Technology Focus Electronics/Computers · 2013-04-10 · 16 Strain-Gauge Measurement of Weight of Fluid in a Tank 16 Advanced Docking System With Magnetic Initial Capture 19 Machinery/Automation

Technology Focus

Electronics/Computers

Software

Materials

Mechanics

Machinery/Automation

Manufacturing

Bio-Medical

Physical Sciences

Information Sciences

Books and Reports

03-04 March 2004

https://ntrs.nasa.gov/search.jsp?R=20110016643 2020-05-09T09:56:32+00:00Z

INTRODUCTIONTech Briefs are short announcements of innovations originating from research and develop-

ment activities of the National Aeronautics and Space Administration. They emphasizeinformation considered likely to be transferable across industrial, regional, or disciplinary linesand are issued to encourage commercial application.

Availability of NASA Tech Briefs and TSPsRequests for individual Tech Briefs or for Technical Support Packages (TSPs) announced herein shouldbe addressed to

National Technology Transfer CenterTelephone No. (800) 678-6882 or via World Wide Web at www2.nttc.edu/leads/

Please reference the control numbers appearing at the end of each Tech Brief. Information on NASA’s Commercial Technology Team, its documents, and services is also available at the same facility or on the World Wide Web at www.nctn.hq.nasa.gov.

Commercial Technology Offices and Patent Counsels are located at NASA field centers to providetechnology-transfer access to industrial users. Inquiries can be made by contacting NASA field centersand program offices listed below.

Ames Research CenterCarolina Blake(650) [email protected]

Dryden Flight Research CenterJenny Baer-Riedhart(661) [email protected]

Goddard Space Flight CenterNona Cheeks(301) [email protected]

Jet Propulsion LaboratoryArt Murphy, Jr.(818) [email protected]

Johnson Space CenterCharlene E. Gilbert(281) [email protected]

Kennedy Space CenterJim Aliberti(321) [email protected]

Langley Research CenterJesse Midgett(757) [email protected]

John H. Glenn Research Center atLewis FieldLarry Viterna(216) [email protected]

Marshall Space Flight CenterVernotto McMillan(256) [email protected]

Stennis Space CenterRobert Bruce(228) [email protected]

Carl RaySmall Business InnovationResearch Program (SBIR) &Small Business TechnologyTransfer Program (STTR)(202) 358-4652 [email protected]

Benjamin NeumannInnovativeTechnology TransferPartnerships (Code RP)(202) [email protected]

John MankinsOffice of Space Flight (Code MP)(202) 358-4659 [email protected]

Terry HertzOffice of Aero-SpaceTechnology (Code RS)(202) 358-4636 [email protected]

Glen MucklowOffice of Space Sciences(Code SM)(202) 358-2235 [email protected]

Roger CrouchOffice of Microgravity ScienceApplications (Code U)(202) 358-0689 [email protected]

Granville PaulesOffice of Mission to Planet Earth(Code Y) (202) 358-0706 [email protected]

NASA Field Centers and Program Offices

NASA Program Offices

At NASA Headquarters there are seven major program offices that develop and oversee technology projects of potential interest to industry:

NASA Tech Briefs, March 2004 1

5 Technology Focus:

5 Advanced Signal Conditioners for Data-Acquisition Systems

5 Downlink Data Multiplexer

6 Viewing ISS Data in Real Time via the Internet

7 Autonomous Environment-Monitoring Networks

8 Readout of DSN Monitor Data

9 Electronics/Computers

9 Parallel-Processing Equalizers for Multi-GbpsCommunications

10 AIN-Based Packaging for SiC High-TemperatureElectronics

11 Software

11 Software for Optimizing Quality Assurance ofOther Software

11 The TechSat 21 Autonomous SciencecraftExperiment

11 Software for Analyzing Laminar-to-Turbulent Flow Transitions

13 Materials

13 Elastomer Filled With Single-Wall CarbonNanotubes

15 Mechanics

15 Modifying Ship Air-Wake Vortices for AircraftOperations

16 Strain-Gauge Measurement of Weight of Fluid ina Tank

16 Advanced Docking System With Magnetic InitialCapture

19 Machinery/Automation

19 Blade-Pitch Control for Quieting Tilt-RotorAircraft

20 Solar Array Panels With Dust-Removal Capability

21 Manufacturing

21 Aligning Arrays of Lenses and Single-ModeOptical Fibers

22 Automatic Control of Arc Process for MakingCarbon Nanotubes

22 Curved-Focal-Plane Arrays Using Deformed-Membrane Photodetectors

25 Physical Sciences

25 Role of Meteorology in Flights of a Solar-PoweredAirplane

26 Model of Mixing Layer With MulticomponentEvaporating Drops

27 Solution-Assisted Optical Contacting

29 Information Sciences

29 Improved Discrete Approximation of Laplacian ofGaussian

30 Utilizing Expert Knowledge in Estimating FutureSTS Costs

31 Books & Reports

31 Study of Rapid-Regression Liquefying HybridRocket Fuels

31 More About the Phase-SynchronizedEnhancement Method

03-04 March 2004

This document was prepared under the sponsorship of the National Aeronautics and Space Administration. Neither the United States Governmentnor any person acting on behalf of the United States Government assumes any liability resulting from the use of the information contained in thisdocument, or warrants that such use will be free from privately owned rights.



Technology Focus:

Advanced Signal Conditioners for Data-Acquisition Systems“Smart” circuitry repairs itself by switching in spare parts as needed.John F. Kennedy Space Center, Florida

Signal conditioners embodying ad-vanced concepts in analog and digitalelectronic circuitry and software havebeen developed for use in data-acquisi-tion systems that are required to be com-pact and lightweight, to utilize electricenergy efficiently, and to operate withhigh reliability, high accuracy, and highpower efficiency, without interventionby human technicians. These signal con-ditioners were originally intended foruse aboard spacecraft. There are alsonumerous potential terrestrial uses —especially in the fields of aeronauticsand medicine, wherein it is necessary tomonitor critical functions.

Going beyond the usual analog anddigital signal-processing functions ofprior signal conditioners, the new signalconditioner performs the following ad-ditional functions:• It continuously diagnoses its own elec-

tronic circuitry, so that it can detectfailures and repair itself (as describedbelow) within seconds.

• It continuously calibrates itself on thebasis of a highly accurate and stablevoltage reference, so that it can con-tinue to generate accurate measure-ment data, even under extreme envi-ronmental conditions.

• It repairs itself in the sense that it con-tains a microcontroller that reroutes

signals among redundant componentsas needed to maintain the ability toperform accurate and stable measure-ments.

• It detects deterioration of compo-nents, predicts future failures, and/ordetects imminent failures by means ofa real-time analysis in which, amongother things, data on its present stateare continuously compared with lo-cally stored historical data.

• It minimizes unnecessary consumptionof electric energy.The design architecture divides the sig-

nal conditioner into three main sections:an analog signal section, a digital mod-ule, and a power-management section.The design of the analog signal sectiondoes not follow the traditional approachof ensuring reliability through total re-dundancy of hardware: Instead, followingan approach called “spare parts — toolbox,” the reliability of each component isassessed in terms of such considerationsas risks of damage, mean times betweenfailures, and the effects of certain failureson the performance of the signal condi-tioner as a whole system. Then, fewer ormore spares are assigned for each af-fected component, pursuant to the re-sults of this analysis, in order to obtainthe required degree of reliability of thesignal conditioner as a whole system.

The digital module comprises one ormore processors and field-programma-ble gate arrays, the number of each de-pending on the results of the aforemen-tioned analysis. The digital moduleprovides redundant control, monitor-ing, and processing of several analog sig-nals. It is designed to minimize unneces-sary consumption of electric energy,including, when possible, going into alow-power “sleep” mode that is imple-mented in firmware. The digital modulecommunicates with external equipmentvia a personal-computer serial port. Thedigital module monitors the “health” ofthe rest of the signal conditioner by pro-cessing defined measurements and/ortrends. It automatically makes adjust-ments to respond to channel failures,compensate for effects of temperature,and maintain calibration.

This work was done by Angel Lucena andJose Perotti of Kennedy Space Center, andAnthony Eckhoff and Pedro Medelius of Dy-nacs, Inc. Further information is containedin a TSP (see page 1).

This invention is owned by NASA, and apatent application has been filed. Inquiriesconcerning nonexclusive or exclusive licensefor its commercial development should be ad-dressed to the Technology Programs and Com-mercialization Office, Kennedy Space Center,(321) 867-8130. Refer to KSC-12301.

Downlink Data MultiplexerBandwidth is allocated as needed among four data streams of various rates.Lyndon B. Johnson Space Center, Houston, Texas

A multiplexer/demultiplexer systemhas been developed to enable the trans-mission, over a single channel, of fourdata streams generated by a variety ofsources at different (including variable)bit rates. In the original intended appli-cation, replicas of this multiplexer/de-multiplexer system would be incorpo-rated into the spacecraft-to-groundcommunication systems of the spaceshuttles. The multiplexer of each systemwould be installed in the spacecraft,where it would acquire and process data

from such sources as commercial digitalcamcorders, video tape recorders, andthe spacecraft telemetry system. The de-multiplexer of each system would be in-stalled in a ground station. Purely ter-restrial systems of similar design couldbe attractive for use in situations inwhich there are requirements to trans-mit multiple streams of high-qualityvideo data and possibly other data oversingle channels.

The figure is a block diagram of themultiplexer as configured to process

data received via three fiber-optic chan-nels like those of the InternationalSpace Station and one electrical-cablechannel that conforms to the Instituteof Electrical and Electronic Engineers(IEEE) 1394 standard. (This standardconsists of specifications of a high-speedserial data interface, the physical layerof which includes a cable known in theart as “FireWire.” An IEEE 1394 interfacecan also transfer power between thecomponents to which it is connected.)The fiber-optic channels carry packet

6 NASA Tech Briefs, March 2004

and/or bit-stream signals that conformto the standards of the ConsultativeCommittee for Space Data Systems(CCSDS). The IEEE 1394 interface ac-cepts an isochronous signal like thatfrom a digital camcorder or a video taperecorder.

The processing of the four input datastreams to combine them into one out-put stream is governed by a statisticalmultiplexing algorithm that features aflow-control capability and makes it pos-sible to utilize the transmission channelwith nearly 100-percent efficiency. Thisalgorithm allocates the available band-width of the transmission channel to the

data streams according to a combinationof data rates and preassigned priorities.Incoming data streams that demand toomuch bandwidth are blocked. Band-width not needed for a transmission of agiven data stream is allocated to otherstreams as available. Priority is given tothe IEEE 1394 stream.

In addition to the four incoming datastreams, the multiplexer transmits dataon the status of the system. An operatorcan monitor and control the multi-plexer via displays and controls on themultiplexer housing. The output of themultiplexer is connected via a coaxialcable with an impedance of 50 Ω to an

interface circuit compatible with thespace-shuttle high-speed digital down-link, which operates at a rate of 48 Mb/s.

This work was done by S. Douglas Hol-land, Glen F. Steele, Denise M. Romero, andRobert David Koudelka of Johnson SpaceCenter. Further information is containedin a TSP (see page 1).

This invention is owned by NASA, and apatent application has been filed. Inquiriesconcerning nonexclusive or exclusive licensefor its commercial development should be ad-dressed to the Patent Counsel, Johnson SpaceCenter, (281) 483-0837. Refer to MSC-23303.

This Data Multiplexer combines four input streams (three fiber-optic TAXI interface channels and one IEEE 1394 channel) into a single output stream.

F/ORXInput

Channel0

RX Channel-ControlEPLD

TAXI

TAXI

TAXI

128K × 9FIFO

F/OTX TX

F/ORXInput

Channel1


256K × 9FIFO

F/OTX TX

F/ORXInput

Channel2


128K × 9FIFO

F/OTX TX

IEEE1394PHY

IEEE1394LLC

InputChannel

3

Notes: 1. “EPLD” signifies “erasable programmable logic device.” 2. “F/O” signifies “fiber/optic.” 3. “TX” signifies “transmitter.” 4. “RX’ signifies “receiver.” 5. “TAXI” signifies“transparent asynchronous transceiver interfacer.” 6. “FIFO” signifies “first in/first out memory.” 7. “PHY” signifies “physical layer.” 8. “LLC” signifies “link layer controller.”

Channel-ControlLogic Circuit

128K × 9FIFO

Master InputEPLD

StatusDual-PortMemory

Microcontroller

RS-232Interface

TEMP Real-TimeClock

Reed-SolomonError-Detection and-Correction Encoder

Interleaver

50-ΩDrivers

Randomizer

Parallel-to-SerialConverter

24-MHzClock

Clock-Signal-DistributingPhase-Lock

Loop

6 MHz

12 MHz

48 MHz

48-MHzClock Clock Signal Out

Data Signal Out

MainBus

Rate Selector

EZStream is a computer program thatenables authorized users at diverse ter-restrial locations to view, in real time,data generated by scientific payloadsaboard the International Space Station(ISS). The only computation/communi-cation resource needed for use ofEZStream is a computer equipped with

standard Web-browser software and aconnection to the Internet. EZStreamruns in conjunction with the TReK soft-ware, described in a prior NASA TechBriefs article, that coordinates multiplestreams of data for the ground commu-nication system of the ISS. EZStream in-cludes server components that interact

with TReK within the ISS ground com-munication system and client compo-nents that reside in the users' remotecomputers. Once an authorized clienthas logged in, a server component ofEZStream pulls the requested data froma TReK application-program interfaceand sends the data to the client. Future

Viewing ISS Data in Real Time via the InternetMarshall Space Flight Center, Alabama


EZStream enhancements will include(1) extensions that enable the server toreceive and process arbitrary datastreams on its own and (2) a Web-basedgraphical-user-interface-building sub-

program that enables a client who lacksprogramming expertise to create cus-tomized display Web pages.

This program was written by Gerry Myersand Jim Chamberlain of AZ Technology, Inc.,

for Marshall Space Flight Center. For fur-ther information, contact the company’s NewTechnology Representative, David O’Neil, at(256) 837-9877.MFS-31836

Autonomous Environment-Monitoring NetworksThese neural networks recognize novel features in streams of input data.NASA’s Jet Propulsion Laboratory, Pasadena, California

Autonomous environment-monitor-ing networks (AEMNs) are artificialneural networks that are specialized forrecognizing familiarity and, conversely,novelty. Like a biological neural net-work, an AEMN receives a constantstream of inputs. For purposes of com-putational implementation, the inputsare vector representations of the infor-mation of interest. As long as the mostrecent input vector is similar to the pre-vious input vectors, no action is taken.Action is taken only when a novel vectoris encountered. Whether a given inputvector is regarded as novel depends onthe previous vectors; hence, the sameinput vector could be regarded as famil-iar or novel, depending on the contextof previous input vectors. AEMNs havebeen proposed as means to enable ex-ploratory robots on remote planets torecognize novel features that couldmerit closer scientific attention. AEMNscould also be useful for processing datafrom medical instrumentation for auto-mated monitoring or diagnosis.

The primary substructure of an AEMNis called a spindle. In its simplest form, aspindle consists of a central vector (C), ascalar (r), and algorithms for changing Cand r. The vector C is constructed fromall the vectors in a given continuous

stream of inputs, such that it is minimallydistant from those vectors. The scalar r isthe distance between C and the most re-mote vector in the same set.

The construction of a spindle involvesfour vital parameters: setup size, spindle-population size, and the radii of two nov-elty boundaries. The setup size is thenumber of vectors that are taken into ac-count before computing C. The spindle-population size is the total number ofinput vectors used in constructing thespindle — counting both those that ar-rive before and those that arrive afterthe computation of C. The novelty-boundary radii are distances from C thatpartition the neighborhood around Cinto three concentric regions (see Fig-ure 1). During construction of the spin-dle, the changing spindle radius is de-noted by h. It is the final value of h,reached before beginning constructionon the next spindle, that is denoted by r.

During construction of a spindle, if anew vector falls between C and the innerboundary, the vector is regarded as com-pletely familiar and no action is taken. Ifthe new vector falls into the region be-

tween the inner and outer boundaries, itis considered unusual enough to war-rant the adjustment of C and r by use ofthe aforementioned algorithms, but notunusual enough to be considered novel.If a vector falls outside the outer bound-ary, it is considered novel, in which caseone of several appropriate responsescould be initiation of construction of anew spindle.

An AEMN comprises a collection ofspindles that represent a typical historyor range of behaviors of a system thatone seeks to monitor. An AEMN can berepresented as a familiarity map, onwhich successive spindles are repre-sented by adjacent circles that are addedas construction proceeds. A familiaritymap could be simple or complex, de-pending on the monitored system. Forexample, the range of behaviors of acomplex system might be represented bya networklike familiarity map that couldeven include dead-end branches thatlead to the demise of the system. An au-tomated monitoring system based on theAEMN corresponding to the familiaritymap could recognize that the system was

Figure 1. The Central Vector and The NoveltyBoundaries play major roles in the constructionof a spindle.

InnerNovelty Boundary

OuterNovelty Boundary

C

Central Vector

Figure 2. A Familiarity Map comprises a sequence of overlapping circles that represent spindles con-structed from data acquired in observation or simulation of a system to be monitored.

Disaster 1

Disaster 2Disaster 3


progressing along a dead-end branchand respond by generating an alarm ortriggering control action to move the sys-tem away from the dead-end condition.

This work was done by Charles Hand ofCaltech for NASA’s Jet Propulsion Labo-ratory. Further information is contained ina TSP (see page 1).

This software is available for commerciallicensing. Please contact Don Hart of the Cal-ifornia Institute of Technology at (818) 393-3425. Refer to NPO-30408.

DSN Monitor Data Reader is a com-puter program that, as its name suggests,reads file of monitor data from the DeepSpace Network (DSN). The monitor dataconstitute information on the status andperformance of tracking, telemetry, com-mand, and pointing equipment at theDSN antennas. The DSN has recently in-troduced a new, more advanced monitordata format, denoted 0158-Mon, that isbased on the standard formatted dataunit (SFDU) and compressed header

data objects (CHDO) of the ConsultativeCommittee for Space Data Systems(CCSDS). The 0158-Mon data format is avery flexible generic format that providesfor specific variable-length formats andfor self-identifying parameters that obvi-ate the proprietary NASA Communica-tions (NASCOM) bit-packed formats ofthe past. The monitor data SFDUs arealso encapsulated in Standard DSNBlocks and routed to DSN customers forprocessing at their local mission control

centers. This program helps a DSN cus-tomer to read and parse the monitor datato assess the statuses of the DSN stationsin support of spacecraft flight operations.

This program was written by Katherine Lev-ister and May Tran of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).


Readout of DSN Monitor DataNASA’s Jet Propulsion Laboratory, Pasadena, California


Electronics/Computers

Parallel-Processing Equalizers for Multi-Gbps CommunicationsOne can compromise among computational efficiency, complexity of circuitry, and processing rates.NASA’s Jet Propulsion Laboratory, Pasadena, California

Architectures have been proposedfor the design of frequency-domainleast-mean-square complex equalizersthat would be integral parts of paral-lel-processing digital receivers ofmultigigahertz radio signals and otherquadrature-phase-shift-keying (QPSK)or 16-quadrature-amplitude-modula-tion (16-QAM) of data signals at ratesof multiple gigabits per second.“Equalizers” as used here denotes re-ceiver subsystems that compensate fordistortions in the phase and frequencyresponses of the broad-band radio-fre-quency channels typically used to conveysuch signals. The proposed architec-tures are suitable for realization in very-large-scale integrated (VLSI) circuitryand, in particular, complementary metaloxide semiconductor (CMOS) applica-tion-specific integrated circuits (ASICs)operating at frequencies lower thanmodulation symbol rates.

A digital receiver of the type to whichthe proposed architecture applies (seeFigure 1) would include an analog-to-digital converter (A/D) operating at arate, fs, of 4 samples per symbol period.To obtain the high speed necessary forsampling, the A/D and a 1:16 demulti-plexer immediately following it wouldbe constructed as GaAs integrated cir-cuits. The parallel-processing circuitrydownstream of the demultiplexer, in-cluding a demodulator followed by anequalizer, would operate at a rate ofonly fs/16 (in other words, at 1/4 ofthe symbol rate). The output from theequalizer would be four parallelstreams of in-phase (I) and quadrature(Q) samples.

The proposed architectures wouldimplement subconvolution (see Figure2), fast-Fourier-transform/inverse-fast-Fourier-transform (FFT-IFFT), and dis-crete-Fourier-transform/inverse-discrete-Fourier-transform (DFT-IDFT) overlap-and-save filter algorithms. A key prop-erty of the proposed architectures isthat one can make engineering com-promises among computational effi-ciency, complexity of circuitry, andprocessing rates. Such trades are made

Figure 1. A Parallel-Processing Digital Receiver would include a parallel-processing equalizer.

Figure 2. A Parallel Subconvolution Filter Bank would perform R subconvolutions, each of length L+1,at 1/M of the input sample rate. The symbol z–1 denotes a delay of one sample period, “↓16” signifiesdecimation by a factor of 16, and Hi denotes a frequency-domain digital filter.

Anti-AliasingFilter A/D

Demultiplexer1:16

ParallelDemodulatorCMOS ASIC

ParallelEqualizer

Intermediate-FrequencySignal In

fs fs1/16

I1

I4

Q1

Q4

H1(0)

H1(L)

H2(L)

z–1z–1

z–1z–1

z–1z–1

z–1

z–1

↓M

↓M

HR (0)

HR (L)

HR (L)

DFT IDFT

x(n) 0

(L–1)/2

(L+1)/2

L

y(n)

y(n +M –1)


possible, in part, by utilizing subconvo-lutions and relatively simple digital sig-nal-processing methods in such a man-ner as to eliminate a lower boundimposed on FFT-IFFT lengths by equal-izer tap lengths. For a given receiver,the equalizer tap length would theo-retically be unlimited, and the FFT-IFFT length could be chosen com-pletely independently of the equalizertap length. The FFT-IFFT length couldbe determined on the basis of the de-

sired reduction in the processing rate.The specific values chosen for the pro-posed architectures are an equalizertap length of 32, with an FFT-IFFTlength of 8 chosen to enable process-ing at 1/4 of the symbol rate.

This work was done by Andrew Gray, Par-minder Ghuman, Scott Hoy, and Edgar H.Satorius of Caltech for NASA’s Jet Propul-sion Laboratory. Further information iscontained in a TSP (see page 1).

In accordance with Public Law 96-517,

the contractor has elected to retain title to thisinvention. Inquiries concerning rights for itscommercial use should be addressed to

Intellectual Assets OfficeJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109(818) 354-2240E-mail: [email protected] to NPO-30246, volume number and

page number.

Packaging made primarily of aluminumnitride has been developed to enclose sili-con carbide-based integrated circuits(ICs), including circuits containing SiC-based power diodes, that are capable ofoperation under conditions more severethan can be withstood by silicon-based in-tegrated circuits. A major objective of thisdevelopment was to enable packaged SiCelectronic circuits to operate continuouslyat temperatures up to 500 °C. AlN-pack-aged SiC electronic circuits have commer-cial potential for incorporation into high-power electronic equipment and intosensors that must withstand high tempera-tures and/or high pressures in diverse ap-plications that include exploration inouter space, well logging, and monitoringof nuclear power systems. This packagingembodies concepts drawn from flip-chippackaging of silicon-based integrated cir-cuits. One or more SiC-based circuit chipsare mounted on an aluminum nitridepackage substrate or sandwiched betweentwo such substrates. Intimate electrical

connections between metal conductors onthe chip(s) and the metal conductors onexternal circuits are made by direct bond-ing to interconnections on the packagesubstrate(s) and/or by use of holesthrough the package substrate(s). This ap-proach eliminates the need for wirebonds, which have been the most vulnera-ble links in conventional electronic cir-cuitry in hostile environments. Moreover,the elimination of wire bonds makes itpossible to pack chips more densely thanwas previously possible.

Especially notable components ofpackaging of this type are the following:• AlN substrates that have high thermal

conductivity [170 W/(m⋅K)] and a co-efficient of thermal expansion (CTE)that matches that of SiC;

• Thick gold conductor film circuit traces,the adhesion and sheet resistance of whichdo not change measurably at 500 °C overtime periods as long as 1,000 hours; and

• Glass passivation/sealing layers thathave a breakdown potential of 2,575 V

at room temperature and, at 500 °C,breakdown potentials of 1,100 V for en-capsulation of Au conductors and 1,585V for encapsulation of Pt conductors.The matching of CTEs minimizes ther-

mal stresses. Packaging interconnectionsare monometallic or bimetallic and able towithstand high temperatures. These andother features are known to contribute toreliability at high temperatures and are ex-pected to extend the high-temperaturefunctionality of the packaged electronicdevices. Further research will be necessaryto characterize the long-term reliability ofSiC-based circuits in AlN-based packages.

This work was done by Ender Savrun of Si-enna Technologies, Inc., for Glenn Re-search Center. For further information, ac-cess http://www.siennatech.com.

Inquiries concerning rights for the commer-cial use of this invention should be addressed toNASA Glenn Research Center, CommercialTechnology Office, Attn: Steve Fedor, Mail Stop4-8, 21000 Brookpark Road, Cleveland , Ohio44135. Refer to LEW-17478.

AIN-Based Packaging for SiC High-Temperature ElectronicsElectronic packaging can withstand a continuous temperature of 500 °C.John H. Glenn Research Center, Cleveland, Ohio


Software

Software for OptimizingQuality Assurance of Other Software

Software assurance is the plannedand systematic set of activities that en-sures that software processes and prod-ucts conform to requirements, stan-dards, and procedures. Examples ofsuch activities are the following: codeinspections, unit tests, design reviews,performance analyses, construction oftraceability matrices, etc. In practice,software development projects haveonly limited resources (e.g., schedule,budget, and availability of personnel)to cover the entire development effort,of which assurance is but a part. Pro-jects must therefore select judiciouslyfrom among the possible assurance ac-tivities. At its heart, this can be viewedas an optimization problem; namely, todetermine the allocation of limited re-sources (time, money, and personnel)to minimize risk or, alternatively, tominimize the resources needed to re-duce risk to an acceptable level. Theend result of the work reported here isa means to optimize quality-assuranceprocesses used in developing software.This is achieved by combining twoprior programs in an innovative man-ner:• First Program: The first of these pro-

grams is the Advanced Risk ReductionTool (ARRT), which can be used tocalculate the costs and benefits of aset of assurance activities on a givensoftware project. ARRT is itself basedon a risk-management tool, DefectDetection and Prevention (DDP).DDP uses a detailed mathematicalmodel of requirements, risks, and mit-igations.

• Second Program: The second of theseprograms is the TAR2 “treatmentlearner,” which can be used to deter-mine from a large set of factors thosefactor settings most critical to attaininga given objective.

• Innovative Combination: The majorcontribution of this work is the com-bination of these two programs. Theyare combined so as to operate in aniterative procedure, as follows: Ineach cycle of the iteration, TAR2 istuned to identify the most criticalsoftware assurance activities, boththose most critical to perform (be-cause they contribute to cost-effective

risk reduction), and those most criti-cal to not perform (because they de-tract from cost-effective risk reduc-tion).These identified activities are then

set accordingly in ARRT, and the cost-benefit calculations rerun. Repeatingthis cycle determines more and moreactivities to perform (and/or to notperform), culminating in a solutionthat is (near) optimal. An important as-pect of this approach is that it allowsfor human experts to add further guid-ance during each iteration of the cycle.For example, if the experts observethat two of the recommended activitiesare actually incompatible (say, becausethey would both require use of thesame limited resource at the sametime), they can reject the TAR2 recom-mendations involving this pair of activ-ities, and instead ask for the next-bestsolution. This makes good use of theexperts’ time, since they are only askedfor guidance pertinent to promising so-lutions.

This innovation was developed by MartinFeather and Steven Cornford of Caltech andTim Menzies of the University of British Co-lumbia for NASA’s Jet Propulsion Labora-tory. Further information is contained in aTSP (see page 1).


The TechSat 21 AutonomousSciencecraft Experiment

Software has been developed to per-form a number of functions essential toautonomous operation in the Au-tonomous Sciencecraft Experiment(ASE), which is scheduled to be demon-strated aboard a constellation of threespacecraft, denoted TechSat 21, to belaunched by the Air Force into orbitaround the Earth in January 2006. Aprior version of this software was re-ported in “Software for an AutonomousConstellation of Satellites” (NPO-30355), NASA Tech Briefs, Vol. 26, No. 11(November 2002), page 44.

The software includes the followingcomponents:• Algorithms to analyze image data, gen-

erate scientific data products, and de-tect conditions, features, and events of

potential scientific interest;• A program that uses component-based

computational models of hardware toanalyze anomalous situations and togenerate novel command sequences,including (when possible) commandsto repair components diagnosed asfaulty;

• A robust-execution-management com-ponent that uses the Spacecraft Com-mand Language (SCL) software to en-able event-driven processing andlow-level autonomy; and

• The Continuous Activity Scheduling,Planning, Execution, and Replanning(CASPER) program for replanning ac-tivities, including downlink sessions,on the basis of scientific observationsperformed during previous orbit cy-cles.This program was written by Robert Sher-

wood, Russell Knight, Gregg Rabideau, SteveChien, Daniel Tran, Benjamin Cichy, Re-becca Castaño, Timothy Stough, and AshleyDavies of Caltech for NASA’s Jet Propul-sion Laboratory. Further information iscontained in a TSP (see page 1).


Software for Analyzing Laminar-to-Turbulent Flow Transitions

Langley Stability and TransitionAnalysis Codes (LASTRAC) is a set ofengineering software tools developedwith the C++ language and modernsoftware technologies for use in analyz-ing transition from laminar to turbu-lent flows. LASTRAC is a product of on-going NASA Langley research projectsrelated to transition flow physics mod-eling and simulations. It is intended tobe a set of easy-to-use engineering toolsthat can be applied to routine engi-neering design studies. At the currentstage, LASTRAC is capable of perform-ing transition calculations based on lin-ear stability theory (LST) or linear andnonlinear parabolized stability equa-tions (PSE) for a broad range of flowregimes and configurations of interestfor the design of low-speed as well as su-personic and hypersonic vehicles. Atpresent, LASTRAC is limited to two-di-mensional, axisymmetric, or infinite


swept-wing boundary layers. Optionsfor general three-dimensional bound-ary layers are currently under develop-ment. The LST option makes it possibleto perform traditional N-factor transi-tion correlation. Linear and nonlinearPSE are used to track instability waveevolution from small-amplitude till

early transition stage in a high-fidelitymanner. It is planned to incorporatemodules in LASTRAC that models thereceptivity (the process by which per-turbations are introduced into laminarboundary-layer flow) and late stage ofthe transition process. These softwaremodules are intended to enable LAS-

TRAC to perform computations for dif-ferent stages of laminar-to-turbulenttransition in an integrated fashion.

This program was written by Chau-LyanChang of Langley Research Center. Fur-ther information is contained in a TSP (seepage 1).LAR-16260


Materials

Elastomer Filled With Single-Wall Carbon NanotubesStrength and stiffness increase with SWNT content.Lyndon B. Johnson Space Center, Houston, Texas

Experiments have shown that compos-ites of a silicone elastomer with single-wall carbon nanotubes (SWNTs) are sig-nificantly stronger and stiffer than is theunfilled elastomer. The large strength-ening and stiffening effect observed inthese experiments stands in contrast tothe much smaller strengthening effectobserved in related prior efforts to rein-force epoxies with SWNTs and to rein-force a variety of polymers with multiple-wall carbon nanotubes (MWNTs). Therelative largeness of the effect in the caseof the silicone-elastomer/SWNT com-posites appears to be attributable to (1)a better match between the ductility ofthe fibers and the elasticity of the matrixand (2) the greater tensile strengths of

SWNTs, relative to MWNTs.For the experiments, several composites

were formulated by mixing various pro-portions of SWNTs and other filling mate-rials into uncured RTV-560, which is a sili-cone adhesive commonly used inaerospace applications. Specimens of astandard “dog-bone” size and shape fortensile testing were made by casting theuncured elastomer/filler mixtures intomolds, curing the elastomer, then pressingthe specimens from a “cookie-cutter” die.

The results of tensile tests of the spec-imens showed that small percentages ofSWNT filler led to large increases in stiff-ness and tensile strength, and that theseincreases were greater than those af-forded by other fillers. For example, the

incorporation of SWNTs in a proportionof 1 percent increased the tensilestrength by 44 percent and the modulusof elasticity (see figure) by 75 percent.However, the relative magnitudes of theincreases decreased with increasing nan-otube percentages because more nan-otubes made the elastomer/nanotubecomposites more brittle. At an SWNTcontent of 10 percent, the tensilestrength and modulus of elasticity were125 percent and 562 percent, respec-tively, greater than the correspondingvalues for the unfilled elastomer.

This work was done by Bradley S. Files andCraig R. Forest of Johnson Space Center.Further information is contained in a TSP(see page 1). MSC-23301

The Stiffness of a Silicone Elastomer filled with several different kinds and proportions of reinforcing materials was measured in standard tensile tests.

Control Specimen:Unfilled Elastomer

1% SICWhiskers

1% UnpurifiedSWNT,

Manually Mixed

5% GroundCarbon Black

5% SICWhiskers

10% SICWhiskers

5% UnpurifiedSWNT,

Manually Mixed

Composite Specimens Containing Fillers as Noted

10% UnpurifiedSWNT,

Mechanically Mixed

10% GroundCarbon Black

0.89 0.85

1.56

0.931.21

2.05

2.57

5.89

1.00

7

6

5

4

3

2

1

0

Mod

ulus

of E

last

icity

, kps

i


Mechanics

Modifying Ship Air-Wake Vortices for Aircraft OperationsTakeoffs and landings would be safer.Langley Research Center, Hampton, Virginia

Columnar-vortex generators (CVG)have been proposed as means to in-crease the safety of takeoffs and landingsof aircraft on aircraft or helicopter carri-ers and other ships at sea. According tothe proposal, CVGs would be installed atcritical edge locations on ships to modifythe vortices in the air wakes of the ships.The desired effects of modifications areto smooth airflows over takeoff andlanding deck areas and divert vorticesfrom takeoff and landing flight paths.

With respect to aircraft operations,the wake flows of primary interest arethose associated with the bow and sideedges of aircraft-carrier decks and withsuperstructures of ships in general (seeFigure 1). The bow and deck-edge vor-tices can adversely affect airplane andhelicopter operations on carriers, whilethe superstructure wakes can primarilyaffect operations of helicopters.

The concept of the CVG is not new;what is new is the proposed addition ofCVGs to ship structures to effect favorablemodifications of air wakes. Figure 2 de-picts a basic CVG, vertical and horizontalCVGs installed on a simple superstruc-ture, and horizontal CVGs installed onthe bow and deck edges. The verticalCVGs would be closed at the deck butopen at the top. Each horizontal CVG

would be open at both ends. The dimen-sions of the CVGs installed on the aftedges of the superstructure would be cho-sen so that the portion of the flow modi-fied by the vertical CVGs would interactsynergistically with the portion of the flowmodified by the horizontal CVG to movethe air wake away from the takeoff-and-landing zone behind the superstructure.

The deck-edge CVGs would bemounted flush with, and would extendslightly ahead of the bow of, the flightdeck. The overall length of each tubewould exceed that of the flight deck.Each deck-edge CVG would capture thatportion of the airflow that generates adeck-edge vortex and would generate acolumnar vortex of opposite sense tothat of the unmodified vortex. The vor-tex generated by the CVG could be dis-persed at its base, thereby removing un-wanted turbulence in the path of anapproaching airplane. The deck-edge

CVGs would promote smooth flow overthe entire flight deck. In the case of aNimitz-class aircraft carrier like that ofFigure 1, there would be a CVG on eachof the outer edges of the two left por-tions of the flight deck and a single CVGon the right side of the flight deck. Theforwardmost CVG on the left side wouldtake the generated vortex underneaththe angled flight deck.

A CVG could also be installed on thebow of the flight deck to smooth theflow of air onto the flight deck. In thecase of wind incident on the deck froman azimuth other than straight ahead,the vortex generated by the bow CVGcould, perhaps, be used to feed theCVG(s) of the leeward side edge of theflight deck.

This work was done by John E. Lamar ofLangley Research Center. Further infor-mation is contained in a TSP (see page 1).LAR-16281

Figure 1. Air Wakes of ship structures can ad-versely affect operations of aircraft.

Superstructure

Air Wake

HelicopterLanding Pad

RelativeWind

VORTICES GENERATED AT SIDE EDGES OF AIRCRAFT-CARRIER DECK

TYPICAL WAKE FLOW BEHIND A SUPERSTRUCTURE

RelativeWind

Figure 2. Columnar-Vortex Generators would modify air wakes to provide smoother flows and divertvortices from paths of aircraft.

COLUMNAR VORTEX GENERATOR ON BOW OF FLIGHT DECK DIFFUSER ON VERTICAL CVG

HelicopterLanding Pad

COLUMNAR VORTEX GENERATORS ON SIDE EDGES OF AIRCRAFT-CARRIER DECKLeft-Side View

Airflow

CVGs

Airflow

Front View

BASIC COLUMNAR VORTEX GENERATOR

COLUMNAR VORTEX GENERATOR BEHIND SUPERSTRUCTURE

Columnar Vortex

Plate, Vortex Generator ofOpposite Sense, or Cone

Superstructure

CVGs

Relative Wind

RelativeWind

Bow CVG

Flight Deck

Flight Deck


An advanced docking system is under-going development to enable softer, saferdocking than was possible when usingprior docking systems. This system is in-tended for original use in docking of visit-ing spacecraft and berthing the Crew Re-

turn Vehicle at the International SpaceStation (ISS). The system could also beadapted to a variety of other uses in outerspace and on Earth, including mating sub-mersible vehicles, assembling structures,and robotic berthing/handling of pay-

loads and cargo.Heretofore, two large spacecraft have

been docked by causing the spacecraftto approach each other at a speed suffi-cient to activate capture latches — a pro-cedure that results in large docking

Advanced Docking System With Magnetic Initial CaptureSpeeds of approach, and thus docking forces, could be relatively small.Lyndon B. Johnson Space Center, Houston, Texas

A method of determining the amountof fluid in a tank is based on measure-ment of strains induced in tank supportsby the weight of the fluid. Unlike mostprior methods, this method is nonintru-sive: there is no need to insert instru-mentation in the tank and, hence, noneed to run wires, cables, or tubesthrough the tank wall. Also unlike mostprior methods, this method is applicableeven if the fluid in the tank is at super-critical pressure and temperature, be-cause it does not depend on the pres-ence of a liquid/gas interface (as inliquid-level-measuring methods).

The strain gauges used in this methodmay be of two types: foil and fiber-optic.Four foil gauges (full bridge) aremounted on each of the tank-supportinglegs. As the tank is filled or emptied, thedeformation in each leg increases or de-creases, respectively. Measured deforma-tions of all legs are added to obtain acomposite deformation indicative of thechange in weight of the tank plus fluid.An initial calibration is performed byrecording data at two points (usually,empty and full) for which the mass orweight of fluid is known. It is assumedthat the deformations are elastic, so thatthe line passing through the two pointscan be used as a calibration curve ofmass (or weight) of fluid versus defor-mation.

One or more fiber-optic gauges maybe used instead of the foil gauges. Theresolution of the fiber-optic and foilgauges is approximately the same, butthe fiber-optic gauges are immune toEMI (electromagnetic interference), arelinear with respect to temperature overtheir entire dynamic range (as definedby the behavior of the sample), and mea-

sure thermally induced deformations aspredictable signals. Conversely, longterm testing has demonstrated that thefoil gauges exhibit an erratic behaviorwhenever subjected to direct sun radia-tion (even if protected with a rubberizedcover). Henceforth, for deployment inoutdoor conditions, fiber-optic gaugesare the only option if one is to rely onthe system for an extended period oftime when a recalibration proceduremay not be acceptable.

A set of foil gauges had been tested onthe supports of a 500-gallon (1,900-liter)tank. The gauges were found to be capa-ble of measuring the deformations (upto 22 micro-strain) that occurred duringfilling and emptying of the tank. Thefluid masses calculated from the gaugereadings were found to be accuratewithin 4.5 percent. However, the relia-bility of the foil gauges over a few hourswas not acceptable. Therefore, the foilsensor system is acceptable for use onlyin controlled environments (completeshade, or indoors).

The fiber-optic sensors are reliableand at least as accurate as the foil sensors(possibly more). The fiber-optic systemconsists of one or more sensorsmounted on the structure, a tempera-ture sensor also mounted on the struc-ture, and a reference fiber-optic sensormounted on a plate made of the samematerial as the tank-supporting legs, thatis not subjected to any mechanical load.An important element to consider is thethermal deformation of the structure,which is not going to be exactly the sameas that of the reference plate. This is be-cause the structure has appendages(pipes, etc.) that must be taken into ac-count for temperature compensation.

The procedure to calculate a compen-sated measurement (a measurement re-flecting the mass contents of the tank)using the fiber-optic system is as follows:First, one must characterize the struc-ture as it deforms due solely to thermaleffects by taking measurements over itsoperating temperature range, when thetank is empty. During this characteriza-tion procedure, deformation and tem-perature measurements are taken fromthe sensors attached to the structure(strain and temperature) and the sensorattached to the reference plate (strain).With this information, a tabulated orgraphical tool is developed such that atany given temperature, the referencesignal is subtracted from the structuralsignal, and further modified by a valuethat depends on the structure’s temper-ature. Note that the characterization ofthe structure as it deforms due to tem-perature variations may also be done byanalytical methods that can model theprocess. In that case, the experimentalcharacterization is not needed.

It may be possible to increase accuracyfurther by increasing the signal-to-noiseratio through the use of more de-formable tank-supporting legs, largergauges that measure larger deformations,or other methods to increase the strain insome part of the tank support structure.

This work was done by Jorge Figueroa,William St. Cyr, and Shamim Rahman ofStennis Space Center and Gregory McVay,David Van Dyke, William Mitchell, andLester Langford of Lockheed Martin Corp.

Inquiries concerning rights for the commercialuse of this invention should be addressed to theIntellectual Property Manager, Stennis SpaceCenter, (228) 688-1929. Refer to SSC-00187.

Strain-Gauge Measurement of Weight of Fluid in a TankThis method is nonintrusive and independent of the nature of the fluid.Stennis Space Center, Mississippi


loads and is made more difficult becauseof the speed. The basic design and modeof operation of the present advanceddocking system would eliminate theneed to rely on speed of approach to ac-tivate capture latches, thereby making itpossible to reduce approach speed andthus docking loads substantially.

The system would comprise an activesubsystem on one spacecraft and a pas-sive subsystem on another spacecraftwith which the active subsystem will bedocked. The passive subsystem would in-clude an extensible ring containingmagnetic striker plates and guide petals.The active subsystem would include mat-ing guide petals and electromagnetscontaining limit switches and would be

arranged to mate with the magneticstriker plates and guide petals of the pas-sive assembly. The electromagnets wouldbe carried on (but not rigidly attachedto) a structural ring that would be in-strumented with load sensors. The out-puts of the sensors would be sent, alongwith position information, as feedbackto an electronic control subsystem. Thesystem would also include electro-mechanical actuators that would extendor retract the ring upon command bythe control subsystem.

In preparation for docking, one space-craft would move to a position near (butnot touching) the other spacecraft, withthe docking ports of the two spacecraft inapproximate alignment. Then while one

spacecraft maintained an approximatelyconstant position relative to the otherspacecraft, the actuators of the active sub-system would be made to extend the ring,gently pushing the guide petals and elec-tromagnets toward the passive ring guidepetals and magnetic striker plates: in ef-fect, the active subsystem would reach out,comply, and grab the passive subsystem.

During this reaching out, the hard-ware and software of the feedback con-trol subsystem would command the ac-tuators to respond to sensed loads tocorrect for any misalignments betweenthe docking ports, i.e., to comply. Thereaching-out-and-alignment processwould continue until the limit switchesindicated soft capture — i.e., final petalalignment and magnetic capture of themagnetic striker plates. Once soft cap-ture and alignment was complete, thering would be retracted, then mechani-cal latches would be engaged to securethe docked spacecraft to each other.

The active subsystem ring, electro-magnets, and petals would then be with-drawn, and the latches would continueto hold the spacecraft together. Later,the undocking could be effected by re-leasing the mechanical latches.

This work was done by James L. Lewis ofJohnson Space Center and Monty B. Car-roll, Ray Morales, and Thang Le of LockheedMartin.

This invention has been patented by NASA(U.S. Patent No. 6,354,540). Inquiries con-cerning nonexclusive or exclusive license forits commercial development should be ad-dressed to the Patent Counsel, Johnson SpaceCenter, (281) 483-0837. Refer to MSC-22931.

This Graphical Representation Depicts a Demonstration Version of the advanced docking system foruse in berthing an X-38 spacecraft at the International Space Station.

Soft-Capture, Load-Sensing Ring

Guide Petals

Striker Plates

Magnets

TunnelMechanical Latching System

Electromechanical Actuators

Cross-Sectional View

Androgynous Low-Impact Docking System


Machinery/Automation

Blade-Pitch Control for Quieting Tilt-Rotor AircraftActively induced harmonic blade-pitch oscillations reduce BVI noise.Ames Research Center, Moffett Field, California

A method of reducing the noise gener-ated by a tilt-rotor aircraft during descentinvolves active control of the blade pitchof the rotors. This method is related toprior such noise-reduction methods, of atype denoted generally as higher-har-monic control (HHC), in which the bladepitch is made to oscillate at a harmonic ofthe frequency of rotation of the rotor.

A tilt-rotor aircraft is so named be-cause mounted at its wing tips are motorsthat can be pivoted to enable the aircraftto take off and land like a helicopter orto fly like a propeller airplane. When theaircraft is operating in its helicoptermode, the rotors generate more thrustper unit rotor-disk area than helicopterrotors do, thus producing more blade-vortex interaction (BVI) noise. BVI is amajor source of noise produced by heli-copters and tilt-rotor aircraft during de-scent: When a rotor descends into itsown wake, the interaction of each bladewith the blade-tip vortices generated pre-viously gives rise to large air-pressurefluctuations. These pressure fluctuationsradiate as distinct, impulsive noise.

In general, the pitch angle of therotor blades of a tilt-rotor aircraft is con-trolled by use of a swash plate connectedto the rotor blades by pitch links. Inboth prior HHC methods and the pre-sent method, HHC control signals arefed as input to swash-plate control actua-tors, causing the rotor-blade pitch to os-cillate. The amplitude, frequency, andphase of the control signal can be cho-sen to minimize BVI noise.

In the present method, one typically,chooses a control waveform that causesthe blade pitch to oscillate sinusoidallyat the N+1 harmonic of the rotation fre-quency (where N is the number ofblades on each rotor). The phase of theoscillation is typically chosen such thatthe minimum pitch angle occurs at afixed rotor-blade azimuth angle withinthe range from 60° to 90° (where 0° az-imuth is defined as directly aft). Thephase is critical, but the amplitude isnot: It has been found that pitch-angleoscillation amplitudes up to approxi-mately 0.7° can result in large reduc-tions of noise. Larger pitch amplitudes

can result in larger reductions of noise,but one might prefer to avoid them be-cause they are accompanied by increasesin control loads.

Prior efforts to exploit the HHC con-cept to reduce helicopter vibration andnoise have been oriented toward the de-velopment of complex, closed-loop con-trol systems that would utilize feedbackfrom sensors and that would implementiterative control algorithms to adjustHHC settings to optimize responses overthe full ranges of operating conditions.The development of such systems can beexpensive and time-consuming. In con-trast, a system according to the presentmethod is relatively simple because it isan open-loop system. By selecting a sin-gle harmonic (the N+1) and fixing theamplitude, one can reduce the problemof choosing the control signal to one ofselecting a single open-loop input signal(the phase signal), which can be opti-mized for several descending flight con-ditions. In other words, HHC phase val-ues can be predetermined andscheduled for specific flight conditions.The schedule of phase values can be im-plemented by use of control softwareand hardware.

The pilot can turn the HHC system onor off by means of a switch (see figure).Ordinarily, the HHC system would beused only when reduction of noise wasdesired during descent in the helicoptermode. Because the time spent in use ofthe HHC system would ordinarily be asmall fraction of the total operationaltime of the aircraft, the control loads as-sociated with use of the HHC systemcould be expected to cause little, if anyreduction, in the lifetime of the me-chanical components of the pitch-con-trol system. Because the HHC noise-re-duction system is not needed for safeoperation of the aircraft, it is a fail-safesystem: The system can be switched off atany time without adversely affectingflight, the only penalty being an increasein noise.

This work was done by Mark D. Betzinaand Khanh Q. Nguyen of Ames ResearchCenter. Further information is contained ina TSP (see page 1).

This invention is owned by NASA, and apatent application has been filed. Inquiriesconcerning nonexclusive or exclusive licensefor its commercial development should be ad-dressed to the Patent Counsel, Ames ResearchCenter, (650) 604-5104. Refer to ARC-14606.

This Graph Demonstrates the Effect of HHC on acoustic time history at peak directivity.

Acoustic Time History at Peak Noise Location

-40

-20

0

20

40

60

80

100

0 120 180 240 300 360

HHC Off

HHC On

Mic

roph

one

Out

put (

Pas

cals

)

Master Blade Azimuth (Degrees)

60


Solar Array Panels With Dust-Removal CapabilityInexpensive, low-power piezoelectric buzzers would be built in.NASA’s Jet Propulsion Laboratory, Pasadena, California

It has been proposed to incorporatepiezoelectric vibrational actuators intothe structural supports of solar photo-voltaic panels, for the purpose of occa-sionally inducing vibrations in the panelsin order to loosen accumulated dust. Pro-vided that the panels were tilted, the loos-ened dust would slide off under its ownweight. Originally aimed at preventing ob-scuration of photovoltaic cells by dust ac-cumulating in the Martian environment,the proposal may also offer an option forthe design of solar photovoltaic panels forunattended operation at remote locationson Earth.

The figure depicts a typical lightweightsolar photovoltaic panel comprising abackside grid of structural spars that sup-port a thin face sheet that, in turn, sup-ports an array of photovoltaic cells onthe front side. The backside structure in-cludes node points where several sparsintersect. According to the proposal,piezoelectric buzzers would be attachedto the node points. The process of de-signing the panel would be an iterativeone that would include computationalsimulation of the vibrations by use of fi-nite-element analysis to guide the selec-tion of the vibrational frequency of theactuators and the cross sections of thespars to maximize the agitation of dust.

Although the basic concept of theproposal is a straightforward extensionof a common household cleaning prac-tice, the engineering implementation ofthe proposal would not be trivial. Thefollowing are some of the engineering is-sues that must be addressed:

• Compact, low-power, inexpensive piezo-electric buzzers are commercially avail-able. They may or may not be suitablefor use as the piezoelectric actuators toimplement the proposal. Because typi-cal commercial buzzers operate in thekilohertz frequency range and the nat-ural vibrational frequencies of typicalsolar photovoltaic panels are lower, itmay be necessary to build lower-fre-quency piezoelectric buzzers.

• It may be necessary to cover panels withflat, transparent sheets or else redesignthe panels to eliminate recesses or pro-trusions that could retain dust or preventdust from sliding off during vibration.

• The expected rate of accumulation ofdust must be taken into account in as-sessing the effectiveness of a dust-re-moval design.

• Tests must be performed to determinethe interdependences among tilt an-gles required for interception of solarradiation, the amounts of agitation re-

quired at various vibrational frequen-cies and amplitudes to reduce obscura-tion by dust to an acceptably low levelat those tilt angles, and the differencesin among the rates of removal of dustparticles of different sizes and types.

• Care must be taken to ensure that theenergy recovered by removing dustthat obscures the solar photovoltaicpanel exceeds the energy expended inshaking the dust off. This entails con-sideration of buzzer power levels andagitation times.

• Care must also be taken to ensure thatthe dust-removal design does not ad-versely affect equipment other thanthe solar photovoltaic panel.This work was done by Stephen Dawson,

Nick Mardesich, Brian Spence, and SteveWhite of Caltech for NASA’s Jet PropulsionLaboratory. Further information is con-tained in a TSP (see page 1).NPO-30909

Piezoelectric Buzzers would be mounted at nodes of a grid of spars that support a solar photovoltaic panel.

Piezoelectric Buzzers Back-Side Spars

Back Front

Face Sheet Photovoltaic Cells

BACK VIEW LENGTHWISE VIEW


Manufacturing

Aligning Arrays of Lenses and Single-Mode Optical FibersA procedure for precise alignment involves the use of an interferometer and positioning stages.NASA’s Jet Propulsion Laboratory, Pasadena, California

A procedure now under developmentis intended to enable the precise align-ment of sheet arrays of microscopiclenses with the end faces of a coherentbundle of as many as 1,000 single-modeoptical fibers packed closely in a regulararray (see Figure 1). In the original ap-plication that prompted this develop-ment, the precise assembly of lenses andoptical fibers serves as a single-mode spa-tial filter for a visible-light nulling inter-ferometer. The precision of alignmentmust be sufficient to limit any remainingwavefront error to a root-mean-square

value of less than 1/10 of a wavelength oflight. This wavefront-error limit translatesto requirements to (1) ensure uniformityof both the lens and fiber arrays, (2) en-sure that the lateral distance from thecentral axis of each lens and the corre-sponding optical fiber is no more than afraction of a micron, (3) angularly alignthe lens-sheet planes and the fiber-bun-dle end faces to within a few arc seconds,and (4) axially align the lenses and thefiber-bundle end faces to within tens ofmicrons of the focal distance.

Figure 2 depicts the apparatus used in

the alignment procedure. The beam oflight from a Zygo (or equivalent) inter-ferometer is first compressed by a ratio of20:1 so that upon its return to the inter-ferometer, the beam will be magnifiedenough to enable measurement of wave-front quality. The apparatus includesrelay lenses that enable imaging of thearrays of microscopic lenses in a charge-coupled-device (CCD) camera that ispart of the interferometer. One of the ar-rays of microscopic lenses is mounted ona 6-axis stage, in proximity to the frontface of the bundle of optical fibers. Thebundle is mounted on a separate stage. Amirror is attached to the back face of thebundle of optical fibers for retroreflec-tion of light. When a microscopic lensand a fiber are aligned with each other,the affected portion of the light is re-flected back by the mirror, recollimatedby the microscopic lens, transmittedthrough the relay lenses and the beamcompressor/expander, then split sothat half goes to a detector and half tothe interferometer. The output of the de-tector is used as a feedback control signalfor the six-axis stage to effect alignment.

The alignment procedure, which israther complex, can be summarized asfollows:1. In the absence of a sheet array of mi-

croscopic lenses, the longitudinal axis

Figure 1. The Problem Is To Align two sheet arrays of microscopic lenses with polished end faces of abundle of optical fibers. The assembly of lenses and fibers is meant to act as a spatial filter.

Figure 2. An Interferometer is used along with a six-axis positioning stage (and other positioning stages omitted from this view for clarity) to measure andcorrect the relative position and orientation of the bundle of optical fibers and the sheet array of microscopic lenses.

Sheet Arrays of Microscopic Lenses

Spatially Filtered

Wavefront

Incident Wavefront Spatially Modulated at High Frequency

by Defects in Preceding Optics

Bundle of Optical Fibers

Computer

Interferometer

Detector

Iris at Imaging Plane of Interferometer

Beam Splitter Beam Compressor/Expander

Assembly ofRelay Lenses

Six-Axis Positioning Stage Stage Controller

RetroreflectingMirror

Sheet Array of Microscopic Lenses

Ferrule Containing Bundle of

Optical Fibers


of the fiber-optic bundle is alignedwith the optical axis of the interfer-ometer by use of the reflection fromthe front face of the bundle.

2. One of the sheet arrays of micro-scopic lenses is placed in front of thefiber-optic bundle and similarlyaligned with the interferometer opti-cal axis by use of the reflection fromits front face. As a result, the opticalaxes of the lens array and the fiber-optic bundle are parallel with eachother.

3. The axial position of the lens sheet isadjusted until the interferometricimage of light reflected from the frontface of the fiber-optic bundle indi-cates that the lenses are at the properfocal distance.

4. The lateral (relative to the opticalaxis) position of the lens sheet is ad-justed until the interferometricimage shows that at least one lens is

centered on the end of at least oneoptical fiber. The lateral coordinatesof the six-axis positioner are mea-sured. The lateral position of the lenssheet is further adjusted until an-other lens/fiber pair is thus centered,and the corresponding coordinatesare measured. The two sets of coordi-nates are used to compute the trans-lation and rotation needed to effectthe lateral alignment of the remain-ing lens/fiber pairs.

5. Guided by the foregoing coordinatemeasurements, the final adjustmentsof the lens sheet are made.

6. The lens sheet is bonded to the fiber-optic bundle.

7. The fiber-optic bundle is turnedaround so that what was previously theback face is now the front face.

8. The retroreflecting mirror is alignedwith the optical axis of the interfer-ometer.

9. Steps 1 through 7 are repeated to ef-fect the alignment and bonding of thesecond lens sheet to what is now thefront face of the fiber-optic bundle.This work was done by Duncan Liu of

Caltech for NASA’s Jet Propulsion Labo-ratory. Further information is contained ina TSP (see page 1).

In accordance with Public Law 96-517,the contractor has elected to retain title to thisinvention. Inquiries concerning rights for itscommercial use should be addressed to:

Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099(818) 354-2240E-mail: [email protected] to NPO-40021, volume and number

of this NASA Tech Briefs issue, and thepage number.

Automatic Control of Arc Process for Making CarbonNanotubesLyndon B. Johnson Space Center, Houston, Texas

An automatic-control system has beendevised for a process in which carbonnanotubes are produced in an arc be-tween a catalyst-filled carbon anode anda graphite cathode. The control systemincludes a motor-driven screw that ad-justs the distance between the elec-trodes. The system also includes abridge circuit that puts out a voltageproportional to the difference between(1) the actual value of potential dropacross the arc and (2) a reference value

between 38 and 40 V (corresponding toa current of about 100 A) at which theyield of carbon nanotubes is maximized.Utilizing the fact that the potential dropacross the arc increases with the inter-electrode gap, the output of the bridgecircuit is fed to a motor-control circuitthat causes the motor to move theanode toward or away from the cathodeif the actual potential drop is more orless, respectively, than the reference po-tential. Thus, the system regulates the

interelectrode gap to maintain the opti-mum potential drop. The system also in-cludes circuitry that records the poten-tial drop across the arc and the relativeposition of the anode holder as functionof time.

This work was done by Carl D. Scott ofJohnson Space Center, Robert B. Pulum-barit of Lockheed Martin, and Joe Victor ofHernandez Engineering. Further informationis contained in a TSP (see page 1).MSC-23134

Curved-Focal-Plane Arrays Using Deformed-MembranePhotodetectorsIt would not be necessary to perform fabrication processing of curved substrates.NASA’s Jet Propulsion Laboratory, Pasadena, California

A versatile and simple approach to thedesign and fabrication of curved-focal-plane arrays of silicon-based photodetec-tors is being developed. This approach isan alternative to the one described in“Curved Focal-Plane Arrays Using Back-Illuminated High-Purity Photodetectors”(NPO-30566), NASA Tech Briefs, Vol. 27,No. 10 (October 2003), page 10a.

As in the cited prior article, the basicidea is to improve the performance ofan imaging instrument and simplify theoptics needed to obtain a given level ofperformance by making an image sensor(in this case, an array of photodetectors)conform to a curved focal surface, in-stead of designing the optics to projectan image onto a flat focal surface. There

is biological precedent for curved-focal-surface designs: retinas — the imagesensors in eyes — conform to the natu-rally curved focal surfaces of eye lenses.

The present approach is applicable toboth front-side- and back-side-illumi-nated, membrane photodetector arraysand is being demonstrated on charge-coupled devices (CCDs). The very-large-

scale integrated (VLSI) circuitry of such aCCD or other array is fabricated on thefront side of a silicon substrate, then theCCD substrate is attached temporarily to asecond substrate for mechanical support,then material is removed from the back toobtain the CCD membrane, which typi-cally has a thickness between 10 and 20 µm. In the case of a CCD designed tooperate in back-surface illumination, deltadoping can be performed after thinning to enhance the sensitivity. This approachis independent of the design and methodof fabrication of the front-side VLSI cir-cuitry and does not involve any processingof a curved silicon substrate.

In this approach, a third substratewould be prepared by polishing one ofits surfaces to a required focal-surfacecurvature. A CCD membrane fabricatedas described above would be pressedagainst, deformed into conformity with,and bonded to, the curved surface. Thetechnique used to press and bond theCCD membrane would depend on thenature of the supporting material (seefigure). For example, if the third sub-strate were made of quartz frit, the sub-strate would be prepared by suffusing itwith epoxy. Then one would take advan-tage of the porosity of the frit by apply-ing a partial vacuum to the opposite sur-

face of the frit, causing atmosphericpressure to push the CCD membraneagainst the curved surface. The curingof the epoxy would bond the CCD mem-brane to the curved surface.

Alternatively, if the third substrate weremade of a nonporous material, thecurved substrate surface would be pre-pared by coating it with a wax or an un-cured epoxy. The CCD membrane wouldbe pressed against the coated, curved sur-face by use of a suitably pressurized bal-loon. The CCD membrane would thenbecome bonded to the curved surface bycuring of the epoxy or freezing of the wax.

This work was done by Shouleh Nikzadand Todd Jones of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).

In accordance with Public Law 96-517,the contractor has elected to retain title to thisinvention. Inquiries concerning rights for itscommercial use should be addressed to

Intellectual Assets OfficeJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109(818) 354-2240E-mail: [email protected] to NPO-30580, volume and number

of this NASA Tech Briefs issue, and thepage number.


A Flat Membrane CCD would be pressed against, and bonded to, a curved substrate in either of two ways.

Porous Substrate(Frit)

Vacuum

POROUS SUBSTRATE NONPOROUS SUBSTRATE

NonporousSubstrate

Epoxy or Wax

CCD Membrane

Pressure

Balloon MembraneEpoxy

CCDMembrane

Gasket


Physical Sciences

Role of Meteorology in Flights of a Solar-Powered AirplaneMeteorological support helped ensure safety and success of experimental high-altitude flights.Dryden Flight Research Center, Edwards, California

In the summer of 2001, the Helios pro-totype solar-powered uninhabited aerialvehicle (UAV) [a lightweight, remotelypiloted airplane] was deployed to the Pa-cific Missile Range Facility (PMRF), atKauai, Hawaii, in an attempt to fly to alti-tudes above 100,000 ft (30.48 km). Thegoal of flying a UAV to such high alti-tudes has been designated a level-I mile-stone of the NASA Environmental Re-search Aircraft and Sensor Technology(ERAST) program. In support of thisgoal, meteorologists from NASA DrydenFlight Research Center were sent toPMRF, as part of the flight crew, to pro-vide current and forecast weather infor-mation to the pilots, mission directors,and planners. Information of this kind isneeded to optimize flight conditions forpeak aircraft performance and to enableavoidance of weather conditions thatcould adversely affect safety.

In general, the primary weather dataof concern for ground and flight opera-tions are wind speeds (see Figure 1). Be-cause of its long wing span [247 ft (≈75m)] and low weight [1,500 to 1,600 lb(about 680 to 726 kg)], the Helios air-plane is sensitive to wind speeds exceed-ing 7 kn (3.6 m/s) at the surface. Also,clouds are of concern because they canblock sunlight needed to energize anarray of solar photovoltaic cells that pro-vide power to the airplane. Vertical windshear is very closely monitored in orderto prevent damage or loss of control dueto turbulence.

Two flights were successfully com-pleted during the deployment at PMRF(see Figure 2). The sequence of meteo-rological activities in support of eachflight included the following:• Daily forecasts of surface and upper-

level meteorological conditions wereissued, 48 hours before the plannedflight day.

• Current and forecast weather condi-tions were described at briefings of thecrew.

• A weather briefing was given in earlymorning on the planned flight day tohelp determine whether the airplaneshould be taken out of its hangar andprepared for flight.

Figure 1. These Surface-Wind Histories were recorded at PMRF during intervals that included twoflights. Data like these, plus other data, are needed to increase the likelihood of safe and successfulflight.

Figure 2. The Takeoff of the Helios Prototype Solar-Powered UAV was delayed because of clouds. Theairplane then took off and flew to a record altitude.

00

2

4

6

8

10

5 10 15Time (Hours) After Midnight of the Day Flight Began

Win

d S

peed

, Kno

ts

20 25 30

July 14-15August 13-14

Clouds as Seen From Runway After Sunrise

Helios Prototype Airplane at Takeoff on its Way to a Record Altitude on August 13, 2001


• A final such “go/no-go” briefing wasgiven 2 hours prior to scheduled take-off.

• After takeoff, periodic updates basedof weather-balloon and satellite datawere provided to the pilot and missionplanner.

• Approximately 2 hours before landing,a final weather forecast was issued toenable estimation of the earliest possi-ble landing time and selection of arunway.

• After landing, surface conditions weremonitored until the airplane was safelyin the hangar.

The first successful flight took placeon July 14, 2001. The takeoff was de-layed for 20 minutes because of clouds.Convection over the runway generatedmoderate turbulence during takeoff.The airplane climbed to a maximum al-titude of 76,500 ft (≈23.3 km). The air-plane landed in stable conditions aftermore than 15 hours of flight.

The second successful flight tookplace on August 13, 2001. This time, thetakeoff was delayed 45 minutes becauseof clouds. Strong wind shear due tostrong trade winds and island wake wasobserved at an altitude of 2,000 ft (≈600

m). The airplane then climbed until itreached a world-record altitude for anon-rocket-powered aircraft — 96,863 ft(29,524 m). This altitude is more than11,000 ft (≈3.35 km) higher than therecord set in a flight of the SR-71 air-plane. The airplane landed safely after alast-minute change in runway because ofwinds.

This work was done by Casey Donohue ofAS&M, Inc., for Dryden Flight ResearchCenter. For further information, contact theDryden Commercial Technology Office at(661) 276-3689.DRC-02-25

Model of Mixing Layer With Multicomponent Evaporating DropsEffects of multiple chemical components are represented with computational efficiency.NASA’s Jet Propulsion Laboratory, Pasadena, California

A mathematical model of a three-di-mensional mixing layer laden with evap-orating fuel drops composed of manychemical species has been derived. Thestudy is motivated by the fact that typicalreal petroleum fuels contain hundredsof chemical species. Previously, for thesake of computational efficiency, spraystudies were performed using eithermodels based on a single representativespecies or models based on surrogatefuels of at most 15 species. The presentmulticomponent model makes it possi-ble to perform more realistic simula-tions by accounting for hundreds ofchemical species in a computationally ef-

ficient manner.The model is used to perform Direct

Numerical Simulations in continuingstudies directed toward understandingthe behavior of liquid petroleum fuelsprays. The model includes governingequations formulated in an Eulerian anda Lagrangian reference frame for the gasand the drops, respectively. This repre-sentation is consistent with the expectedvolumetrically small loading of the dropsin gas (of the order of 10–3), although themass loading can be substantial becauseof the high ratio (of the order of 103) be-tween the densities of liquid and gas. Thedrops are treated as point sources of

mass, momentum, and energy; this rep-resentation is consistent with the dropsize being smaller than the Kolmogorovscale. Unsteady drag, added-mass effects,Basset history forces, and collisions be-tween the drops are neglected, and thegas is assumed calorically perfect.

The model incorporates the concept ofcontinuous thermodynamics, according towhich the chemical composition of a fuel isdescribed probabilistically, by use of a dis-tribution function. Distribution functionsgenerally depend on many parameters.However, for mixtures of homologousspecies, the distribution can be approxi-mated with acceptable accuracy as a solefunction of the molecular weight. The mix-ing layer is initially laden with drops in itslower stream, and the drops are colderthan the gas. Drop evaporation leads to achange in the gas-phase composition,which, like the composition of the drops, isdescribed in a probabilistic manner.

The advantage of the probabilistic de-scription is that while a wide range of in-dividual species can be accommodatedin the mixture, the number of governingequations is increased minimally overthat necessary for a single species be-cause the composition is representedonly by the parameter(s) necessary toconstruct the distribution function.Here the distribution function is entirelydefined by the mean and variance. Forthis choice of distribution function, themodel accounts for evaporation-inducedchanges in the composition of fueldrops and the surrounding gas, yet in-volves only two more conservation equa-tions (one for the mean and one for the

Figure 1. The Evolution of Residual Droplet Areas was computed for the single-component and mul-ticomponent cases. Here t* is time in units of a characteristic time calculated from initial parametersof the mixing layer, Do is the initial drop diameter, D is the drop diameter as a function of time, and<< >> denotes an ensemble average.

0

0.6

0.5

0.7

0.8

0.9

1.0

20

<<D

2/D

2 0>>

40t*

60 80

Multicomponent

Single-Component


Figure 2. Drop Number Density is shown in a representative plane at the final computation time: (a) single-component drop layer and (b) multicomponentdrop layer.

–15

–10

–5

0

0 10 20

5

10

15

X1/δω,0

X2/

δ ω,0

–15

–10

–5

0

0 10 20

5

10

15

X1/δω,0

X2/

δ ω,0

ρn

1.5E+101.0E+108.0E+095.0E+093.0E+091.5E+091.4E+09

ρn

3.5E+101.5E+101.0E+105.0E+093.0E+091.6E+09

variance) than does an equivalent modelfor a single-component fuel. The initialmathematical form of the distributionfunction is postulated to be retainedduring the drop lifetime, but with evolv-ing mean and variance as the dropsevaporate.

In a test, a mixing-layer simulationwas performed for drops of single-com-ponent-fuel and another such simula-tion for drops of a multicomponentfuel. Analysis of the results revealedthat although the global layer charac-teristics were similar in the single-com-ponent and multicomponent cases, thedrops evaporated more slowly in the

multicomponent than in the single-component case (see Figure 1). Theslower evaporation of the multicompo-nent drops was primarily attributed tothe lower volatility of higher molar-weight species and to condensation ofthese species on drops transported inregions of different gas composition.The more volatile species released inthe gas phase earlier during the droplifetime were found to be entrained inthe mixing layer, whereas the heavierspecies that evaporated later during thedrop lifetime tended to reside in re-gions of high drop-number density.This behavior was found to lead to seg-

regation of species in the gas phase onthe basis of the relative evaporationtime from the drops. The slower evapo-ration of multicomponent fuel dropswas found to lead to regions of higherdrop-number density in the drop-ladenlayer and to permit greater interactionof the drops with the flow, resulting in amore developed small-scale structure(see Figure 2).

This work was done by Josette Bellan andPatrick Le Clercq of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).NPO-30505

Solution-Assisted Optical ContactingComponents in optical contact can be adjusted for about a minute.NASA’s Jet Propulsion Laboratory, Pasadena, California

A modified version of a conventionaloptical-contact procedure has beenfound to facilitate alignment of opticalcomponents. The optical-contact proce-dure (called simply “optical contacting”in the art) is a standard means of bond-ing two highly polished and cleanedglass optical components without usingepoxies or other adhesives. In its un-modified form, the procedure does notinvolve the use of any foreign substancesat all: components to be optically con-tacted are dry. The main disadvantage ofconventional optical contacting is that itis difficult or impossible to adjust the

alignment of the components once theyhave become bonded.

In the modified version of the proce-dure, a drop of an alcohol-based opticalcleaning solution (isopropyl alcohol orsimilar) is placed at the interface betweentwo components immediately beforeputting the components together. The so-lution forms a weak bond that graduallystrengthens during a time interval of theorder of tens of seconds as the alcoholevaporates. While the solution is present,the components can be slid, without lossof contact, to perform fine adjustmentsof their relative positions.

After about a minute, most of the al-cohol has evaporated and the opticalcomponents are rigidly attached to eachother. If necessary, more solution can beadded to enable resumption or repeti-tion of the adjustment until the compo-nents are aligned to the required preci-sion.

This work was done by Daniel Shaddockand Alexander Abramovici of Caltech forNASA’s Jet Propulsion Laboratory. Fur-ther information is contained in a TSP (seepage 1).NPO-30731


Information Sciences

Improved Discrete Approximation of Laplacian of GaussianThis method reduces the amount of circuitry needed for filtering of video data.Lyndon B. Johnson Space Center, Houston, Texas

An improved method of computing adiscrete approximation of the Laplacianof a Gaussian convolution of an imagehas been devised. The primary advan-tage of the method is that without sub-stantially degrading the accuracy of theend result, it reduces the amount of in-formation that must be processed andthus reduces the amount of circuitryneeded to perform the Laplacian-of-Gaussian (LOG) operation.

Some background information is nec-essary to place the method in context.The method is intended for applicationto the LOG part of a process of real-timedigital filtering of digitized video datathat represent brightnesses in pixels in asquare array. The particular filteringprocess of interest is one that convertspixel brightnesses to binary form,thereby reducing the amount of infor-mation that must be performed in sub-sequent correlation processing (e.g.,correlations between images in a stereo-scopic pair for determining distances orcorrelations between successive framesof the same image for detecting mo-tions). The Laplacian is often includedin the filtering process because it em-phasizes edges and textures, while theGaussian is often included because itsmooths out noise that might not beconsistent between left and right images

or between successive frames of thesame image.

For the purpose of processing digi-tized image data, the Gaussian andLaplacian values of a pixel of interest areapproximated as weighted sums ofbrightnesses of the pixels in a squaresubarray (typically 3 × 3) of pixels cen-tered on the pixel of interest. Theweights are represented by coefficientmatrices, the elements of which corre-spond to pixels in the subarray. For ex-ample,

gives a reasonable approximation of theLaplacian for a 3 × 3 subarray.

A typical prior state-of-the-art LOG algo-rithm operates in a sequential raster-ori-ented pixel stream, and the kernel is fac-tored as much as possible into 1 × 3 and 3× 1 components. To apply a 1 × 3 or a 3 ×3 kernel to the pixel in a given row andcolumn, it is necessary to retain the pixelstream in two raster-length delay lines sothat the pixels in the adjacent rows andcolumns remain available for the compu-tation. Heretofore, 12 bits of precisionhave been needed to maintain accuracysufficient for reasonable convolution re-

sults through several stages. Usually, theLaplacian operation is performed first tonormalize the image data about 0. Thisconcludes the background information.

In the present method, intermediateresults are approximated in such a waythat only 6 or possibly even as few as fourbits are retained, yet the final result isstill reasonable. If only 6 bits of precisionare needed rather than 12, the size ofthe memory circuitry needed to imple-ment the delay lines can be halved.Thus, it should be possible to buildsmaller, lower-power filtering circuits.

Heretofore, it has been common prac-tice in limited-precision arithmetic cir-cuitry to approximate large values bytruncating them, eliminating the leastsignificant bits. However, a detailedanalysis of the arithmetic process showsthat eliminating the bits of lowest ordercan lead to errors in the final conversionto binary representation. In the presentmethod, the least significant bits are re-tained and large values are approxi-mated by a saturation technique basedon the unconventional approach of dis-carding the highest-order bits. If themagnitude of a pixel value is larger thanthe largest magnitude that can be repre-sented in the result, then the pixel valueis replaced by a value of the same sign(positive or negative) and the largestrepresentable magnitude.

The figure depicts an example of a cir-cuit that utilizes 2’s-complement encod-ing of negative numbers and that imple-ments saturation of a six-bit quantity to afour-bit quantity. The exclusive-OR gatesexamine the number of high-order bitsto be eliminated, plus one, to determinewhether they are already all alike. If theexamined input bits are not all alike, amaximum positive or negative quantity isderived from the sign bit and gatedthrough the multiplexer. This is easily ex-tended to handle a reduction of anynumber of bits.

This work was done by Robert L. Shuler, Jr.,of Johnson Space Center. Further infor-mation is contained in a TSP (see page 1).MSC-22954

0 1 0

1 4 1

0 1 0

−− −

−

i0

i1

i2

i3

i4

i5

Input Bits

Least Significant Bit

Multiplexer

Output

A

B A if s = 0B if s = 1

Most Significant Bit

The Input Bits Are Passed through to the multiplexer as long as the magnitude of the input quantitydoes not exceed a preset four-bit saturation value; if it does exceed that value, then the output hasthe saturation magnitude and the same sign as that of the input.


Utilizing Expert Knowledge in Estimating Future STS CostsJohn F. Kennedy Space Center, Florida

A method of estimating the costs of fu-ture space transportation systems (STSs)involves classical activity-based cost(ABC) modeling combined with system-atic utilization of the knowledge andopinions of experts to extend theprocess-flow knowledge of existing sys-tems to systems that involve new materialsand/or new architectures. The expertknowledge is particularly helpful in fillinggaps that arise in computational modelsof processes because of inconsistencies inhistorical cost data. Heretofore, the costsof planned STSs have been estimated fol-lowing a “top-down” approach that tends

to force the architectures of new systemsto incorporate process flows like those ofthe space shuttles. In this ABC-basedmethod, one makes assumptions aboutthe processes, but otherwise follows a“bottoms up” approach that does notforce the new system architecture to in-corporate a space-shuttle-like processflow. Prototype software has been devel-oped to implement this method.Through further development of soft-ware, it should be possible to extend themethod beyond the space program to al-most any setting in which there is a needto estimate the costs of a new system and

to extend the applicable knowledge basein order to make the estimate.

This work was done by David B. Fortnerof Command and Control Technologies,Inc., and Alex J. Ruiz-Torres of the Univer-sity of Texas at El Paso for KennedySpace Center. For further information,contact

Kevin BrownCommand and Control Technologies1425 Chaffee Drive, Suite 1Titusville, FL 32780(321) 264-1193E-mail: [email protected]

KSC-12512


Study of Rapid-RegressionLiquefying Hybrid RocketFuels

A report describes experiments di-rected toward the development of paraf-fin-based hybrid rocket fuels that burn atregression rates greater than those of con-ventional hybrid rocket fuels like hydroxyl-terminated butadiene. The basic ap-proach followed in this development is touse materials such that a hydrodynami-cally unstable liquid layer forms on themelting surface of a burning fuel body. En-trainment of droplets from the liquid/gasinterface can substantially increase therate of fuel mass transfer, leading to sur-face regression faster than can be achievedusing conventional fuels. The higher re-gression rate eliminates the need for thecomplex multi-port grain structures ofconventional solid rocket fuels, making itpossible to obtain acceptable perfor-mance from single-port structures. Thehigh-regression-rate fuels contain no toxicor otherwise hazardous components andcan be shipped commercially as non-haz-ardous commodities. Among the experi-

ments performed on these fuels werescale-up tests using gaseous oxygen. Thedata from these tests were found to agreewith data from small-scale, low-pressureand low-mass-flux laboratory tests and toconfirm the expectation that these fuelswould burn at high regression rates, cham-ber pressures, and mass fluxes representa-tive of full-scale rocket motors.

This work was done by Greg Zilliac andShane DeZilwa of Ames Research Centerand M. Arif Karabeyoglu, Brian J. Cantwell,and Paul Castellucci of Stanford University.Further information is contained in a TSP(see page 1).

Inquiries concerning rights for the commer-cial use of this invention should be addressedto the Patent Counsel, Ames Research Center,(650) 604-5104. Refer to ARC-14486-2.

More About the Phase-Synchronized En-hancement Method

A report presents further details regard-ing the subject matter of “Phase-Synchro-nized Enhancement Method for EngineDiagnostics” (MFS-26435), NASA Tech

Briefs, Vol. 22, No. 1 (January 1998), page54. To recapitulate: The phase-synchro-nized enhancement method (PSEM) in-volves the digital resampling of a quasi-pe-riodic signal in synchronism with theinstantaneous phase of one of its spectralcomponents. This resampling transformsthe quasi-periodic signal into a periodicone more amenable to analysis. It is par-ticularly useful for diagnosis of a rotatingmachine through analysis of vibrationspectra that include components at thefundamental and harmonics of a slightlyfluctuating rotation frequency. The reportdiscusses the machinery-signal-analysisproblem, outlines the PSEM algorithms,presents the mathematical basis of thePSEM, and presents examples of applica-tion of the PSEM in some computationalsimulations.

This work was done by Jen-Yi Jong of AISignal Research, Inc., for Marshall SpaceFlight Center. For further information, con-tact the company at [email protected] or(256) 551-0008.MFS-31409

Books & Reports

National Aeronautics andSpace Administration

Technology Focus Electronics/Computers · 2013-04-10 · 16 Strain-Gauge Measurement of Weight of Fluid in a Tank 16 Advanced Docking System With Magnetic Initial Capture 19 Machinery/Automation

Documents