analog vs Digital Receiver for Magnetron SystemsClassic Analog Receiver for MagnetronAnalog IF IFDigital AttenControl BitsIAGCSplitLOGLinear AmpQuad PhaseDetectorCOHORVP8 Magnetron InterfaceAnalog IFFrom IAGC LogicIFDigitizer Fiber OpticRF Tx BurstSplitXSplitAFCIF Tx SampleIF Tx SampleXDigital AFC ControlCOAXUplinkAFC SignalIQLOGLine DriversLow Q Locking COHOIF Filters Matched toPulse WidthsBPFBPFBPFIFDDownlinkDAFC+–10V Analog AFCXXAnalog RFFrom LNABPFAnalog RFFrom LNASplit24 bitsPhaseLockedRVP8/RxOptionalSTALORF Tx BurstSTALO
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–201.2.3 Klystron or TWT Receiver and Transmit RF ExampleA typical analog receiver for a klystron system is shown in the top portion of Figure 1–2. Thearrangement of components is similar to the magnetron case, except that the COHO operates at afixed phase and frequency, a phase shifter is included for 2nd trip echo filtering and there is noAFC feedback required. The phase stability of a Klystron system is better than a magnetron, butthe system is still constrained by limited linear dynamic range, IAGC inaccuracy, quad phasedetector asymmetries, phase shifter inaccuracies, etc.The RVP8/Tx card now plays the role of a programmable COHO. The digitally synthesizedtransmit waveform can be phase, frequency and amplitude modulated (no separate phase shifteris required) and even produce multiple simultaneous transmit frequencies. These capabilities areused to support advanced algorithms, e.g., range/velocity ambiguity resolution or pulsecompression for low power TWT systems.Figure 1–2: Analog vs Digital Receiver for Klystron SystemsClassic Receiver and Transmit RF for KlystronCOHORVP8 Klystron InterfaceTransmit RFXTo KlystronAnalog IF IFDigital AttenControl BitsIAGCSplitLOGLinear AmpQuad PhaseDetectorFrom IAGC LogicSplitIQLOGLine DriversBPFBPFBPFXAnalog RFFrom LNABPFAnalog IFIFDigitizer Fiber OpticIF Tx SampleRF Tx SampleXDigital Synthesized IF (smart COHO)COAXUplinkIFDDownlinkXAnalog RFFrom LNASplitXTransmit RFTo KlystronOptional CLKRVP8/TxRVP8/RxΔφPhase ShifterControl BitsSTALOSTALO
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–211.3 RVP8 IF Signal Processing1.3.1 IFD Data Capture and TimingThe RVP8 design concept is to perform very little signal processing within the IFD digitizermodule itself. This is to minimize the presence of digital components that might interfere withthe clean capture of the IF signals.The digitized IF and burst pulse samples are multiplexed onto the fiber channel link whichprovides the digital data to the RVP8/Main board at approximately 540-MBits/sec. The 14-bitsamples are encoded for transmission over a fiber channel link. This optical link allows the IFDto be as far as 100 meters away from the RVP8/Main board and provides an added degree ofnoise immunity and isolation.The uplink input from the RVP8/Main board provides the timing for multiplexing the burst pulsesample with the IF signal. In addition, it is used to set the AFC DAC or digital output level, andto perform self tests.The sample clock oscillator in the IFD is selected to be very stable. The sample clock serves asimilar function to the COHO on a traditional Klystron system, i.e., it is the master time keeper.Because of this the IFD sample clock is used to phase lock the entire RVP8, i.e., the Rx, Tx,IO-62 boards and the SoftPlane are all phase locked to the IFD sample clock. Designers havetwo choices for factory configuration of the IFD sample clock:SA fixed crystal frequency selected to achieve a desired range resolution. The standardrange resolution corresponds to 25 m increments.SA very narrow band VCXO (50 ppm) selected to lock to an input reference signal fromthe radar, and provide a desired range resolution. SIGMET stocks VCXO’s for 25 mrange resolution increments for reference inputs of 10, 20, 30 and 60 MHz. Customfrequency VCXO’s are available on request. Examples of external reference signalsources are an external COHO, external STALO reference or perhaps even a GPS clock).
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–221.3.2 Burst Pulse Analysis for Amplitude/Frequency/PhaseThe burst pulse analysis provides the ampli-tude, frequency and phase of the transmittedpulse. The phase measurement is analogous tothe COHO locking that is performed by a tradi-tional magnetron radar. The difference is thatthe phase is known in the digital technique, sothat range dealiasing using the phase modula-tion techniques is possible. Amplitude mea-surement (not performed by traditional radars)can provide enhanced performance by allow-ing the “I” and “Q” values to be corrected forvariations in the both the average and the pulse-to-pulse transmitted power. In addition, awarning is issued if the burst pulse amplitudefalls below a threshold value.The burst pulse data stream is first analyzed by an adaptive algorithm to locate the burst pulsepower envelope (e.g. 0.8 msec). The algorithm first does a coarse search for the burst pulse in thetime/frequency domain (by scanning the AFC) and then does a fine search in both time andfrequency, to assure that the burst is centered at “range 0” and is at the required IF value. Thepower-weighted phase of the burst pulse and the total burst pulse power is then computed. Thepower weighted average phase is used to make the digital phase correction. Phase jitter formagnetron systems with good quality modulator and STALO is better than 0.5 degrees RMS, asmeasured on actual nearby clutter targets. For Klystron systems, the phase locking is better than0.1 degree RMS.The burst pulse frequency is also analyzed to calculate the frequency error from the nominal IFfrequency. For magnetron systems, the error is filtered with a selectable time constant which istypically set to several minutes to compensate for slow drift of the magnetron. The digitalfrequency error is sent via the uplink to the IFD in the receiver cabinet where a DAC converts itinto an analog output to the magnetron STALO. Optionally, a DAFC unit can be Teed off theuplink cable to interface to Klystron systems do not require the AFC.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–231.3.3 Rx Board and CPU IF to I/Q ProcessingFigure 1–3: IF to I/Q Processing Steps Fiber OpticReceiverDigital IFIF Tx SamplesBandpass FilterDigital Quad Phaseand DecompressionTx Burst PulseAnalysisFrequencyAmplitude/Phase CorrectionInterferenceFilterAFCServoI/QSignal I/Q TxBurstIFD Fiber OpticDownlinkIFD COAXUplinkDigitized IF Signaland Tx Burst SampleTiming and digital AFCIF to I/Q Processing Steps1000 BaseTEthernetUDP Broadcast I/QSamples to recordingsystem or separateprocessing systemDigital FIRRVP8/RxCPUPCII/QDigital AFCI/Q MomentProcessingIF TxSamplesI/QTx BurstI/QSignalI/Q TimeSeriesAPIThe RVP8/Rx board performs the initial processing of the IF digital data stream and outputs “I”and “Q” data values to the host computer via the PCI bus. In addition, the frequency, phase andamplitude of the burst pulse are measured. The functions performed by the processor are:SReception of the digital serial fiber optic data stream.SBand pass filtering of the IF signal using configurable digital FIR filter matched to thepulsewidth.SRange gating and optional coherent averaging (essentially performed during the bandpass filtering step).
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–24SComputation of “I” and “Q” quadrature values (also performed during the band passfiltering step).STransmit burst sample frequency, phase and amplitude calculationSI and Q phase and amplitude correction based on transmit burst sample.SInterference rejection algorithm.SAFC frequency error calculation with output to IFD for digital or analog control ofSTALO (for magnetron systems).The advantage of the digital approach is that the software algorithms for these functions can beeasily changed. Configuration information (e.g., processor major mode, PRF, pulsewidth, gatespacing, etc.) is supplied from the host computer.The digital matched filter that computes “I”and “Q” is designed in an interactive man-ner using a TTY and oscilloscope for graph-ical display. The filter’s passband widthand impulse response length are chosen bythe user, and the RVP8 constructs the filtercoefficients using built-in design software.The frequency response of the filter can bedisplayed and compared to the frequencycontent of the actual transmitted pulse.Microwave energy can come from a variety of transmitters such as ground-based, ship-based orairborne radars as well as communications links. These can cause substantial interference to aweather radar system. Interference rejection is provided as standard in the RVP8. Three differentinterference rejection algorithms are supported.The RVP8/Rx board places the wide dynamic range “I” and “Q” samples directly on the PCIbus where they are sent to the processor section of the PC (e.g., dual Pentium processors on asingle-board computer or motherboard). The I/Q values are then processed on the Pentiumprocessors to extract the moment information (Z, V, W and optional polarization parameters).The I and Q values can also be placed on a gigabit Ethernet line (1000 BaseT) which is provideddirectly on the processor board. This means that there is no second PCI bus “hit” required tosend the data to a recording system or a completely separate processing system.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–251.4 RVP8 Weather Signal ProcessingThe processing of weather signals by the RVP8 is based on the algorithms used in the previousgeneration RVP7 and RVP6. However, the performance of the RVP8 allows a different approachto some of the processing algorithms, especially the frequency domain spectrum processing. Allof the algorithms start with the wide dynamic range I and Q samples that are obtained from theRx card over the PCI bus.The resulting intensity, radial velocity, spectrum width and polarization measurements are thensent to a separate host computer to serve as input for applications such as:SQuantitative Rainfall MeasurementSVertical Wind ProfilingSZDR Hail DetectionSTornado Detection and Microburst DetectionSGust Front DetectionSParticle IdentificationSTarget Detection and TrackingSGeneral Weather MonitoringTo obtain the basic moments, the RVP8 offers the option of several major processing modes:SPulse Pair Mode Time Domain ProcessingSDFT/FFT Mode Frequency Domain ProcessingSRandom Phase Mode for 2nd trip echo filteringSPolarization Mode ProcessingNote that the RVP8 is the first commercial processor to perform discrete Fourier transforms(DFT) as well as fast Fourier transforms (FFT). FFT is more computationally efficient thanDFT, but the sample size is limited to be a power of two (16, 32, 64, ...) This is too restrictive onthe scan strategy for a modern Doppler radar since this means, for example, that a one degreeazimuth radial must be constructed from say exactly 64 input I/Q values. The RVP8 has theprocessing power such that when the sample size is not a power of 2, a DFT is performed insteadof an FFTThese modes share some common features that are described first, followed by descriptions ofthe unique features of each mode.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–261.4.1 General Processing featuresFigure 1–4 shows a block diagram of the processing steps. These are discussed below.AutocorrelationsThe autocorrelations R0, R1 and R2 are produced by all three processing modes. However, theway that they are produced is different for the three modes, particularly with regard to thefiltering that is performed.SPulse Pair Mode — Filtering for clutter is performed in the time domain.Autocorrelations are computed in the time domain.SDFT/FFT Mode — Filtering for clutter is performed in the frequency domain using bothfixed width filters and the Gaussian Model Adaptive Processing (GMAP) technique.Autocorrelations are computed from the inverse transform.SRandom Phase — Filtering for clutter and second trip echo is performed in the frequencydomain by adaptive algorithms. Autocorrelations are computed from the inversetransform.Figure 1–4: I/Q Processing for Weather Moment Extraction Moment andThresholdCalculationsClutter FilterIQAutocorrelationsRVP8 Standard Moment Processing StepsClutter MicroSuppression andRange AveragingThresholdingSpeckle FilterTZVWSQILOGSIGCCORT0R0R1R2T0R0R1R2TZVWPulse Pair, FFT/DFT, Random Phase Modes10/100/1000 BaseT EthernetTo Applications Host ComputerTime SeriesAPIThe use of the R2 lag provides improved estimation of signal-to-noise ratio and spectrum width.Processors that do not use R2 cannot effectively measure the SNR and spectrum width.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–27Time (azimuth) AveragingThe autocorrelations are based on input “I” and “Q” values over a selectable number of pulsesbetween 8, 9, 10, ...,256. Any integer number of pulses in this interval may be used includingDFT/FFT and random phase modes.Selectable angle synchronization using the input AZ and EL tag lines assures that all possiblepulses are used during averaging for each, say, 1 degree interval. This minimizes the number of“wasted” pulses for maximum sensitivity. Azimuth angle synchronization also assures theaccurate vertical alignment of radial data from different elevation angles in a volume scan (seebelow).TAG Angle Samples of Azimuth and Elevation During data acquisition and processing it is usually necessary to associate each output ray withan antenna position. To make this task simpler the RVP8 samples 32 digital input “TAG” lines,once at the beginning and once at the end of each data acquisition period. These samples areoutput in a four-word header of each processed ray. When connected to antenna azimuth andelevation, the TAG samples provide starting and ending angles for the ray, from which themidpoint could easily be deduced. Since the bits are merely passed on to the user, any anglecoding scheme may be used. The processor also supports an angle synchronization mode, inwhich data rays are automatically aligned with a user-defined table of positions. For thatapplication, angles may be input either in binary or BCD.Range Averaging and Clutter Microsuppression To improve the accuracy of the reflectivity measurements, the RVP8 can perform rangeaveraging. When this is done, autocorrelations from consecutive range bins are averaged, andthe result is treated as if it were a single bin. This type of averaging is useful to lower thenumber of range bins that the host computer must process.Range averaging of the autocorrelations may be performed over 2, 3, 4, ..., 16 bins. Prior torange averaging, any bins that exceed the selectable clutter-to-signal threshold are discarded.This prevents isolated strong clutter targets from corrupting the range average, which improvesthe sub-clutter visibility.Moment ExtractionThe autocorrelations serve as the basis for the Doppler moment calculations,SMean velocity – from Arg [ R1 ]SSpectrum width – from |R1| and |R2| assuming Gaussian spectrumSdBZ – from R0 with correction for ground clutter, system noise and gaseous attenuation.Uses calibration information supplied by host computer.SdBT – identical to dBZ except without ground clutter.These are the standard parameters that are output to the host computer on the high-speedEthernet interface.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–28ThresholdingThe RVP8 calculates several parameters that are used to threshold (discard) bins with weak orcorrupted signals. The thresholding parameters are:SSignal quality index (SQI=|R1|/R0)SLOG (or incoherent) signal-to-noise ratio (LOG)SSIG (coherent) signal-to-noise ratioSCCOR clutter correctionThese parameters are computed for each range bin and can be applied in AND/OR logicalexpressions independently for dBZ, V and W.Speckle FilterThe speckle filter can be selected to remove isolated single bins of either velocity/width orintensity. This feature eliminates single pixel speckles which allows the thresholds to be reducedfor greater sensitivity with fewer false alarms (speckles). Both a 1D (single azimuth ray) and 2D(3 azimuth rays by 3 range bins) are supported.Velocity UnfoldingA special feature of the RVP8 processor is its ability to “unfold” mean velocity measurementsbased on a dual PRF algorithm. In this technique two different radar PRF’s are used foralternate N-pulse processing intervals. The internal trigger generator automatically produces thecorrect dual-PRF trigger, but an external trigger can also be applied. In the later case, theENDRAY_ output line provides the indication of when to switch rates. The RVP8 measures thePRF to determine which rate (high or low) was present on a given processing interval, and thenunfolds based on either a 2:3, 3:4 or 4:5 frequency ratio. Table 1–1 gives typical unambiguousvelocity intervals for a variety of radar wavelengths and PRF’s.Table 1–1: Examples of Dual PRF Velocity UnfoldingUnambiguous Velocity (m/s) forVarious Radar WavelengthsPRF1 PRF2 UnambiguousRange (km) 3 cm 5 cm 10 cm500 * 300 3.75 6.25 12.50 NoU f ldi1000 * 150 7.50 12.50 25.00 Unfolding2000 * 75 15.00 25.00 50.00500 333 300 7.50 12.50 25.00 TwoTi1000 667 150 15.00 25.00 50.00 TimesUnfolding2000 1333 75 30.00 50.00 100.00Unfolding
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–29PRF1 10 cm5 cm3 cmUnambiguousRange (km)PRF2500 375 300 11.25 18.75 37.50 Three TimesU f ldi1000 750 150 22.50 37.50 75.00 Unfolding2000 1500 75 45.00 75.00 150.00500 400 300 15.00 25.00 50.00 FourTi1000 800 150 30.00 15.00 100.00 TimesUnfolding2000 1600 75 60.00 100.00 200.00Unfolding1.4.2 RVP8 Pulse Pair Time Domain ProcessingPulse pair processing is done by direct calculation of the autocorrelation. Prior to pulse pairprocessing, the input “I” and “Q” values are filtered for clutter using a a time domain notchfilter. Filters of various selectable widths are available for either 40 or 50 dB stop bandattenuation. The filtered I/Q values are processed to obtain the autocorrelation lags R0, R1 andR2. The unfiltered power is also calculated (T0). The autocorrelations are then sent to the rangeaveraging and moment extraction steps.1.4.3 RVP8 DFT/FFT Processing The DFT/FFT mode allows clutter cancelation to be performed in the frequency domain. DFT isused in general, with FFT’s used if the requested sample size is a power of 2.Three standard windows are supported to provide the best match of window width to thespectrum dynamic range:SRectangularSHammingSBlackmanSExact BlackmanSVon HanAfter the FFT step, clutter cancelation is done using a selectable fixed width filter thatinterpolates across the noise or any overlapped weather or an adaptive filter which automaticallydetermines the optimal width. This technique preserves overlapped weather as compared to timedomain notch filters which will always attenuate overlapped weather to some extent, dependingon the spectrum width. After clutter cancelation, R0, R1 and R2 are computed by inversetransform and these are used for moment estimation.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–301.4.4 Random Phase Processing for 2nd Trip EchoSecond trip echoes can be a serious problem for applications that require operation at a highPRF. Second trip echoes can appear separately or can be overlaid on first trip echoes (second tripobscuration). The random phase technique separates the first and second trip echoes so that:SIn nearly all cases, the 2nd trip echo can be removed from the first trip even in the caseof overlapped 1st and 2nd trip echoes. The benefit is a clean first trip display.SThe 2nd trip echoes can be recovered and placed at their proper range at 1st trip/2nd tripsignal ratios of up to 40 dB difference for overlapped echoes. Because of the widedynamic range of weather echoes, this power limit will sometimes be exceeded.The technique requires that the phase of each pulse be random. Digital phase correction is thenapplied in the processor for the first and second trips. The critical step is the adaptive filterwhich removes the echo of the other trip to increase the SNR. Magnetrons have a naturallyrandom phase. For Klystron radars, a digitally controlled precision IF phase shifter is required.The RVP8 provides an 8-bit RS422 output for the phase shifter.For more information on the technique refer to Joe, et. al., 1995.1.4.5 Polarization Mode ProcessingPolarization processing uses a time domain autocorrelation approach to calculate the variousparameters of the polarization co-variance matrix, i.e., ZDR, LDR, PHIDP, RHOHV, PHIDP(KDP), etc. In addition, the standard moments T, V, Z, W are also calculated. Which parametersare available and which algorithms are used to calculate them depends on the type ofpolarization radar, e.g., single channel switching, simultaneous transmit and receive (STAR),dual channel switching. SIGMET, Inc. is licensed by US National Severe Storms Laboratory(NSSL) to use the STAR hardware and processing techniques and algorithms.Polarization measurements require special calibration of the ZDR and LDR offsets. The use of aclutter filter for the polarization variables can sometimes bias the derived parameters. Because ofthis, the user decides whether or not to use filtered or unfiltered time series.1.4.6 Output DataThe RVP8 output data for standard moment calculations consist of mean radial velocity (V),Spectrum Width (W), Corrected Reflectivity(Z or dBZ) and Uncorrected Reflectivity (T ordBT). Other data outputs include I/Q time series, DFT/FFT power spectrum points andpolarization parameters. The output can be made in either 8 or 16-bit format. 8-bit format ispreferred over 16-bit format for most applications since the accuracy is more than adequate foran operational radar system, and the data communications are reduced by 50%. 16-bit formatsare sometimes used by research customers for data archive purposes. Note that time series andDFT are always 16-bit formats. All data formats are documented in Chapter 6 of this manual.A standard output is the I/Q time series on gigabit network (1000 BaseT). These are sent viaUDP broadcast to an I/Q archiving system or even a completely independent parallel processingsystem.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–311.5 RVP8 Control and Maintenance Features1.5.1 Radar Control FunctionsThe RVP8 also performs several important radar control functions:STrigger generation- up to 6 programmable triggers.SPulsewidth control (four states controlled by four bits).SAngle/data synchronization- to collect data at precise azimuth intervals (e.g., every 0.5,1, 1.5 degrees) based on the AZ/EL angle inputs.SPhase shifter- to control the phase on legacy Klystron systems. New or upgrade Klystronor TWT systems can use the RVP8/Tx card to provide very accurate phase shifting.SZDR switch control- for horizontal/vertical or other polarization switching scheme.SAFC output (digital or analog) based on the burst pulse analysis for magnetron systems.Pulsewidth and trigger control are both built into the RVP8. Four TTL output lines can beprogrammed to drive external relays that control the transmitter pulsewidth. The internal triggergenerator drives six separate lines, each of which can be programmed to produce a desiredwaveform. The trigger generator is unique in that the waveforms are stored in RAM and can bemodified interactively by user software. Thus, precisely delayed and jitter-free strobes and gatescan easily be produced. For each pulsewidth there is a corresponding maximum trigger rate thatcan be generated. Note, however, that the RVP8 can also operate from an external user-suppliedtrigger. In either case, the processor measures the trigger period between pulses so that usersoftware can monitor it as needed.The RVP8 also supports trigger blanking during which one or more (selectable) of the transmittriggers can be inhibited. Trigger blanking is used to avoid interference with other electronicequipment and to protect nearby personnel from radiation hazard. There are two techniques forthis:S2D AZ/EL sector blanking areas can be defined in the RVP8 itself.SAn external trigger blanking signal (switch closure to ground, TTL or RS422) can besupplied, for example from a proximity switch that triggers when the antenna goes belowa safe elevation angle or connected to the radome access hatch.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–321.5.2 Power-Up Setup ConfigurationThe RVP8 stores on disk an extensive set of configuration information. The purpose of thesedata is to define the exact configuration of the RVP8 upon startup. The setup information can beaccessed and modified using either a local keyboard and monitor, or over the network. Formultiple radar networks, the configuration management can be centrally administered bycopying tested “master” configuration files to the various network radars. It is not necessary togo to the radar to change ROM’s as was the case for previous generation processors.1.5.3 Built-In DiagnosticsOn power-up, the RVP8 performs a sequence of internal self-tests. The test sequence requiresabout four seconds to perform, and tests approximately 95% of the internal digital circuitry.Errors are isolated to specific sections of the board as much as possible. If any check fails, theuser can be certain that some component is not functioning correctly. However, there is a verysmall chance that even a defective board may pass all the tests; the failure may be in one of thefew areas that can not be checked.The RVP8 displays the test results on the LED front panel (for a standard SIGMET chassis). Inthis way, there is immediate visual confirmation of the diagnostic tests, even if the host computerhas not yet been connected. The local keyboard and monitor or a networked workstation can beused to see the test results in the TTY menus or even invoke a power–up reset and test.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–331.6 Support Utilities and Available Application SoftwareThe RVP8 system includes a complete set of tools for the calibration, alignment andconfiguration of the RVP8. These includes the following utilities:Sascope- a comprehensive utility for manual signal processor control and data display ofmoments, times series and Doppler spectra. ascope includes a realistic signal simulatorcapable of producing both first and second trip targets. Recording/playback of timeseries and moments is included as well.Sdspx- an ASCII text-based program to access and control the signal processor, includingproviding access to the local setup menus.Sspeed- a performance measuring utility.SDspExport- exports the RVP8 to another workstation over the network. This allowsutilities on a remote network to run locally, as opposed to exporting the utility displaywindow over the network.Ssetup- interactive GUI for creating/editing the RVP8 configuration files.Szauto- calibration utility for use with a test signal generator.These tools can be run locally on the RVP8 itself or over the network from a central maintenancefacility. The DspExport utility improves the performance of the utilities for network applicationsby letting them be run on the workstation that is remote from the RVP8. Note that standardX–Window export is of course supported but requires more bandwidth.In addition, complete radar application software can be purchased from SIGMET:SIRIS/Radar on a separate PC, interfaces to the RVP8 by 100 BaseT Ethernet.IRIS/Radar controls both the RVP8 and the SIGMET RCP8 radar/antenna controlprocessor. The package provides complete local and remote control/monitoring, dataprocessing and communication for a radar system.SIRIS/Analysis (and options) runs on a separate PC, often at a central site. OneIRIS/Analysis can support up to 20 radar systems. This functions as a radar productgenerator (RPG) to provide outputs such as CAPPI, rain accumulations, echo tops,automatic warning and tracking, etc. Optional software packages are provided for specialapplications: wind shear and microburst detection, hydrometeorology with raingagecalibration and subcatchments, composite, dual Doppler and 3D Display.SIRIS/Web provides IRIS displays to network users on standard PC’s (Windows orLinux) running Netscape or Internet Explorer.SIRIS/Display can display products sent to it and, with password authorization, can serveas a remote control and monitoring site for networked radar systems. Features such aslooping, cross–section, track, local warning, annotation, etc. are all provided byIRIS/Display. Note that both IRIS/Analysis and IRIS/Radar have all of the capabilities ofIRIS/Display in addition to their own functions. This means that any IRIS system candisplay products.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–341.7 System Network ArchitectureThe RVP8 provides considerable flexibility for network operation. This allowsremote control and monitoring of the system from virtually anywhere on the network,subject to the user’s particular security restrictions.Unlike the previous generation RVP7, which used a SCSI interface, the RVP8 uses anetwork interface exclusively. The “dsp lib” runs locally on the RVP8 and a utility,called DspExport, exports the library over the network using a TCP/IP socket.Typically this is exported to a local host radar control workstation (RCW) on thenetwork. Perhaps this workstation is running the SIGMET IRIS software. At least10BaseT connection is recommended for this connection.Figure 1–5: Network Architecture for Socket Interface with DspExportKeyboardMouseMonitorUtilitiesLocal Keyboard, Mouse, MonitorLocal Host RCWLAN or WAN TCPIP NetowrkRemote WorkstationSocket Interface ConnectionsRunningDspExportRVP8TCPIP LAN 10BaseT or betterUtilitiesUtilitiesA remote workstation on the network can also use the DspExport technique tocommunicate for configuration, monitoring and diagnostic testing.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–351.8 Open Architecture and Published APISIGMET recognizes that certain users may require the ability to write their own signalprocessing algorithms which will run on the RVP8. To accommodate this, the RVP8 software isorganized to allow separately compiled plug-in modules to be statically linked into the runningcode. The application program interface (API) allows user code to be inserted at the followingstages of processing:STx/Rx waveform synthesis and matched filter generation— The API allows the transmitwaveforms to be defined from pulse to pulse, along with the corresponding FIRcoefficients that will extract (I,Q) from that Tx waveform. This allows users toexperiment with arbitrary waveforms for pulse compression and frequency agility.STime series and spectra processing from (I,Q)- The API allows you to modify the defaulttime series and spectra data, e.g., to perform averaging or windowing in a different way.SParameter generation from (I,Q)- This is probably where the greatest activity will occurfor user–supplied code. The API allows you to redefine how the standard parameters(dBZ, Velocity, Width, PHIDP, etc.) are computed from the incoming (I,Q) time series.You may also create brand new parameter types that are not included in the basic RVP8data set.Note that the standard SIGMET algorithms are not made public in this model. Rather, theinterface hooks and development tools are provided so that users can add their own softwareextensions to the RVP8 framework. Many of the library routines that are fundamental to theRVP8 are also documented and can be called by user code; but the source to these routines is notgenerally released. Development tools which are not under public license must be purchasedseparately by the customer.While most customers will use the signal processing software supplied by SIGMET, the newopen software architecture approach employed by the RVP8 will be very useful to those researchcustomers who want to try innovative new approaches to signal processing, or to those OEMmanufacturers who are interested in having their own “custom” stamp on the product.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–361.9 RVP8 Technical Specifications1.9.1 IFD Digitizer Module, Rev E or laterInput SignalsSIF Received Signal: 50W, + 6.5 dBm full–scale, +20dBm absolute maxSIF Burst or COHO: 50W, +6.5 dBm full–scale, +20dBm absolute maxSOptional Reference Clock: 2–60 MHz –10 to 0 dBmIF Ranges:S12—34 MHz, 38—70 MHzLinear Dynamic RangeS85 to >100dB depending on pulsewidth/bandwidth filterA/D ConversionSResolution 14 bit with jitter <2.5 picosecSSampling rate 67 to 79 MHz (selectable, standard is 71.9364 MHz)AFC OutputSAnalog –10 to +10VSOptional Digital AFC (DAFC) with up to 24 programmable output bits.SAutomatic 2-D (time/frequency) burst pulse search and fine tracking algorithms.IFD LinkSUses shielded CAT 5E cable, non standard signals, requires RVP8/Rx card, rev C or later.Cable length to RVP8/RxS2—25 meters, with automatic calibration of round trip time and range correction.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–371.9.2 RVP8/Rx PCI Card, Rev C or laterPulse Repetition FrequencyS50 Hz to 20 KHz +0.1%, continuously selectable.IF Band Pass FilterSProgrammable Digital FIR with software selectable bandwidth. Built-in filter designsoftware with graphical user interface.Impulse ResponseSUp to 3024 FIR filter taps, corresponding to 75 msec impulse response length for 72 MHzIF samples at 125 meter range resolution. These very long filters are intended for usewith pulse compression.Range ResolutionSMinimum bin spacing of 25 meters selectable in N*8.33 meter steps. Bins can bepositioned in a configurable range mask with resolution of N* the fundamental binspacing, or arbitrarily to an accuracy of ±2.2 meters.Maximum RangeSUp to 1024 kmNumber of Range BinsSFull unambiguous range at minimum resolution or 3096 range bins (whichever is less).The RVP8 processor may only be fast enough to process an average of 50 meter bins.Electrical InterfacesSCAT 5E cable from the IFD, rev E or later.SBNC #1 for trigger output (12V, 75W), or pretrigger input.BNC #2 for trigger output (12V, 75W).S9-pin “D” connector supporting four RS-422 differential signals for miscellaneous inputand output with SoftPlanet support. Each line pair can operate as a transmitter or as areceiver depending on what’s needed. Possible uses are: alternate reference clock input,gating input for CW modes, additional trigger outputs, external phase shift requests, etc.Data Output via PCI BusS16-bit floating I and Q valuesS14-bit raw IF samples
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–381.9.3 RVP8/Tx PCI CardAnalog Waveform ApplicationsSDigitally synthesized IF transmit waveform for pulse compression, frequency agility, andphase modulation applications.SMaster clock or COHO signal to the radar; can be phase locked or free running, arbitraryfrequency.Analog Output Waveform CharacteristicsSTwo independent, digitally synthesized, analog output waveforms (BNC). These twooutputs are electrically identical and logically independent IF waveform synthesizers thatcan produce phase modulated CW signals, finite duration pulses, compressed pulses, etc.SCan drive up to +12dBm into 50W.S14-bit interpolating TxDAC provides 71dB Signal-to-Noise Ratio.SIF center frequency selectable from 8 to 32.4 MHz, and from 48.6 to 75MHz.SSignal bandwidth as large as 15MHz for wideband/multiband Tx applications. Bandwidth is adjustable in software.SContinuous or pulse modulated output with band width limiting on pulse modulationoutput.SPrecise phase shifting with transient band width limiting.STotal harmonic distortion less than –74dB.SWaveform pre-emphasis compensates for both static and dynamic Tx nonlinearities.Other I/O signalsSClock In/Out 50W SMA connector. This can receive a CW reference frequency to whichthe RVP8/Tx can lock to a P/Q frequency multiple (much like the RVP8/IFD can lock toan external reference). This connector can also supply the TxData Clock, optionallydivided by some N between 1 and 16, in order to supply external circuitry with +10dBmclock reference at 50W.S9-pin “D” connector supporting four RS-422 differential signals for miscellaneous inputand output with SoftPlanet support. Each line pair can operate as a transmitter or as areceiver depending on what’s needed. Possible uses are: alternate reference clock input,gating input for CW modes, additional trigger outputs, external phase shift requests, etc.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–391.9.4 SIGMET I/O-62 PCI CardSShort format PCI card with 62-position “D” connector. Multiple cards may be installed.SIncludes D/A, A/D, discrete inputs and outputs (TTL, wide range, RS422, etc.) Seesummary table below.SI/O pin assignment mapping by softplane.conf file.SStandard or custom remote backpanels available.SESD protection using Tranzorbt silicon avalanche diode surge suppression andhigh-voltage tolerant components.SIGMET I/O-62 Summary of Electrical InterfacesQty Description40 Lines configurable in groups of 8 to be either inputs or outputs. The electrical specifications aresoftware defined within each group as follows:SSingle-ended TTL input or output with software–configured pull-up or pull-down resistors for inputs.SWide range inputs (27VDC, threshold +2.5VDC), often used for “lamp voltage” status inputs.SRS-422/485 @ 10 MBit/sec (requires two lines each). RS-422 receivers can be configured in software to have 100W termination between each pair.8A/D convertors configurable as 0, 4, or 8 convertors, 2V, 12 bits @ 10 MHz, These lines areshared with some of the 40 I/O lines listed above.2D/A convertors, 10V 1 MHz update rate, output can drive a 75W load.2SPDT relays on the board. These are often used for switching high power relays. Contacts arediode protected.2RS-232C full duplex lines (Tx and Rx)412V 75W trigger drivers .2Power/Ground pairs of 12V power (filtered, fused) for external equipment or remote backpaneluse (up to 24 W total). Polyfuse technology acts like a circuit breaker with auto reset in the eventof an overload.8Ground wires for signal grounds from the remote back panel.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–401.9.5 I/O-62 Standard Connector PanelSMounts on front or rear of standard 19” EIA rackSConnects to I/O-62 via 1:1 62–pin 1.8–m cable (provided).SProvides standard inputs and outputs required by most weather radars such as triggers,polarization control, pulse width control and antenna angles.SAz and El synchro and reference inputs (nominal 100V 60 Hz)S3 internal relays and 4 12V relay control signals for switching external devices.SProgrammable scope test points with source waveforms selectable in software.SDiagnostic power supply and self test LED’s for troubleshooting.RVP8 Connector Panel Summary J-ID Label Type DescriptionJ1 AZ INPUT DBF25 Up to 16–bits of parallel TTL binary or BCD angleJ2 AZ OUTPUT DBF25 Up to 16–bits of parallel TTL binary or BCD angleJ3 PHASE OUT DBF25 Up to 8–bits of parallel TTL or RS422. Angles are configurable.J4 EL INPUT DBF25 Up to 16–bits of parallel TTL binary or BCD angleJ5 EL OUTPUT DBF25 Up to 16–bits of parallel TTL binary or BCD angleJ6 RELAY DBF25 3 internal relays, contact rating 0.5 A continuous. The switchingload is 0.25 A and 100V, with the additional constraint that the totalpower not exceed 4VA.4, 12V relay control signals, up to 200mA.(Note that external relays should be equipped with proper diodeprotection to shunt the back EMF).J7 SPARE DBF25 20 additional TTL I/O lines each configurable to be input or output.J8 SPARE DBF25 10 differential analog inputs, up to ±20V max multiplexed into A/Dconvertor sampling each at >1000 Hz.J9 MISC I/O DBF25 7 additional RS422 lines and 2 each dedicated (non–multiplexed)A/D inputs (±580V with pot adjust) and D/A outputs (±10V).J10 SERIAL DBF9 RS232CJ11 SERIAL DBF9 RS232CJ12 S–D Modular 3 x 4 matrix connector for AZ and EL synchro and reference inputsJ13 TP–1 BNC Programmable scope test point. 75 OhmsJ14 TP–2 BNC Programmable scope test point. 75 OhmsJ15 TRIG–1 BNC 12V trigger into 75 OhmsJ16 TRIG–2 BNC 12V trigger into 75 OhmsJ17 TRIG–3 BNC 12V trigger into 75 OhmsJ18 TRIG–4 BNC 12V trigger into 75 Ohms
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–411.9.6 RVP8 Processing AlgorithmsInput from Rx BoardS16–bit I/Q samplesSOptional dual-channel I/Q samples (e.g., for polarization systems or dual frequencysystems)IQ Signal Correction OptionsSAmplitude jitter correction based on running average of transmit power from burst pulse.SInterference correction for single pulse interferenceSSaturation correction (3 to 5 dB)Primary Processing ModesSPoly-Pulse Pair (PPP)SDFTSRandom or Phase Coded 2nd trip echo filtering/recoverySOptional Polarization with full co-variance matrix (ZDR, PHIDP, LDR, RHOHV, etc.)SOptional Pulse CompressionProcessing OptionsSFIR Clutter filters (40 and 50 dB) in pulse pair mode.SAdaptive width clutter filters in DFT and phase coded 2nd trip mode.SVelocity De-Aliasing: Dual PRF Velocity unfolding at 3:2, 4:3 and 5:4 PRF ratios orDual PRT Velocity processing for selectable inter-pulse intervals.SRange De-aliasing: Phase coding method (random phase for magnetron)Frequency coding method (not available for magnetron)SScan angle synchronization for data acquisition.SPulse integration up to 1024SCorrections for gaseous attenuation and 1/R2.SUp to 4 pulse widthsData OutputsSdBZ Calibrated equivalent radar reflectivity, 8 or 16 bitsSV Mean radial velocity, 8 or 16 bits
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–42SW Spectrum width, 8 or 16 bitsSI/Q Time series, 16 bits each per sampleSDFT Doppler Spectrum output option in DFT mode, 16 bits per componentSOptional: ZDR, PHIDP, RHOHV, LDR, RHO, 8 or 16 bitsData Quality ThresholdsSSignal–to–noise ratio (SNR) Used to reject bins having weak signals. Typically applied to dBZ.SSignal quality index (SQI) Used to reject bins having incoherent signals. Typically applied to mean velocity and width.SClutter-to-signal ratio (CSR) Used to reject range bins having very strong clutter. Typically applied to mean velocity, width and dBZ.SSpeckle Filter Filter removes single-bin targets such as aircraft or noiseFills isolated missing pixels as well.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–431.9.7 RVP8 Input/Output SummaryEthernet Input/Output from Host ComputerSData output of calibrated dBZ, V and W during normal operation. Full I/Q timeseriesrecording with a separate tsarchive utility, or through a customer’s application using apublic API. Signal processor configuration and verification read–back is performed viathe Ethernet interface.RS-232C Serial Data I/OSFor real time display/monitoring or data remoting.AZ/EL Angle Input OptionsSSerial AZ/EL angle tag input using standard SIGMET RCP format.S16-bit each parallel TTL binary angles via the I/O-62 card.SSynchro angle inputs via the I/O-62 card.SSIGMET network antenna packet protocol.Trigger OutputSUp to 10 total triggers available on various connector pins. Triggers are programmablewith respect to trigger start, trigger width and sense (normal or inverted).Optional Polarization ControlSRS-422 differential control for polarization switch.
Introduction and SpecificationsRVP8 User’s ManualMarch 20061–441.9.8 Physical and Environmental CharacteristicsPackagingSMotherboard Configuration 4U rack mount with 6 PCI slotsSCustom PC configurations available or packaged by customer.SDimensions of standard 4U chassis43.2 wide x 43.2 long x 17.8 cm high17 wide x 17 long x 7.00 inch highSDimensions IF Digitizer2.5 wide x 10.9 long x 23.6 cm high 1 wide x 4.3 long x 9.3 inch highSRedundant Power Supplies. Three hot–swap modules with audio failure alarm.Input PowerSIFD 100–240 VAC 47–63 Hz auto–rangingSMain Chassis 60/50 Hz 115/230 VAC Manual SwitchesPower ConsumptionSRVP8/Main Processor 180 Watts with Rx and MotherboardSRVP8/IFD IF Digitizer 12 WattsEnvironmentalSTemperature 0C (32F) to 50C (122F)SHumidity 0 to 95% non–condensingReliabilitySMTBF>50,000 hours (based on actual RVP7 field data).

Navigation menu