`
`85
`AD— A26?.
`I1‘!|!E1I!'!'!!11!!
`
`o“.
`
`s
`
`'
`
`-
`
`4
`‘i
`
`WHOl-92-44
`
`Woods Hole
`Oceanographic
`Institution
`
`
`
`Array Data Acquisition wlth Wireless LAN TeIemetry as
`
`applied to Shallow Water Tomography In the Barents Sea
`
`K. von der Heydt, J. Kemp, J. Lynch, J. Miller and 0.8. Chiu
`
`by
`
`’
`
`1
`
`December 1992
`
`Technical Report
`
`meamwmwmmmmmmwcmmCmmmmmtm
`Contact Mamas-920007151116 the Office of Naval Research mmm N00014-9N-1246.
`
`Approved for public release; distribution unlimited
`
`
`
`' “W 17 078
`
`05634
`\\N!N!
`Pet1t10nerMotorolaMobilityLflfiExhibit 1011-Page 1
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 1
`
`
`
`
`
`
`mscmm NOTICE
`
`
`
`THIS
`
`DOCUMENT
`
`IS
`
`BEST
`
`QUALITY AVAILABLE. THE COPY
`
`FURNISHED TO DTIC CONTAINED
`
`A SIGNIFICANT NUMBER OF
`
`COLOR PAGES WHICH DO NOT
`
`REPRODUCE LEGIBLY ON BLACK
`
`AND WHITE MICROFICHE.
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 2
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 2
`
`
`
`
`
`WHOI-92-44
`
`Array Data Acquisition with Wireless LAN Telemetry as applied to
`Shallow Water Tomography in the Barents Sea
`
`by
`
`K. von der Heydt. J. Kemp. and]. Lynch
`
`Woods Hole Oceanographic Institution
`Woods Hole. Massachusetts
`
`and
`
`
`
`J. Miller. and cs. Chin
`
`Naval Postgraduate School
`
`
`
`
`
`NTlS CRA&l
`DT'C TAB
`Unannounced
`Justification
`
`Monterey, California
` Technical Report
`
`December 1992
`
`
`
`
`Avail and] or
`SpeCial
`
`Funding was provided by the Long Beach Naval Regional Contracting Center Detachment
`under Contract N00123-92—C-00’7 l and the Office of Naval Research under Contract N000] 4-91-1-1246.
`
`Reproduction in whole or in part is permitted for any purpose of the
`United States Government. This report should be cited as:
`Woods Hole Ooeanog. Inst Tech. Rept., WHOl-92-44.
`
`Approved for publication; distribution unlimited.
`
`Approved for Distribution:
`
`a“
`
`"
`
`George V. Frisk, Chairman
`Department of Applied Ocean Physics and Engineering
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 3
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 3
`
`
`
`
`
`
`
`Array Data Acquisition with Wireless LAN Telemetry as applied to
`
`Shallow Water Tomography in the Barents Sea
`
`K. von der Heydt, J. Kemp, J. Lynch, Woods Hole Oceanographic Institution, Woods Hole. MA
`J. Miller, C. 3. Chin, Naval Postgraduate School, Monterey, CA
`
`December 29, 1992
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 4
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 4
`
`
`
`
`Contents
`
`1 Scientific Motivation
`
`1.1 The Barents Sea Polar Front Experiment ........................
`
`2 Technical Backround
`
`3
`
`3
`
`4
`
`6
`3 System Design
`3.1 Array ............................................ 6
`3.2 Mooring ...........................................
`T
`3.3 Buoy ............................................. 8
`3.4 The Shipboard System ................................... 10
`
`3.5 Acquisition Software ................................... 10
`
`11
`4 Operation
`4.1 Deployment and Recovery ................................. 11
`4.2 Acquisition and Telemetry ................................. 1'2
`
`5 Results
`
`13
`
`5.1 Mooring Performance ................................... 13
`5.2 Telemetry Performance ................................... 13
`5.3 Scientific Results ...................................... 14
`
`6 Future Plans
`
`7 Acknowledgments
`
`8 Figures
`
`15
`
`16
`
`16
`
`O
`
`‘
`
`.
`
`'
`
`
`
`2
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 5
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 5
`
`
`
`
`1 Scientific Motivation
`
`By examining the details of the propagation of a pulse of sound sent from a source to a receiver
`some distance away, it is possible to learn about the medium through which the sound propagated.
`Use of sound pulses to study the structure of the water column is generally called “acoustic to~
`mography". a technique which has been developed by a number of researchers over the past fifteen
`years. Mid frequency (50 to 1000 Hz) studies of the top kilometer of the ocean bottom are often
`referred to as "bottom acoustics”.
`
`For both civilian and naval purposes, the study of the water column and bottom in shallow water
`coastal regions (H _<_ 500m) has become increasingly important.
`It has long been known that. in
`looking at the details of the acoustic field from 50 to 1000 Hz in shallow water environments. an
`acoustic normal mode representation is more useful than a ray representation (Figure 1). The
`modal representation is not only exact mathematically, but it is also efficient. i.e. one needs only
`a few normal modes to represent the acoustic field in shallow water whereas one would need many
`rays. (The opposite is true in deep water.) In studying the ocean bottom or the water column using
`pulses of sound (broadband signals), it has been common practice to look at multi-path arrivals.
`be they rays or modes, which are cleanly separated in arrival time (Figure ‘2). This is relatively
`easy to do at long ranges, where small differences in the group velocities separate the arrivals in
`time. The equation
`
`ATnmil = R/Avgnil a
`
`(1)
`
`where AT, AvG and R are the arrival time difference, the group velocity difference, and the range
`separation respectively, shows the simple proportionality to distance. However. at the short ranges
`to which one is generally limited in shallow water due to the high propagation loss incurred by
`boundary reflections, time separation of different modal arrivals can become impossible. Since
`the modal arrival times are basic data for bottom and water column inverses. one often needs to
`
`resort to spatial array processing methods to separate the modes. Both vertical and horizontal
`acoustic arrays can be used to filter modes in shallow water. However, in terms of mode filtration
`efficiency per unit length of array, vertical arrays are preferred. Moreover, one is able (to a first
`approximation) to separate as many normal modes as there are hydrophones in the array.
`
`1.1 The Barents Sea Polar Front Experiment
`
`A vertical line array (VLA), played a vital role in the Barents Sea Polar Front Experiment
`conducted in the coastal waters of Svalbard in August 1992.
`In this experiment. a combination
`of CTD hydrography, moored oceanographic sensors, broadband acoustic bottom measurements.
`and acoustic tomography were employed to study the structure and dynamics of the Polar Front.
`Participants in the experiment included R. Bourke, C.S. Chin, J. Miller and M. Stone from N PS. R.
`Muench from SAIC, J. Bouthillette, J. Kemp, S. Liberatore, J. Lynch, A. Newhall. R. Pawlowicz.
`K. von der Heydt, and N. Witzell from WHOI.
`This report is focused on the technical aspects of the development and use for the Barents Sea
`Polar Front Experiment of a VLA coupled to a buoy using a new radio telemetry technique to
`acquire data in realtime aboard a nearby ship. Scientific results of the experiment will be reported
`in pending journal articles by a. subset of the above mentioned personnel collectively referred to as
`the Barents Sea Polar Front Group.
`The Polar Front is formed by the confluence of cold, fresh Arctic water and warmer. saltier
`North Atlantic water and is bound topographically to the 200m isobath. The topography of the
`experiment area is found in Figure 3a, along with the three tomographic paths along which we
`transmitted. It is seen that the eastern half of our experiment region (which actually extended 10
`
`
`
`
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 6
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 6
`
`
`
`km further east), was a rather simple E-W shelf, whereas the western half had a significant valley
`(called Fingerdjupet) which could make the oceanography more complicated, e.g. by creating eddies
`which propagate up the valley and mix waters across the shelf. A topview of the front is seen in
`Figure 3b, which displays the water temperature at 70m, just below the mixed layer. One sees the
`first order adherence of the front to the bathymetry, including the valley. A side view of the front
`in terms of temperature is shown in Figure 3c. Many interesting features are evident. At about
`75m depth, one most clearly sees the cold Arctic water on the right and the warmer Atlantic water
`on the left. Protrusions of Atlantic water into the Arctic water are also seen from 50-100m depth.
`These are thought to be significant in frontal mixing. The surface layer is clearly seen from zero to
`35m depth, and interestingly enough, does n_ot show a distinct front in temperature (though it does
`in salinity). A flow of colder Arctic water underneath the Atlantic water is seen near the bottom.
`The temperature component of the experiment was included to look at the temporal and spatial
`structure which is aliased by the CTD survey and moored sensors. For instance. the front can move
`5-10 km over a 12 hour tidal cycle, which is the order of time it takes to complete a north-south CTD
`transect. By sampling every five minutes, tomography can avoid aliasing this frontal movement.
`which also may account for some of the E-W corregations seen in Figure 3b. The baroclinic tilting
`mode of the front, the time evolution of the interleaving structure, the internal waves and internal
`tides are also of interest and should be observable by tomography.
`
`2 Technical Backround
`
`Initially our experimental plan was for the USN S Bartlett to remain on station tethered to the
`array so that data could be acquired on board to. verify operation of the VLA system and to process
`data in real time to adapt the experiment to unforseen environmental changes. Our main concerns
`with this scheme were twofold:
`i) that noise radiated form the ship could occlude the acoustic
`signatures that we hoped to record, and ii) that it would be impractical to tether an affordable
`array to the Bartlett which offered no facilities for multi~point mooring.
`Ship noise has been a chief concern with attempts to record low SNR data from tethered vertical
`arrays and is the reason for extraordinary efforts towards noise abatement in the design of ships
`such as the R/V Alliance operated by the SACLANT center in La Spezia, Italy. Also, experiments
`have demonstrated that it is no small feat to remain successfully connected to a bottom founded
`array for days, even with much smaller and more maneuverable craft than the Bartlett. These two
`issues have always been strong motivation to entirely decouple vertical arrays from a ship by using
`a buoy equipped with the means either to store data for retrieval when the system is recovered.
`transmit data to the ship in realtime or some combination of the two. Our design had to address
`these issues.
`
`Though efforts are underway at WHOI to design hydrophone arrays employing very high res-
`olution analog-to—digital conversion at sensor sites and digital transmission of data on a minimum
`of conductors up the cable, the Barents Sea project enjoyed neither the time nor the resources to
`pursue that goal. It was clear at the outset that we had to utilize an existing array architecture to
`minimize risk and cost.
`
`The ocean acoustics community has assembled arrays of hydrophones in many different configo
`urations with a wide variety of objectives. Our requirements were for a system that included an
`array that could be deployed vertically, with emphasis on the near bottom region in 200-300m of
`water with a minimum ”watch circle” of lateral motion, as well as the means to record digitized
`signals aboard the ship. To separate the roughly ‘20 acoustic normal modes which we expected
`from our primary cross-front tomographic track, a 16 channel hydrophone array was considered a
`
`
`
`-
`
`'
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 7
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 7
`
`
`
`
`minimum configuration.
`The acoustics component of the science plan called for use of tomographic acoustic sources
`operating at center frequencies of 224 and 400 Hz, the latter with up to 100 Hz of bandwidth.
`In addition, SUS charges were to be dropped from both the USNS Bartlett and a plane which
`would clearly inject energy at frequencies beyond 400 Hz; however it would haVe been too difficult
`to monitor the position of array sensors with sufficient accuracy to permit coherent processing of
`data beyond 500 Hz. Tomographic signals in previous experiments have typically been sampled
`at a convenient multiple of 4 times the center frequency which eases the quadrature demodulation
`process that is usually the first step in processing such data. A sampling rate of 1600 Hz was chosen
`since 2 of the 3 sources were to be centered at 400 Hz with 100 Hz bandwidth, allowing anti—alias
`filtering at 500 Hz. Data at the 224 Hz center frequency is thus interpolated and resampled prior
`to processing.
`
`We were also keen on recording the data aboard ship rather than with a self-contained system
`internal to the buoy, both to be confident of data quality in realtime and because we had no suitable
`autonomous system. Various broadband FM transmission systems have been invented or adapted to
`multiplex individual channels from acoustic arrays for transmission to distant receiving platforms.
`The dynamic range, overall data quality and reliability of these systems has been variable and often
`in the field, some or all of these characteristics have been less than ideal. Though such systems
`have been used successfully, we were not in a position to quickly borrow or buy and adapt one
`for our needs. Similarly, digital transmission using broadband FM links, such as those available
`from Aydin-Vector, to send serial bit streams at megabit /sec rates can and haVe been configured.
`These systems require a receiver and transmitter at both ends to achieve a bidirectional link. as
`well as signal conditioners, bit synchronizers and forward error correction hardware/ software to
`achieve these rates with reliable reception of data. Despite the high quality of these products. a
`large amount of time and money can be spent developing protocols that are able to detect errors
`and request retransmission of lost or erroneous data. It became clear that for our purposes. this
`method was risky and simply too expensive.
`During the summer of 1991, we had begun to experiment with wireless Local Area Networks
`(LAN’s), which are now becoming available with raw rates up to 5 megabits/sec. These are radio
`or infrared linked systems that are designed primarily for use in office spaces, warehouses and in
`some cases between buildings where cables are expensive to run or change. The data transport
`layer of the radio based systems is designed for shorter range and lower bit rate but otherwise
`similar to microwave links used to transmit digital data over long distances for telephone traffic
`and dedicated network connections between distant sites. These new products are typically spread
`spectrum, operate in the ISM bands, (though at least one is licensed at 18 GHz) and offer true
`LAN topologies in that they provide multi-node network capability without cables.
`When we first became interested, the NCR WaveLAN had just become available. The WaveLAN
`hardware is a board that plugs into a PC, (16 bit ISA bus), and achieves an Ethernet connection
`to additional PC’s each with WaveLAN boards, using spread spectrum techniques in the .902 ~ 928
`MHz radio band rather than coaxial or twisted pair cable. WaveLAN is advertised as having a max-
`imum range of 5 miles using ‘20 dB of antenna gain and a power level of 24dBm. For applications
`such as ours, the compelling feature of these systems is that they allow use of software that over the
`years has been developed and refined for the LAN world to assure reliable data transmission be-
`tween systems connected on the world—wide Internet network, through use of Transmission Control
`Protocol (TCP). Though this softWare protocol has been designed for use over cables in a Collision
`Sense Multiple Access, (CSMA) environment, they are equally effective over radio links. it matters
`not to the TCP protocol whether transmitted data is lost due to radio link fades, interference or
`”nobody home”. The nature of TCP guarantees that if the data is delivered, there is a very high
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 8
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 8
`
`
`
`
`
`
`
`probability that it is 100% correct.
`‘
`Our early field tests over water not only confirmed that we could use the same Ethernet appliv
`cations such as "ftp" for reliable file transfer over the WaveLAN hardware (as we are accustomed
`with cabled networks), but that with antennas of modest gain, the system was indeed useful over
`water at line—of—sight ranges of at least a 1 km. These were very promising results that encouraged
`our application of this emerging technology to remote acquisistion of data from a buoy to which
`instrumentation was attached. Data rates using ”ftp" between ‘2 mid—speed 286 and 386 PC‘S were
`typically 30 to 40 kbytes/sec to/from slow disk drives and twice that to/from memory.
`Some months later after funding for the VLA portion of the Barents Sea experiment had been
`obtained, further testing was done. From FTP, a network software vender in Cambridge MA. we
`acquired their software package PC/TCP, for developing Ethernet TCP applications for PC‘s. This.
`along with the Network Data Interface Standard (NDIS) driver that NCR was supplying with the
`WaveLAN hardware. allowed us to write code that directly accessed TCP functionality. “is were
`able to demonstrate that with our software, the WaveLAN hardware was capable of as much as 150
`kbytes/sec over a TCP stream connection. This was quite acceptable for a link rated at a maximum
`of '2 megabits/sec and comfortably greater than the 50 kbytes/sec we estimated was necessary to
`continuously acquire the data from a 16 channel VLA during the Barents Sea experiment.
`
`3 System Design
`
`3.1 Array
`
`Fabrication of the array cable with 16 breakouts for hydrophones was contracted to Neptune
`Technologies in Picayune MS. Although hydrophones could have been purchased as well. we chose
`to manufacture our own as we have done for other projects. The cable was 1.5 km in length.
`contained 39 individually insulated, stranded #26 copper conductors. (of which 32 were used).
`layed up in 3 concentric layers. The conductors were covered with a layer of Kevlar with a nominal
`12000 lb breaking strength. covered by a polyester sheath interw0ven with a hairy fairing for strum
`suppression. The finished, relaxed outer diameter is about 5/8”. The basic cable was made by Yale
`Cordage. Figure 4 is a plot of showing stretch and yield strength of the Kevlar component of the
`cable.
`
`Neptune installed 16 breakouts at 10 m spacing starting from the outboard end of the cable.
`leaving a ‘2 pin connector and aluminum tube protective cover for a hydrophone at each site. The
`pair of wires to any given sensor is disconnected from the remaining length of that pair beyond
`the breakout. Table 1 details the conductor numbering and colors, the channel spacing. sensor
`numbering and connector pinout cross references.
`
`A current mode preamplifier was used with each sensor such that a single independent pair of
`wires carries power to and signal from each hydrophone. Voltage driven designs can result in fewer
`wires: however, they are less desirable as breakage of conductors can affect. more than one sensor.
`The preamps, housed internal to the cylindrical sensor elements, draw approximately 3.6 ma from
`a power supply that must be at least 10 volts, but can range up to about 30 volts DC. The nominal
`sensitivity of the hydrophones is -160 dBV referenced to 1 nPascal. The equivalent sound pressure
`noise (ESPN) of the preamp in dB re 1 pPa is 34 © 100 Hz, and approximately ‘2? =0: 1000 Hz.
`The ~3db point of the sensors was set at 50 Hz to minimize signal levels from array and sea surface
`motion. Figure 5 is a power spectral density (PSD) plot of the preamp self noise level referred to
`the input (for sensor 15), relating ESPN to frequency.
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 9
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 9
`
`
`
`The sensor consists of two PZT-l ceramic cylinders in series with stainless steel endcaps .25“
`thick. The cylinders and endcaps were assembled with isocyanurate glue. The power/signal lead
`was brought through a tight hole in one endcap and the ground lead was simply connected with a
`screw on each side ofthat endcap. A Brantner MAW-2 connector (to mate with the array breakout).
`was then attached. The assembly was dipped in urethane for a thin insulating coating and allowed
`to cure.
`It was then dipped in a nickel based compound (Acheson Electrodag 550) to provide a
`complete electrostatic shield connected to ground. Finally, the shielded assembly was inserted into
`a mold for an outer layer of urethane. The finished size of the sensor was approximately 4" long
`with a 1.5" diameter and fit snugly into the protective aluminum housings attached at each array
`breakout.
`
`Despite the modest bandwidth (500 Hz) of signals to be transmitted up 1.5 km wire pairs.
`we were concerned about crosstalk. particularly between adjacent conductor pairs. This led us to
`specify conductor assignments such that physically distant sensors used correspondingly separated
`wire pairs in the cable.
`It is our experience that it is difficult to model crosstalk as a function of
`frequency whether a cable contains twisted or untwisted conductor pairs. We did some simple tests
`of this cable in air as well as a 300 m piece in both air and sea water. The following measurements
`were made in air using the 1500 m Barents Sea cable with the signal injected from the sensor site
`at the end of the cable (CH 0) and channels 1 thmugh 15 terminated at the receiveing end. Similar
`measurements taken with the 300 m length in air and sea water suggest that crosstalk is generally
`lower by 15 to 20 dB in sea water.
`
`IN AIR --->
`
`FREQUENCY (Hz)
`50
`100
`
`200
`
`500
`1000
`
`(dB)
`CH 1
`‘50.6
`*44.6
`
`~38.6
`
`-30.3
`-22.2
`
`CH 15 (dB)
`-51.2
`—45.6
`
`-39.5
`
`-31.0
`-22.0
`
`We estimate that signal coupling from one channel to any other. while the array is deployed.
`would be attenuated by 45 to 50 dB at 500 Hz. As one might expect, crosstalk increases with cable
`length and frequency with approximately a “'20 log” relationship. After observing the crosstalk
`levels on a number of channels, it would appear that there is little difference with respect
`to
`physical proximity of one pair of conductors to another in the cable bundle.
`
`3.2 Mooring
`
`The U—shaped mooring system as shown in Figure 6 was designed for a short term deployment of
`the VLA. This design took advantage of previous sub-surface and “S-tether" technology developed
`at WHOI to accommodate a limited budget and a compressed schedule. Emphasis was placed on
`minimizing strain applied to the cable during deployment and on maximizing the distance between
`the array and the telemetry buoy. The array cable between the vertical active array section and the
`telemetry buoy was continuous to eliminate the risk and expense of using underwater connectors.
`Mechanical attachment points for anchors and instrumentation were fashioned using Kevlar grips.
`The section of the mooring containing the 16 element hydrophone array was modeled using the
`NOYFB software developed at WHO]. It was designed to maintain inclinations from the vertical
`of less than 3 degrees in a linear current profile of ‘25 cm/sec, (0.5 knot).
`An acoustic navigator was installed in-line at the array top and used to track the mooring
`motion by interrogating 3 bottom mounted transponders. Each of the 3 expendable transponders.
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 10
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 10
`
`
`
`
`
`operating in the 10 kHz region. were surveyed from the ship using a combination of GPS and
`acoustic tra» itime measurements to obtain estimates of their absolute positions. TWO-way travel
`times. rc arded within the navigator at 10 minute intervals throughout the experiment. were later
`processed to estimate the “watch circle” motion of the array.
`The surface leg of the mooring between the second bottom anchor and the buoy was designed
`as a low~forcing “S-tether“ with a scope of 1.5. WHOI‘S SURFMOOR design program was used to
`evaluate the mooring performance. This design in which a slack tether mechanically decouples buoy
`motion from the anchor. reduces the dynamic loading on the array cable.
`In addition. a flexible
`chain-in-urethane section was used directly beneath the buoy to reduce cable fatigue due to surface
`induced motion. The 39 conductor array cable was simply attached to the lower end of the chain
`section with a grip and passed upward around the outside of the buoy with chatting protection.
`
`3.3 Buoy
`
`The Telemetry Surface Buoy, for which the cross section is shown in Figure 7. was originally
`designed as a coastal radio/ marker buoy. The hull is a rolled S irlyn foam flotation collar made by
`Gilman Corp. The foam collar is sandwiched between a removable galvanized steel upper structure
`and the lower ballast stand using 1/2" threaded rod. A 12 incn diameter. steel pipe instrument
`well extends centrally through the foam collar. A removable aluminum tube mast with supports
`provided mounts for the antenna. light and radar reflector approximately 10 feet all" the water. The
`buoy specifications are as follows:
`
`0 Air weight. 906 lbs
`
`0 Reserve buoyancy, 1870 lbs
`
`0 Overall height. 14 ft
`
`0 Maximum diameter. 6‘2 in
`
`0 Well inside depth. 70 in
`
`0 Well inside diameter, 12 in
`
`Referring again to Figure 7, the central 12” diameter tube of the buoy housed alkaline batteries
`for 10 days operation, an accurate self contained timebase and realtime clock. a 6 slot. 16 bit ISA
`backplane within a chassis containing a 16 channel amplifier and filter, a DC to DC converter. and
`4 plug-in cards: i) a 16 channel ADC board, ii) an interface to the external clock. iii) the WaveLAN
`board, and iv) a single board ’C286 based DOS PC. Figure 8 is a block diagram of the acquisition
`and telemetry system contained within the buoy. The array signals were connected via a 37 pin
`waterproof Impulse M/N MSAM-37-BCR,CCP connector mounted on the cover of the buoy tube.
`A ten foot mast supported a radar reflector, flasher and the antenna for the WaveLAN radio link.
`The dynamic range requirements of sampling a combination of nearby explosively generated
`signals and ambient level signals would normally lead us to use an auto-ranging or floating point
`amplifier/digitizing system that was developed at WHOI for other projects. This was impractical
`for this application as that system is bulky and would have used far too much battery power.
`The radio LAN link is bidirectional so a simple front end was designed that allows "manual" gain
`changes to accommodate the dramatically different signal level requirements of shots and the low
`SNR tomographic signals. Signal conditioning for each channel consisted of a differential input.
`current mode receiver with software programmable gain selections of 0. ‘20, 40. 8; 60 dB. followed
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 11
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 11
`
`
`
`by an 8 pole Butterworth low pass filter at 500 Hz. Figure 9 is the schematic of the current mode.
`differential input. programmable gain recevier and low pass filter applied to each channel. Figure 10
`is a plot showing the shape of the broadband response of the amplifier and filter. Sixteen channels
`were configured using ‘2 identical 8 channel printed circuit boards designed to fit into the cover
`(with shielding), of the 6~slot lSA bus chassis.
`With a fixed gain system, a 16 bit sample word was desirable. An 8 bit PC bus ADC board
`(Datel PC411) that multiplexes the 16 VLA channels through a single converter to achieve the
`1600 Hz sample rate was used. The Datel board was upgraded from 14 to 16 bits by changing
`the A to D converter part after tests of the board indicated that the increased resolution would
`be meaningful. The First In First Out memory (FIFO) on the ADC board was upgraded from 1
`to 8 kbytes to decrease the software overhead of data transfers to system memory. Our concern
`was that the high rate of interrupts from continuous network transactions and ADC data transfers
`would be a significant software overhead that might become the limiting factor for speed rather
`than the WaveLAN link itself.
`
`A 1 MHz timebase and IRIG-B timecode was supplied from a self-contained unit that was
`borrowed from another project. This package has internal batteries but was used with an external
`battery as well. The timebase source uses an Austron 1115 crystal which drifted approximately
`10 microsecond per hour over the period of the experiment. Stability and predictable drift of the
`timebase are crucial to accurate estimation of travel times of tomographic signals. which in turn
`determines how well the oceanographic parameters can be inferred. The lRlG-B code was used to
`supply a realtime sync anytime the buoy system was restarted either due to cycling of the power.
`a reset or by command over the LAN link. The operation of this clock unit was independent of the
`PC system in the buoy.
`Within the ISA chassis was a board, designed for another project and for which software existed.
`having a clock with microsecond resolution that could lie synchronized through software to the
`lRIG-B timecode from the external timebase unit. This clock was regularly read by the acquisition
`program with microsecond resolution for the current time value that was placed in data records
`transmitted to the shipboard system. In this way, there was never any doubt regarding the precise
`time associated with the data stream.
`
`The WaveLAN board is full length and has an inconvenient F-type, 75 ohm connector at the
`edge for an antenna feed cable. The antenna provided by N CR is omnidirectional with no gain and
`suitable for mounting on a wall near a desktop computer. On the buoy, an F to BNC‘ adapter was
`used to connect the antenna with about 12 feet of RG58 cable. The antenna and cable has a 50
`
`ohm characteristic impedance; however, the mismatch had no discernable effect. The RG58 cable
`was made as short as possible to minimize attenuation which at 900 MHz is about 14 dB/IOO ft.
`When coupled with the losses through connectors, the total attenuation was nearly equal to the
`3 dB gain of the buoy antenna, which was a 5/8 over 1/4 wavelength ungrounded whip.
`It was
`necessary to electrically isolate the antenna from the buoy or the radiated power would be very
`much diminished.
`
`The implementation of a DOS PC was easily achieved using a single board computer (SBC).
`that included a 16 MHz 800286, 2 MB of memory, a PROMDISK, standard interfaces for keyboard.
`parallel 85 serial ports, floppy and IDE disk drives. Program development was done using all the
`on-board facilities with disks and a separate video adapter; however, during operation in the buoy.
`no peripherals were connected to the CPU board and the video adapter was removed. The SBC was
`configured to boot as well as execute the acquisition program from the PROM DISK. The SEC was
`purchased from Micro Computer Specialists, Inc. and performed adequately. A later. faster version
`using a 3863K processor would have been more convenient and may permit faster transmissions
`with the WaveLAN, though at the expense of at least another watt of power dissipation.
`'26 A
`
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 12
`Petitioner Motorola Mobility LLC - Exhibit 1011 - Page 12
`
`
`
`
`
`
`
`
`
`commercially available DC/DC converter was mounted within the chassis to supply power to all
`plug-in cards. The efficiency of the conversion was about 82%, leading to a total power dissipation
`for the buoy system of about 28 watts as summarized below. The amplifier/filter section and the
`hydrophone preamps were supplied from a separate battery without regulation. The system ground
`was isolated from the buoy. The battery was a 11.75 inch diameter “puck" of 60 “D" alkaline cells
`configured as il‘ZVDC. The main battery powering the DC/DC converter consisted of 5 ~ 11.75 inch
`diameter "pucks", each made of 120 alkaline "D“ cells, stood 25 inches high and represented about
`Skw hours at zero degrees C. This battery was designed to be adequate for 9 days of continuous
`operation. Figure 11 shows the cell arrangement in one of the Dvcell pucks.
`The total power dissipation while transmitting at 50 kbytes/sec is summarized below:
`
`0 Array nominally 18VDC. 12v minimum *"Q 56ma. approx.
`
`1 watt
`
`0 16 CH Differential amp 8.: filter. nominally i 12VDC 2Q 100ma. approx 2.4 watt
`
`0 Timecode/timebase interface board, 0.5 watt
`
`0 16 channel ADC board. 3.? watts
`
`a WaveLAN board. 10 watts on average
`
`0 MCSI 'C'286 computer board,