Technology Focus Computers/Electronics · (256) 544-2615 [email protected] Stennis Space Center Robert Bruce (228) 688-1929 [email protected] Carl Ray Small Business

Technology Focus

Computers/Electronics

Software

Materials

Mechanics

Machinery/Automation

Manufacturing

Bio-Medical

Physical Sciences

Information Sciences

Books and Reports

08-04 August 2004

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 CenterGregory Poteat(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, August 2004 1

5 Technology Focus: Data Acquisition

5 Data Relay Board With Protocol for High-Speed,Free-Space Optical Communications

6 Software and Algorithms for Biomedical ImageData Processing and Visualization

7 Rapid Chemometric Filtering of Spectral Data

8 Prioritizing Scientific Data for Transmission

8 Determining Sizes of Particles in a Flow FromDPIV Data

9 Faster Processing for Inverting GPS Occultation Data

11 Electronics/Computers

11 FPGA-Based, Self-Checking, Fault-TolerantComputers

12 Ultralow-Power Digital Correlator for MicrowavePolarimetry

12 Grounding Headphones for Protection Against ESD

13 Lightweight Stacks of Direct Methanol Fuel Cells

14 Highly Efficient Vector-Inversion Pulse Generators

15 Software

15 Estimating Basic Preliminary Design Performancesof Aerospace Vehicles

15 Framework for Development of Object-OrientedSoftware

15 Analyzing Spacecraft Telecommunication Systems

15 Collaborative Planning of Robotic Exploration

16 Tools for Administration of a UNIX-Based Network

16 Preparing and Analyzing Iced Airfoils

16 Evaluating Performance of Components

17 Materials

17 Fuels Containing Methane or Natural Gas inSolution

17 Direct Electrolytic Deposition of Mats of MnxOyNanowires

19 Mechanics

19 Bubble Eliminator Based on Centrifugal Flow

19 Inflatable Emergency Atmospheric-Entry Vehicles

20 Lightweight Deployable Mirrors With TensegritySupports

23 Machinery/Automation

23 Centrifugal Adsorption Cartridge System

23 Ultrasonic Apparatus for Pulverizing BrittleMaterial

25 Bio-Medical

25 Transplanting Retinal Cells Using Bucky Paper forSupport

26 Using an Ultrasonic Instrument to SizeExtravascular Bubbles

27 Physical Sciences

27 Coronagraphic Notch Filter for RamanSpectroscopy

28 On-the-Fly Mapping for Calibrating DirectionalAntennas

29 Working Fluids for Increasing Capacities of HeatPipes

31 Information Sciences

31 Computationally-Efficient Minimum-Time AircraftRoutes in the Presence of Winds

33 Books & Reports

33 Liquid-Metal-Fed Pulsed Plasma Thrusters

33 Personal Radiation Protection System

33 Attitude Control for a Solar-Sail Spacecraft

08-04 August 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: Data Acquisition

In a free-space optical communica-tion system, the mitigation of transientoutages through the incorporation oferror-control methods is of particularconcern, the outages being caused byscintillation fades and obscurants. Thefocus of this innovative technology isthe development of a data relay systemfor a reliable high-data-rate free-space-based optical-transport network. Thedata relay boards will establish the link,maintain synchronous connection,group the data into frames, and providefor automatic retransmission (ARQ) oflost or erred frames. A certain Quality ofService (QoS) can then be ensured,compatible with the required data rate.The protocol to be used by the datarelay system is based on the draftCCSDS standard data-link protocol“Proximity-1,” selected by orbiters tomultiple lander assets in the Mars net-work, for example. In addition to pro-viding data-link protocol capabilities forthe free-space optical link and bufferingthe data, the data relay system will inter-face directly with user applications overGigabit Ethernet and/or with high-speed storage resources via Fibre Chan-

nel. The hardware implementation isbuilt on a network-processor-based ar-chitecture. This technology combinesthe power of a hardware switch capableof data switching and packet routing atGbps rates, with the flexibility of a soft-ware-driven processor that can hosthighly adaptive and reconfigurable pro-tocols used, for example, in wirelesslocal-area networks (LANs).

The system will be implemented in amodular multi-board fashion. The

main hardware elements of the datarelay system are the new data relayboard developed by Rockwell Scien-tific, a COTS Gigabit Ethernet boardfor user interface, and a COTS FibreChannel board that connects to localstorage. The boards reside in a cPCIback plane, and can be housed in aVME-type enclosure.

A block diagram of the data relay sys-tem is shown in Figure 1. The data relayboard, shown in Figure 2, controls the

Data Relay Board With Protocol for High-Speed, Free-SpaceOptical CommunicationsA fade-tolerant data relay system is proposed to ensure reliable delivery of data across anoptical channel.NASA’s Jet Propulsion Laboratory, Pasadena, California

Data Relay Board

Gigabit EthernetInterface Board (COTS)

Fiber ChannelInterface Board (COTS)

cPC

I Bac

kpla

ne

Free-Space Optical

Gigabit Ethernet

Fiber Channel Storage

cPC

I Bus

Bac

kpla

ne

NetworkProcessor

InterfaceAdapter

SDRAM256 MBytes

PCI Bus

2 Gbps

5 Gbps

4 Gbps 16 Bits156 MHz

16 Bits156 MHz

Data Relay Board

FIFO-Oriented,Scheduled Bus

Serial-toParallel

Converter

Parallel-to-Serial

Converter

8-B

it/10

-Bit

Dec

oder

8-B

it/10

-Bit

Enc

oder Transmitter

(Laser)

Receiver

2.5 Gbps

2.5 Gbps

Figure 1. The Data Relay System block diagram shows the interfaces.

Figure 2. A Data Relay Board would be installed in a free-space optical communication terminal to ensure a reliable high-speed data link. The board wouldregulate the flow of data between the user application and the optical transceiver.

6 NASA Tech Briefs, August 2004

data flow between the cPCI bus on theone hand and the transmitter and re-ceiver on the other hand once the free-space optical link has been established.The data rates in transmission and re-ception need not be equal and couldeven differ by as much as several ordersof magnitude. The data relay boardwould contain a commercially availablenetwork processor programmed to per-form the primitive data handling func-tion required by the protocol. Using amemory buffer, the network processorwould accept, from the user applicationor storage through the cPCI bus, astream of data to be transmitted to thelaser. The network processor would formthe data into appropriately sized frameswith headers and frame sequence infor-mation to identify frames for the ARQprocess. The frames would then be sentto an interface adaptor for frame acqui-

sition and synchronization. The inter-face adaptor would then format the datainto 16-bit words, add error check bits,and send the data to the serializer andencoder for transmission to the laser.

As successful receipt of frames is ac-knowledged using the free-space opticallink in the reverse direction, the corre-sponding data are cleared from the localmemory so that capacity for new stream-ing data is made available. In the eventof missed or corrupted data frames, thenetwork processor will reconstruct andretransmit the data frames over the free-space optical link.

On the receiving side, the interfaceadapter will check for errors, while thenetwork processor will check for framesout of sequence. For each receivedframe, the network processor will gen-erate the appropriate ARQ controlframe and pass it to the reverse channel

free-space optical-link interface fortransmission.

This work was done by Malcolm Wrightand Loren Clare of Caltech and Gary Gouldand Maxim Pedyash of Rockwell ScientificCenter for NASA’s Jet Propulsion Labora-tory. Further information is contained in aTSP (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-30610, volume and num-

ber of this NASA Tech Briefs issue, andthe page number.

Software and Algorithms for Biomedical Image Data Processingand VisualizationPlaqTrak automatically assesses plaque deposits on teeth.NASA’s Jet Propulsion Laboratory, Pasadena, California

A new software equipped with novelimage processing algorithms and graphi-cal-user-interface (GUI) tools has beendesigned for automated analysis and pro-cessing of large amounts of biomedicalimage data. The software, called Plaq-Trak, has been specifically used for analy-

sis of plaque on teeth of patients. New al-gorithms have been developed and im-plemented to segment teeth of interestfrom surrounding gum, and a real-timeimage-based morphing procedure isused to automatically overlay a grid ontoeach segmented tooth. Pattern recogni-

tion methods are used to classify plaquefrom surrounding gum and enamel,while ignoring glare effects due to the re-flection of camera light and ambientlight from enamel regions. The PlaqTraksystem integrates these components intoa single software suite with an easy-to-use

Figure 1. PlaqTrak System Utilities are showing some of the GUI tools.


GUI (see Figure 1) that allows users todo an end-to-end run of a patient’srecord, including tooth segmentation ofall teeth, grid morphing of each seg-mented tooth, and plaque classificationof each tooth image.

The automated and accurate process-ing of the captured images to segmenteach tooth [see Figure 2(a)] and thendetect plaque on a tooth-by-tooth basis isa critical component of the PlaqTrak sys-tem to do clinical trials and analysis withminimal human intervention. These fea-tures offer distinct advantages over othercompeting systems that analyze groups ofteeth or synthetic teeth. PlaqTrak divideseach segmented tooth into eight regions

using an advanced graphics morphingprocedure [see results on a chippedtooth in Figure 2(b)], and a patternrecognition classifier is then used to lo-cate plaque [red regions in Figure 2(d)]and enamel regions. The morphing al-lows analysis within regions of teeth,thereby facilitating detailed statisticalanalysis such as the amount of plaquepresent on the biting surfaces on teeth.

This software system is applicable to ahost of biomedical applications, such ascell analysis and life detection, or roboticapplications, such as product inspectionor assembly of parts in space and industry.

This work was done by Ashit Talukder,James Lambert, and Raymond Lam of Cal-

tech for NASA’s Jet Propulsion Labora-tory. Further information is contained in aTSP (see page 1).

In accordance with Public Law 96-517, thecontractor has elected to retain title to this in-vention. Inquiries concerning rights for its com-mercial use should be addressed to:

Intellectual Assets OfficeJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109(818) 354-2240E-mail: [email protected] to NPO-30417, volume and num-


(a) (b) (c) (d)

Figure 2. Examples are shown of Biomedical Image segmentation, morphing, and classification.

A method of rapid, programmable fil-tering of spectral transmittance, re-flectance, or fluorescence data to mea-sure the concentrations of chemicalspecies has been proposed. By “pro-grammable” is meant that a variety ofspectral analyses can readily be per-formed and modified in software,firmware, and/or electronic hardware,without need to change optical filters orother optical hardware of the associatedspectrometers. The method is intendedto enable real-time identification of sin-gle or multiple target chemical speciesin applications that involve high-throughput screening of multiple sam-ples. Examples of such applications in-clude (but are not limited to)combinatorial chemistry, flow cytometry,bead assays, testing drugs, remote sens-ing, and identification of targets.

The basic concept of the proposedmethod is to perform real-time cross-

correlations of a measured spectrumwith one or more analytical function(s)of wavelength that could be, for exam-ple, the known spectra of targetspecies. Assuming that measured spec-tral intensities are proportional to con-centrations of target species plus back-ground spectral intensities, then aftersubtraction of background levels, itshould be possible to determine target-species concentrations from cross-cor-relation values. Of course, the problemof determining the concentrations ismore complex when spectra of differ-ent species overlap, but the problemcan be solved by use of multiple analyt-ical functions in combination withcomputational techniques that havebeen developed previously for analysesof this type.

The method is applicable to the de-sign and operation of a spectrometer inwhich spectrally dispersed light is mea-

sured by means of an active-pixel sensor(APS) array. The row or column dimen-sion of such an array is generally cho-sen to be aligned along the spectral-dis-persion dimension, so that each pixelintercepts light in a narrow spectralband centered on a wavelength that is aknown function of the pixel position.The proposed method admits of twohardware implementations for comput-ing cross-correlations in real time.

One hardware implementation wouldexploit programmable circuitry withineach pixel of an APS array. The analogspectral-intensity reading of the pho-todetector in each pixel would be multi-plied by a gain proportional to value ofthe analytical function for the wave-length that corresponds to the pixel po-sition. As a result, the output from eachpixel would be proportional to contri-bution of the pixel to the cross-correla-tion (plus background). The outputs of

Rapid Chemometric Filtering of Spectral DataTarget species would be identified in real time.NASA’s Jet Propulsion Laboratory, Pasadena, California


all the pixels along the spectral-disper-sion dimension would be summed toobtain the value of the cross-correlation(plus background).

Such on-chip cross-correlation couldbe performed rapidly because the ana-lytical function could be statically pro-grammed into the APS array and themultiplications could be done simulta-neously or nearly so. All of the addi-tions could be done simultaneously bymeans of a single binning instruction.The charge wells of all the pixels couldbe connected simultaneously, collect-ing all the charge outputs from multi-plication operations into one “super-pixel,” the single readout value ofwhich would constitute the cross-corre-lation value for the given analyticalfunction. For an instrument in whichthe APS rows were aligned along thespectral-dispersion dimension and inwhich the image of a spectrograph slitwas aligned along the pixel columnsand spanned multiple pixel rows, it

would be possible to perform simulta-neous cross-correlations for multipletarget species by applying, to each pixelrow, the analytical function correspond-ing to one of the target species. A sepa-rate readout would be needed for eachtarget species.

In the other hardware implementa-tion, cross-correlations would be com-puted externally to the APS array. Themultiplications and additions would beperformed in pipeline fashion. If theAPS-array outputs were analog, thenprogrammable analog signals repre-senting the analytical functions wouldbe synthesized in phase with the corre-sponding stream of analog APS-arrayoutputs and the multiplications and ad-ditions would be performed by rela-tively inexpensive, commercially avail-able analog mixing and filteringcircuits, respectively. If the APS-arrayoutputs were digital, the cross-correla-tions could be computed by a digitalsignal processor. Ordinarily, the analog

approach would be preferable becausethe analog operations can be per-formed much more rapidly than canthe corresponding digital multiplica-tions and additions.

This work was done by Gregory Bearman,Michael Pelletier, and Suresh Seshadri ofCaltech 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-30912, volume and num-


A software system has been developedfor prioritizing newly acquired geologi-cal data onboard a planetary rover. Thesystem has been designed to enable effi-cient use of limited communication re-sources by transmitting the data likely tohave the most scientific value. This soft-ware operates onboard a rover by ana-lyzing collected data, identifying poten-tial scientific targets, and then usingthat information to prioritize data fortransmission to Earth. Currently, the sys-tem is focused on the analysis of ac-quired images, although the generaltechniques are applicable to a widerange of data modalities. Image prioriti-zation is performed using two main

steps. In the first step, the software de-tects features of interest from eachimage. In its current application, thesystem is focused on visual properties ofrocks. Thus, rocks are located in eachimage and rock properties, such asshape, texture, and albedo, are ex-tracted from the identified rocks. In thesecond step, the features extracted froma group of images are used to prioritizethe images using three different meth-ods: (1) identification of key target sig-nature (finding specific rock featuresthe scientist has identified as impor-tant), (2) novelty detection (findingrocks we haven’t seen before), and (3)representative rock sampling (finding

the most average sample of each rocktype). These methods use techniquessuch as K-means unsupervised cluster-ing and a discrimination-based kernelclassifier to rank images based on theirinterest level.

This program was written by Rebecca Cas-tano, Robert Anderson, Tara Estlin, Dennis De-Coste, Daniel Gaines, Dominic Mazzoni, ForestFisher, and Michele Judd of Caltech for NASA’sJet Propulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).

This software is available for commercial li-censing. Please contact Don Hart of the Cali-fornia Institute of Technology at (818) 393-3425. Refer to NPO-40265.

Prioritizing Scientific Data for TransmissionNASA’s Jet Propulsion Laboratory, Pasadena, California

Determining Sizes of Particles in a Flow From DPIV DataThe same equipment would be used to measure sizes as well as velocities.John H. Glenn Research Center, Cleveland, Ohio

A proposed method of measuring thesize of particles entrained in a flow of aliquid or gas would involve utilization ofdata from digital particle-image ve-locimetry (DPIV) of the flow. That is tosay, with proper design and operation of

a DPIV system, the DPIV data could beprocessed according to the proposedmethod to obtain particle sizes in addi-tion to particle velocities. As an addi-tional benefit, one could then computethe mass flux of the entrained particles

from the particle sizes and velocities.As in DPIV as practiced heretofore, a

pulsed laser beam would be formed intoa thin sheet to illuminate a plane of in-terest in a flow field and the illuminatedplane would be observed by means of a

charge-coupled device (CCD) cameraaimed along a line perpendicular to theilluminated plane. Unlike in DPIV aspracticed heretofore, care would betaken to polarize the laser beam so thatits electric field would lie in the illumi-nated plane, for the reason explained inthe next paragraph.

The proposed method applies, morespecifically, to transparent or semitrans-parent spherical particles that have anindex of refraction different from that ofthe fluid in which they are entrained.The method is based on the establishedMie theory, which describes the scatter-ing of light by diffraction, refraction, andspecular reflection of light by such parti-cles. In the case of a particle illuminatedby polarized light and observed in thearrangement described in the precedingparagraph, the Mie theory shows that theimage of the particle on the focal planeof the CCD camera includes two glarespots: one attributable to light reflectedtoward the camera and one attributableto light refracted toward the camera.The distance between the glare spots is aknown function of the size of the parti-cle, the indices of refraction of the parti-cle material, and design parameters ofthe camera optics. Hence, the size of aparticle can be determined from the dis-tance between the glare spots.

The proposed method would be im-plemented in an algorithm that wouldautomatically identify, and measure thedistance between, the glare spots foreach particle for which a suitable imagehas been captured in a DPIV imageframe. The algorithm (see figure)would begin with thresholding of datafrom the entire image frame to reducenoise, thereby facilitating discrimina-tion of particle images from the back-ground and aiding in the separation ofoverlapping particles. It is importantnot to pick a threshold level so highthat the light intensity between a givenpair of glare spots does not fall belowthe threshold value, leaving the glarespots disconnected.

The image would then be scanned in asequence of rows and columns of pixelsto identify groups of adjacent pixels thatcontain nonzero brightnesses and that

are surrounded by pixels of zero bright-ness. Each such group would be assumedto constitute the image of one particle.Each such group would be further ana-lyzed to determine whether the imagewas saturated; saturated particle imagesmust be rejected because the locations ofglare spots in saturated images cannotaccurately be determined. Within eachunsaturated particle image, the cen-troids (deemed to be the locations) ofthe glare spots would be determined bymeans of gradients of brightness distrib-utions and three-point horizontal andthree-point vertical Gaussian estimatesbased on the brightness values of thebrightest pixels and the pixels adjacentto them. If the brightness of a given par-ticle image contained only one peak,then it would be assumed that a secondglare spot did not exist and that imagewould be rejected.

Once the centroids had been esti-mated for all particle images for whichit was possible to do so, the positions ofthe particles and the distances betweentheir centroids would be computed. Asdescribed above, the size of each parti-cle would then be computed from thedistance between its centroids. Finally,the distribution, mean, and standarddeviation of sizes would be computedfor the collection of particle imagesthat survived to the final stage of thecentroid-estimation process.

This work was done by M. P. Wernet ofGlenn Research Center and A. Mielkeand J. R. Kadambi of Case Western ReserveUniversity. Further information is con-tained in a TSP (see page 1).

Inquiries concerning rights for the com-mercial use of this invention should be ad-dressed to NASA Glenn Research Center,Commercial Technology Office, Attn: SteveFedor, Mail Stop 4–8, 21000 BrookparkRoad, Cleveland, Ohio 44135. Refer toLEW-17340.


The Algorithm of the Proposed Method wouldprocess DPIV data to extract information on par-ticle sizes. The size information would be in ad-dition to the velocity information extractedfrom the data by use of previously developedDPIV software.

Read in 12-bit image file.

Apply global thresholdto eliminate image noise.

Scan image for nonzeropixels and group nonzero neighbors into images of

particles.

Determine glare-spot positions and calculate

particle diameters.

Calculate positions of particles, the distribution of

particle sizes, and the mean and standard

deviation of particle sizes.

Faster Processing for Inverting GPS Occultation DataNASA’s Jet Propulsion Laboratory, Pasadena, California

A document outlines a computa-tional method that can be incorpo-rated into two prior methods used toinvert Global Positioning System(GPS) occultation data [signal data ac-quired by a low-Earth-orbiting satelliteas either this or the GPS satellite risesabove or falls below the horizon] toobtain information on altitude-depen-

dent properties of the atmosphere.The two prior inversion methods,known as back propagation and canon-ical transform, are computationally ex-pensive because for each occultation,they involve numerical evaluation of alarge number of diffraction-like spatialintegrals. The present method involvesan angular-spectrum-based phase-ex-

trapolation approximation in whicheach data point is associated with aplane-wave component that propa-gates in a unique direction from theorbit of the receiving satellite to inter-sect a straight line tangent to the orbitat a nearby point. This approximationenables the use of fast Fourier trans-forms (FFTs), which apply only to data


collected along a straight-line trajec-tory. The computation of the diffrac-tion-like integrals in the angular-spec-trum domain by use of FFTs takes onlyseconds, whereas previously, it tookminutes.

This work was done by Chi Ao of Caltechfor NASA’s Jet Propulsion Laboratory.Further information is contained in a TSP(see page 1).

The software used in this innovation isavailable for commercial licensing. Please

contact Don Hart of the California Instituteof Technology at (818) 393-3425. Refer toNPO-30791.


Electronics/Computers

A proposed computer architecturewould exploit the capabilities of com-mercially available field-programmablegate arrays (FPGAs) to enable computersto detect and recover from bit errors.The main purpose of the proposed ar-chitecture is to enable fault-tolerant com-puting in the presence of single-eventupsets (SEUs). [An SEU is a spurious bitflip (also called a soft error) caused by asingle impact of ionizing radiation.] Thearchitecture would also enable recoveryfrom some soft errors caused by electri-cal transients and, to some extent, fromintermittent and permanent (hard) er-rors caused by aging of electronic com-ponents.

A typical FPGA of the current genera-tion contains one or more completeprocessor cores, memories, and high-speed serial input/output (I/O) chan-nels, making it possible to shrink aboard-level processor node to a singleintegrated-circuit chip. Custom, highlyefficient microcontrollers, general-pur-pose computers, custom I/O processors,and signal processors can be rapidly andefficiently implemented by use ofFPGAs. Unfortunately, FPGAs are sus-ceptible to SEUs. Prior efforts to miti-gate the effects of SEUs have yielded so-lutions that degrade performance of thesystem and require support from exter-nal hardware and software.

In comparison with other fault-toler-ant-computing architectures (e.g., triplemodular redundancy), the proposed ar-chitecture could be implemented withless circuitry and lower power demand.Moreover, the fault-tolerant computingfunctions would require only minimalsupport from circuitry outside the cen-tral processing units (CPUs) of comput-ers, would not require any software sup-port, and would be largely transparentto software and to other computer hard-ware.

There would be two types of modules:a self-checking processor module and amemory system (see figure). The self-checking processor module would beimplemented on a single FPGA andwould be capable of detecting its owninternal errors. It would contain two

CPUs executing identical programs inlock step, with comparison of their out-puts to detect errors. It would also con-tain various cache local memory cir-cuits, communication circuits, andconfigurable special-purpose processorsthat would use self-checking checkers.(The basic principle of the self-checkingchecker method is to utilize logic cir-cuitry that generates error signals when-ever there is an error in either thechecker or the circuit being checked.)

The memory system would comprise amain memory and a hardware-con-trolled check-pointing system (CPS)based on a buffer memory denoted therecovery cache. The main memorywould contain random-access memory(RAM) chips and FPGAs that would, inaddition to everything else, implementdouble-error-detecting and single-error-correcting memory functions to enablerecovery from single-bit errors.

The main purpose served by thememory system as a whole would be toenable the computer to return to a validstate — a known good point reached inthe computations before the occur-rence of a detected error. In operation,the checkers in the self-checkingprocessor module would signal errors tothe memory system. Recovery would in-volve halting the operation of the self-checking processor module, correctingits configuration bits if necessary, re-loading its registers, and returning con-trol to a previous, known good point inthe program. The CPUs could then re-sume correct computations.

The known good point in the compu-tations would be provided by the CPS ina procedure denoted, variously, as check-pointing and checkpoint recovery orcheckpoint rollback. The CPS would pe-riodically command each CPU to storethe contents of its registers in the recov-ery cache and clear its caches. This ac-tion would establish a checkpoint. Thenthe original value and the address of anyclean RAM block that was subsequentlyoverwritten by the CPU would be storedin a special RAM within the recoverycache. Subsequent writes to that blockwould be carried out normally (that is,without intervention by the recoverycache). If an error in the CPU were de-tected, the data in the special recovery-cache RAM could be used to restore thecorresponding data in the main memoryto their prior correct values, the proces-sor configuration would be reloaded, thecaches in the processor module would becleared, and the processor registers re-stored to their prior values.

A new checkpoint could be orderedwhen the recovery cache became filled tocapacity. Alternately, checkpoints couldbe forced at strategic points in the soft-ware. Another alternative would be toforce checkpoints periodically, at inter-vals short enough to ensure that rollbacktime did not exceed a value that could bespecified by design.

This work was done by Raphael Some andDavid Rennels of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).NPO-30806

FPGA-Based, Self-Checking, Fault-Tolerant ComputersNo software support and little hardware support would be needed for fault tolerance.NASA’s Jet Propulsion Laboratory, Pasadena, California

Self-CheckingProcessor Module

Main Memory CPS RecoveryCache

Memory System

Save/Restore

Registers, Etc.

Error Detected

The Memory System would store the states of computations at checkpoint intervals. Upon detectionof an error in the self-checking processor module, the memory system would provide the informationneeded to roll back the computations to the immediately preceding checkpoint.


A recently developed high-speed digitalcorrelator is especially well suited for pro-cessing readings of a passive microwave po-larimeter. This circuit computes the auto-correlations of, and the cross-correlationsamong, data in four digital input streamsrepresenting samples of in-phase (I) andquadrature (Q) components of two inter-mediate-frequency (IF) signals, denoted Aand B, that are generated in heterodyne re-ception of two microwave signals.

The IF signals arriving at the correlatorinput terminals have been digitized tothree levels (-1,0,1) at a sampling rate up to500 MHz. Two bits (representing sign andmagnitude) are needed to represent theinstantaneous datum in each input chan-nel; hence, eight bits are needed to repre-sent the four input signals during any givencycle of the sampling clock. The accumu-lation (integration) time for the correla-tion is programmable in increments of 28

cycles of the sampling clock, up to a maxi-mum of 224 cycles.

The basic functionality of the correla-tor is embodied in 16 correlation slices,each of which contains identical logic cir-cuits and counters (see figure). The firststage of each correlation slice is a logicgate that computes one of the desired cor-relations (for example, the autocorrela-tion of the I component of A or the nega-tive of the cross-correlation of the Icomponent of A and the Q component ofB). The sampling of the output of thelogic gate output is controlled by the sam-pling-clock signal, and an 8-bit counter in-crements in every clock cycle when the

logic gate generates output. The most sig-nificant bit of the 8-bit counter is sampledby a 16-bit counter with a clock signal at 2-

8 the frequency of the sampling clock.The 16-bit counter is incremented everytime the 8-bit counter rolls over.

The correlator is designed for use witha microprocessor. The microprocessorcontrols the function of the correlator,sets the desired integration time by writ-ing appropriate values to registers in thecorrelator, and reads the correlation out-puts as described next. At the end of theintegration period, the contents of the16-bit counter are copied to a 16-bitbuffer, and the 16-bit counter is clearedto begin a new accumulation cycle. Atthe same time, the correlator generates asignal to indicate, to the microprocessor,that new correlation data are available.The correlator and the microprocessorcommunicate via a simple 3.3-V bus-ori-ented interface, such that from the per-spective of the microprocessor, the cor-relator acts much like a smallrandom-access memory containing 32

16-bit words. Hence, the microprocessorreads the correlation-slice buffers by sup-plying five-bit addresses to select thebuffers as a group of memory locations.

The correlator has been imple-mented as a complementary metaloxide/semiconductor (CMOS) inte-grated circuit, following 0.35-µm radia-tion tolerant design rules. The main ad-vantage of this high-speed digitalcorrelator over prior ones is ultralow-power dissipation: whereas a previoushigh-speed digital correlator dissipates apower of about 10 W (and processes onlytwo input data streams), this correlatordissipates a power of 2 mW or less, theexact value depending on the samplingrate. To achieve such ultralow-poweroperation a logic level of only 0.5 V isused, necessitating the use of specialsignal-conditioning circuits.

This work was done by Jeffrey R. Piepmeier ofGoddard Space Flight Center and K. JosephHass of the University of Idaho. Further infor-mation is contained in a TSP (see page 1).GSC-14746-1

Ultralow-Power Digital Correlator for Microwave PolarimetryThis circuit overcomes disadvantages of prior digital correlators.Goddard Space Flight Center, Greenbelt, Maryland

8-Bit Counter 16-Bit Counter 16-Bit Buffer

Clock ÷ 28

Data Out16

Clock

A

B

A Correlation Slice is one of 16 identical units that embody the basic functionality of the high-speed,ultralow-power digital correlator.

Grounding Headphones for Protection Against ESDJohn F. Kennedy Space Center, Florida

A simple alternative technique hasbeen devised protecting delicate equip-ment against electrostatic discharge(ESD) in settings in which workers wearcommunication headsets. In the originalsetting in which the technique was de-vised, the workers who wear the headsetsalso wear anti-ESD grounding straps ontheir wrists. The alternative techniqueeliminates the need for the wrist ground-ing straps by providing for groundingthrough the headsets. In place of theelectrically insulating foam pads on the

headsets, one installs pads made of elec-trically conductive foam like that com-monly used to protect electronic compo-nents. Grounding wires are attached tothe conductive foam pads, then possiblyto the shielding cable which may begrounded to the backshell on the con-nector. The efficacy of this technique inprotecting against ESD has been verifiedin experiments. The electrical resistanceof the pads is a few megohms — aboutthe same as that of a human body be-tween the fingers of opposite hands and,

hence, low enough for grounding. Theonly drawback of the technique is thatcare must be taken to place the foampads in contact with the user’s skin: anyhair that comes between the foam padsand the skin must be pushed aside be-cause hair is electrically insulating andthus prevents adequate grounding.

This work was done by John Peters andRobert C. Youngquist of Kennedy SpaceCenter. For further information, contact theKennedy Commercial Technology Office at321-867-1463. KSC-12295


An improved design concept for di-rect methanol fuel cells makes it possi-ble to construct fuel-cell stacks that canweigh as little as one-third as much asdo conventional bipolar fuel-cell stacksof equal power. The structural-supportcomponents of the improved cells andstacks can be made of relatively inex-pensive plastics. Moreover, in compari-son with conventional bipolar fuel-cellstacks, the improved fuel-cell stacks canbe assembled, disassembled, and diag-nosed for malfunctions more easily.These improvements are expected tobring portable direct methanol fuelcells and stacks closer to commercial-ization.

In a conventional bipolar fuel-cellstack, the cells are interspersed withbipolar plates (also called biplates),which are structural components thatserve to interconnect the cells and dis-tribute the reactants (methanol andair). The cells and biplates are sand-wiched between metal end plates. Usu-ally, the stack is held together underpressure by tie rods that clamp the endplates. The bipolar stack configurationoffers the advantage of very low internalelectrical resistance. However, when thepower output of a stack is only a fewwatts, the very low internal resistance ofa bipolar stack is not absolutely neces-sary for keeping the internal power lossacceptably low.

Typically, about 80 percent of themass of a conventional bipolar fuel-cellstack resides in the biplates, end plates,and tie rods. The biplates are usuallymade of graphite composites and mustbe molded or machined to containflow channels, at a cost that is usually amajor part of the total cost of the stack.In the event of a malfunction in onecell, it is necessary to disassemble theentire stack in order to be able to diag-nose that cell. What is needed is a de-sign that reduces the mass of the stack,does not require high pressure to en-sure sealing, is more amenable to trou-bleshooting, and reduces the cost ofmanufacture.

The present improved design satis-fies these needs and is especially suit-able for applications in which thepower demand is ≤20 W. This designeliminates the biplates, end plates, andtie rods. In this design, the basic build-

ing block of a stack is a sealed unit thatcontains an anode plate, two cathodeplates, and two back-to-back cells (seefigure). The structural-support andflow-channeling components of the

units are made from inexpensive plas-tics. Each unit is assembled and testedseparately, then the units are assem-bled into the stack. The units arejoined by simple snap seals similar to

Partly Sealed Units Are Joined by Snap Seals and their current collectors are connected together toform a fuel-cell stack. Each sealed unit contains a back-to-back pair of fuel cells.

Anode PlateCathode Plate

Cathodes Anodes GasketSolid

Electrolyte

Methanol In

PARTLY SEALED UNIT (BUILDING BLOCK)

STACK WITHOUT ELECTRICAL CONNECTIONS

AirIn

AirOut

Methanol Out

Methanol In Methanol Out

SnapSeals

Cathode Plate

CELLS ELECTRICALLY CONNECTED IN SERIES

Current Collectors

Lightweight Stacks of Direct Methanol Fuel CellsPower-to-mass ratios can be tripled.NASA’s Jet Propulsion Laboratory, Pasadena, California


the zipperlike seals on plastic bagscommonly used to store food. Thecathode and anode plates include cur-rent collectors, the inside ends ofwhich are electrically connected to theelectrodes and the outer ends of whichcan be used to form the desired seriesand/or parallel electrical connectionsamong the cells. Because the stackneed not be clamped or otherwise heldtogether under pressure, the stack can

easily be disassembled to replace a mal-functioning sealed unit.

This work was done by SekharipuramNarayanan and Thomas Valdez of Cal-tech for NASA’s Jet Propulsion Labora-tory. Further information is contained ina TSP (see page 1).

In accordance with Public Law 96-517, the contractor has elected to retaintitle to this invention. Inquiries concern-ing rights for its commercial use should be

addressed toIntellectual Assets OfficeJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109(818) 354-2240E-mail: [email protected] to NPO-30570, volume and number

of this NASA Tech Briefs issue, and thepage number.

Improved transmission-line pulse gen-erators of the vector-inversion type arebeing developed as lightweight sources ofpulsed high voltage for diverse applica-tions, including spacecraft thrusters,portable x-ray imaging systems, impulseradar systems, and corona-discharge sys-tems for sterilizing gases. In this develop-ment, more than the customary attentionis paid to principles of operation and de-tails of construction so as to the maximizethe efficiency of the pulse-generation

process while minimizing the sizes ofcomponents. An important element ofthis approach is segmenting a pulse gen-erator in such a manner that the electricfield in each segment is always below thethreshold for electrical breakdown. Onedesign of particular interest, a completedescription of which was not available atthe time of writing this article, involvestwo parallel-plate transmission lines thatare wound on a mandrel, share a com-mon conductor, and are switched in such

a manner that the pulse generator is di-vided into a “fast” and a “slow” section. Amajor innovation in this design is the ad-dition of ferrite to the “slow” section toreduce the size of the mandrel neededfor a given efficiency.

This work was done by M. Franklin Roseof Radiance Technologies, Inc., for Mar-shall Space Flight Center. Further in-formation is contained in a TSP (seepage 1).MFS-31870

Highly Efficient Vector-Inversion Pulse GeneratorsMarshall Space Flight Center, Alabama


Estimating Basic Prelimi-nary Design Performancesof Aerospace Vehicles

Aerodynamics and Performance Esti-mation Toolset is a collection of foursoftware programs for rapidly estimatingthe preliminary design performance ofaerospace vehicles represented by doingsimplified calculations based on ballistictrajectories, the ideal rocket equation,and supersonic wedges through stan-dard atmosphere. The program consistsof a set of Microsoft Excel worksheetsubprograms. The input and outputdata are presented in a user-friendly for-mat, and calculations are performedrapidly enough that the user can iterateamong different trajectories and/orshapes to perform “what-if” studies. Esti-mates that can be computed by theseprograms include:1. Ballistic trajectories as a function of de-

parture angles, initial velocities, initialpositions, and target altitudes; assum-ing point masses and no atmosphere.The program plots the trajectory intwo-dimensions and outputs the posi-tion, pitch, and velocity along the tra-jectory.

2. The “Rocket Equation” program calcu-lates and plots the trade space for a ve-hicle’s propellant mass fraction over arange of specific impulse and missionvelocity values, propellant mass frac-tions as functions of specific impulsesand velocities.

3. “Standard Atmosphere” will estimatethe temperature, speed of sound, pres-sure, and air density as a function of al-titude in a standard atmosphere, prop-erties of a standard atmosphere asfunctions of altitude.

4. “Supersonic Wedges” will calculate thefree-stream, normal-shock, oblique-shock, and isentropic flow propertiesfor a wedge-shaped body flying super-sonically through a standard atmos-phere. It will also calculate the maxi-mum angle for which a shock remainsattached, and the minimum Machnumber for which a shock becomes at-tached, all as functions of the wedgeangle, altitude, and Mach number.This work was done by Paul L. Luz and Regi-

nald Alexander of Marshall Space Flight Cen-ter. For further information, contact CarolineWang, MSFC Software Release Authority, at(256) 544-3887 or [email protected] to MFS-31795.

Framework for Developmentof Object-Oriented Software

The Real-Time Control (RTC) Appli-cation Framework is a high-level soft-ware framework written in C++ that sup-ports the rapid design andimplementation of object-oriented ap-plication programs. This framework pro-vides built-in functionality that solvescommon software development prob-lems within distributed client-server,multi-threaded, and embedded pro-gramming environments. When usingthe RTC Framework to develop softwarefor a specific domain, designers and im-plementers can focus entirely on the de-tails of the domain-specific softwarerather than on creating custom solu-tions, utilities, and frameworks for thecomplexities of the programming envi-ronment. The RTC Framework was orig-inally developed as part of a Space Shut-tle Launch Processing System (LPS)replacement project called Checkoutand Launch Control System (CLCS). Asa result of the framework’s develop-ment, CLCS software development timewas reduced by 66 percent. The frame-work is generic enough for developingapplications outside of the launch-pro-cessing system domain. Other applicablehigh-level domains include commandand control systems and simulation/training systems.

This software was written by Gus Perez-Povedaof 360 Software Corporation, Tony Ciavarella ofUnited Space Alliance, and Dan Nieten ofKennedy Space Center. For further informa-tion, access http://www.360SoftwareCorp.com.


360 Software Corporation12472 Lake Underhill Rd. #174Orlando, FL 32828Phone: (407) 694-2227Refer to KSC-12499, volume and number

of this NASA Tech Briefs issue, and thepage number.

Analyzing SpacecraftTelecommunication Systems

Multi-Mission Telecom Analysis Tool(MMTAT) is a C-language computer pro-gram for analyzing proposed spacecrafttelecommunication systems. MMTAT uti-

lizes parameterized input and computa-tional models that can be run on standarddesktop computers to perform fast andaccurate analyses of telecommunicationlinks. MMTAT is easy to use and can easilybe integrated with other software applica-tions and run as part of almost any com-putational simulation. It is distributed aseither a stand-alone application programwith a graphical user interface or a link-able library with a well-defined set of ap-plication programming interface (API)calls. As a stand-alone program, MMTATprovides both textual and graphical out-put. The graphs make it possible to un-derstand, quickly and easily, how telecom-munication performance varies withvariations in input parameters. A delim-ited text file that can be read by anyspreadsheet program is generated at theend of each run. The API in the linkable-library form of MMTAT enables the userto control simulation software and tochange parameters during a simulationrun. Results can be retrieved either at theend of a run or by use of a function call atany time step.

This program was written by Mark Kor-don, David Hanks, Roy Gladden, and EricWood of Caltech for NASA’s Jet PropulsionLaboratory. Further information is con-tained in a 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-40298.

Collaborative Planning of Robotic Exploration

The Science Activity Planner (SAP)software system includes an uplink-plan-ning component, which enables collabo-rative planning of activities to be under-taken by an exploratory robot on aremote planet or on Earth. Included inthe uplink-planning component is theSAP-Uplink Browser, which enables usersto load multiple spacecraft activity plansinto a single window, compare them, andmerge them. The uplink-planning com-ponent includes a subcomponent thatimplements the Rover Markup LanguageActivity Planning format (RML-AP),based on the Extensible Markup Lan-guage (XML) format that enables therepresentation, within a single docu-ment, of planned spacecraft and roboticactivities together with the scientific rea-

Software


sons for the activities. Each such docu-ment is highly parseable and can be vali-dated easily. Another subcomponent ofthe uplink-planning component is theActivity Dictionary Markup Language(ADML), which eliminates the need fortwo mission activity dictionaries — one ina human-readable format and one in amachine-readable format. Style sheetsthat have been developed along with theADML format enable users to edit onedictionary in a user-friendly environmentwithout compromising the machine-readability of the format.

This program was written by Jeffrey Norris,Paul Backes, Mark Powell, Marsette Vona,and Robert Steinke of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).


Tools for Administration of a UNIX-Based Network

Several computer programs havebeen developed to enable efficient ad-ministration of a large, heterogeneous,UNIX-based computing and communi-cation network that includes a variety ofcomputers connected to a variety of sub-networks. One program provides securesoftware tools for administrators to cre-ate, modify, lock, and delete accounts ofspecific users. This program also pro-vides tools for users to change theirUNIX passwords and log-in shells. Thesetools check for errors. Another programcomprises a client and a server compo-nent that, together, provide a securemechanism to create, modify, and queryquota levels on a network file system(NFS) mounted by use of the VERITASFile SystemJ software. The client soft-ware resides on an internal secure com-puter with a secure Web interface; onecan gain access to the client softwarefrom any authorized computer capableof running web-browser software. The

server software resides on a UNIX com-puter configured with the VERITAS soft-ware system. Directories where VERITASquotas are applied are NFS-mounted.Another program is a Web-based,client/server Internet Protocol (IP) ad-dress tool that facilitates maintenancelookup of information about IP ad-dresses for a network of computers.

These programs were written by StephenLeClaire and Edward Farrar of Netlander,Inc., for Kennedy Space Center. For fur-ther information, contact the Kennedy Com-mercial Technology Office at (321) 867-1463.KSC-12269/68/71

Preparing and AnalyzingIced Airfoils

SmaggIce version 1.2 is a computerprogram for preparing and analyzingiced airfoils. It includes interactive toolsfor (1) measuring ice-shape characteris-tics, (2) controlled smoothing of iceshapes, (3) curve discretization, (4) gen-eration of artificial ice shapes, and (5) de-tection and correction of input errors.Measurements of ice shapes are essentialfor establishing relationships betweencharacteristics of ice and effects of ice onairfoil performance. The shape-smooth-ing tool helps prepare ice shapes for usewith already available grid-generationand computational-fluid-dynamics soft-ware for studying the aerodynamic effectsof smoothed ice on airfoils. The artificialice-shape generation tool supports para-metric studies since ice-shape parameterscan easily be controlled with the artificialice. In such studies, artificial shapes gen-erated by this program can supplementsimulated ice obtained from icing re-search tunnels and real ice obtained fromflight test under icing weather condition.SmaggIce also automatically detectsgeometry errors such as tangles or dupli-cate points in the boundary which may beintroduced by digitization and providestools to correct these. By use of interac-tive tools included in SmaggIce version1.2, one can easily characterize ice shapes

and prepare iced airfoils for grid genera-tion and flow simulations.

This program was written by Mary B. Vicker-man, Marivell Baez, Donald C. Braun, BarbaraJ. Cotton, Yung K. Choo, Rula M. Coroneos,James A. Pennline, Anthony W. Hackenberg,and Herbert W. Schilling of Glenn ResearchCenter. John W. Slater with Yung K. Choo con-tributed to the conception of the code. In addition,Kevin M. Burke, Gerald J. Nolan, and DennisBrown of InDyne, Inc., contributed to developingtraining material. For further information, accesshttp://icebox-esn.grc.nasa.gov/ext/design/smaggice.html.

Inquiries concerning rights for the commer-cial use of this invention should be addressedto NASA Glenn Research Center, CommercialTechnology Office, Attn: Steve Fedor, MailStop 4-8, 21000 Brookpark Road, Cleveland,Ohio 44135. Refer to LEW-17399.

Evaluating Performance of Components

Parallel Component PerformanceBenchmarks is a computer program devel-oped to aid the evaluation of the CommonComponent Architecture (CCA) — a soft-ware architecture, based on a componentmodel, that was conceived to foster high-performance computing, including parallelcomputing. More specifically, this programcompares the performances (principallyby measuring computing times) of compo-nentized versus conventional versions ofthe Parallel Pyramid 2D Adaptive Mesh Re-finement library — a software library that isused to generate computational meshesfor solving physical problems and that istypical of software libraries in use at NASA’sJet Propulsion Laboratory.

This program was written by DanielKatz, Edwin Tisdale, and Charles Nortonof Caltech for NASA’s Jet PropulsionLaboratory. Further information is con-tained in a TSP (see page 1).



Materials

Fuels Containing Methane or Natural Gas in SolutionA blend of gasoline and natural gas offers advantages over alternative fuels.Lyndon B. Johnson Space Center, Houston, Texas

While exploring ways of producingbetter fuels for propulsion of a space-craft on the Mars sample return mission,a researcher at Johnson Space Center(JSC) devised a way of blending fuel bycombining methane or natural gas witha second fuel to produce a fuel that canbe maintained in liquid form at ambienttemperature and under moderate pres-sure. The use of such a blended fuelwould be a departure for both spacecraftengines and terrestrial internal combus-tion engines. For spacecraft, it would en-able reduction of weights on longflights. For the automotive industry onEarth, such a fuel could be easily distrib-uted and could be a less expensive, moreefficient, and cleaner-burning alterna-tive to conventional fossil fuels.

The concept of blending fuels is notnew: for example, the production ofgasoline includes the addition of liquidoctane enhancers. For the future, it hasbeen commonly suggested to substitutemethane or compressed natural gas foroctane-enhanced gasoline as a fuel forinternal-combustion engines. Unfortu-nately, methane or natural gas must bestored either as a compressed gas (if keptat ambient temperature) or as a cryo-genic liquid. The ranges of automobileswould be reduced from their present val-

ues because of limitations on the capaci-ties for storage of these fuels. Moreover,technical challenges are posed by theneed to develop equipment to handlethese fuels and, especially, to fill tanks ac-ceptably rapidly. The JSC alternative —to provide a blended fuel that can bemaintained in liquid form at moderatepressure at ambient temperature — hasnot been previously tried.

A blended automotive fuel accordingto this approach would be made by dis-solving natural gas in gasoline. The auto-genous pressure of this fuel would elimi-nate the need for a vehicle fuel pump,but a pressure and/or flow regulatorwould be needed to moderate the effectsof temperature and to respond to chang-ing engine power demands. Because thefuel would flash as it entered enginecylinders, relative to gasoline, it woulddisperse more readily and thereforewould mix with air more nearly com-pletely. As a consequence, this fuel wouldburn more nearly completely (and,hence, more cleanly) than gasoline does.

The storage density of this fuel wouldbe similar to that of gasoline, but its en-ergy density would be such that themileage (more precisely, the distancetraveled per unit volume of fuel) wouldbe greater than that of either gasoline or

compressed natural gas. Because thepressure needed to maintain the fuel inliquid form would be more nearly con-stant and generally lower than thatneeded to maintain compressed naturalgas in liquid form, the pressure rating ofa tank used to hold this fuel could belower than that of a tank used to holdcompressed natural gas.

A mixture of natural gas and gasolinecould be distributed more easily thancould some alternative fuels. A massiveinvestment in new equipment would notbe necessary: One could utilize the pre-sent fuel-distribution infrastructure andcould blend the gasoline and natural gasat almost any place in the production ordistribution process — perhaps even atthe retail fuel pump. Yet another advan-tage afforded by use of a blend of gaso-line and natural gas would be a reduc-tion in the amount of gasoline consumed.Because natural gas costs less than gaso-line does and is in abundant supply inthe United States, the cost of automotivefuel and the demand for imported oilcould be reduced.

This work was done by Thomas A. Sulli-van of Johnson Space Center. For furtherinformation, contact the Johnson CommercialTechnology Office at (281) 483-3809.Refer to MSC-22873.

Direct Electrolytic Deposition of Mats of MnxOy NanowiresThese mats of nanowires can be used as electrodes for batteries and capacitors.NASA’s Jet Propulsion Laboratory, Pasadena, California

Mats of free-standing manganeseoxide (MnxOy) nanowires have beenfabricated as experimental electrodematerials for rechargeable electrochem-ical power cells and capacitors. Becausethey are free-standing, the wires in thesemats are electrochemically accessible.The advantage of the mat-of-nanowiresconfiguration, relative to other configu-rations of electrode materials, arisesfrom the combination of narrownessand high areal number density of thewires. This combination offers both high

surface areas for contact with elec-trolytes and short paths for diffusion ofions into and out of the electrodes,thereby making it possible to charge anddischarge at rates higher than wouldotherwise be possible and, consequently,to achieve greater power densities.

The nanowires are fabricated in anelectrolytic process in which there is noneed for an electrode binder material.Moreover, there is no need to incorpo-rate an electrically conductive additiveinto the electrode material; the only

electrically conductive material thatmust be added is a thin substrate contactfilm at the anchored ends of thenanowires. Hence, the mass fraction ofactive electrode material is close to 100percent, as compared with about 85 per-cent in conventional electrodes madefrom a slurry of active electrode mater-ial, binder, and conductive additivepressed onto a metal foil.

The locations and sizes of thenanowires are defined by holes in tem-plates in the form of commercially avail-


able porous alumina membranes. In ex-periments to demonstrate the presentprocess, alumina membranes of variouspore sizes and degrees of porosity wereused. First, a film of Au was sputteredonto one side of each membrane. Themembranes were then attached, vari-ously, to carbon tape or a gold substrateby use of silver or carbon paste. Oncethus attached, the membranes were im-mersed in a plating solution comprising0.01 M MnSO4 + 0.03 M (NH4)2SO4. ThepH of the solution was kept constant at 8by addition of H2SO4 or NH4OH asneeded. MnxOy nanowires were poten-

tiostatically electrodeposited in the poresin the alumina templates. Depending onthe anodic deposition potentials, MnxOywas deposited in various oxidation states[divalent (Mn3O4), trivalent (Mn2O3),or tetravalent (MnO2)]. The MnxOywires were made free-standing (see fig-ure) by dissolving the alumina templates,variously, in KOH or NaOH at a concen-tration of 20 volume percent.

This work was done by Nosang Myung,William West, Jay Whitacre, and Ratnaku-mar Bugga of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).


Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099(818) 354-2240E-mail: [email protected] to NPO-30655, volume and number

of this NASA Tech Briefs issue, and the pagenumber.

These Free-Standing MnxOy Wires were made by electrodeposition of MnxOy in 20-nm-wide pores in an alumina membrane, then dissolving the membranein a concentrated alkaline solution.

Cross-Sectional View (Low Magnification) Cross-Sectional View (High Magnification) Top View


Mechanics

Bubble Eliminator Based on Centrifugal FlowThis device contains no moving parts.Lyndon B. Johnson Space Center, Houston, Texas

The fluid bubble eliminator (FBE) is adevice that removes gas bubbles from aflowing liquid. The FBE contains nomoving parts and does not require anypower input beyond that needed topump the liquid. In the FBE, the buoy-ant force for separating the gas from theliquid is provided by a radial pressuregradient associated with a centrifugalflow of the liquid and any entrainedbubbles. A device based on a similarprinciple is described in “CentrifugalAdsorption Cartridge System” (MSC-22863), which appears on page 48 of thisissue. The FBE was originally intendedfor use in filtering bubbles out of a liq-uid flowing relatively slowly in a bioreac-tor system in microgravity. Versions thatoperate in normal Earth gravitation atgreater flow speeds may also be feasible.

The FBE (see figure) is constructed as acartridge that includes two concentriccylinders with flanges at the ends. Theouter cylinder is an impermeable housing;the inner cylinder comprises a gas-perme-able, liquid-impermeable membrane cov-ering a perforated inner tube. Multiplespiral disks that collectively constitute aspiral ramp are mounted in the space be-tween the inner and outer cylinders.

The liquid enters the FBE through anend flange, flows in the annular space be-tween the cylinders, and leaves throughthe opposite end flange. The spiral diskschannel the liquid into a spiral flow, thecircumferential component of whichgives rise to the desired centrifugal effect.The resulting radial pressure gradientforces the bubbles radially inward; that is,

toward the inner cylinder. At the innercylinder, the gas-permeable, liquid-imper-meable membrane allows the bubbles toenter the perforated inner tube whilekeeping the liquid in the space betweenthe inner and outer cylinders. The gasthus collected can be vented via an end-flange connection to the inner tube.

The centripetal acceleration (and thusthe radial pressure gradient) is approxi-mately proportional to the square of theflow speed and approximately inverselyproportional to an effective radius of theannular space. For a given FBE geometry,one could increase the maximum rate atwhich gas could be removed by increasingthe rate of flow to obtain more centripetalacceleration. In experiments and calcula-tions oriented toward the original micro-

gravitational application, centripetal accel-erations between 0.001 and 0.012 g [whereg ≡ normal Earth gravitation (≈9.8 m/s2)]were considered. For operation in normalEarth gravitation, it would likely be neces-sary to choose the FBE geometry and therate of flow to obtain centripetal accelera-tion comparable to or greater than g.

This work was done by Steve R. Gonda ofJohnson Space Center and Yow-Min D. Tsaoand Wenshan Lee of Wyle Laboratories. For fur-ther information, contact the Johnson Commer-cial Technology Office at 281-483-3809.

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-22996.

Outer CylinderSpiral Disks

Flange

Liquid In

Inner Cylinder: Membrane on Perforated Inner Tube

Flange

Vented Gas

Liquid Out

The Spiral Disks Channel the Liquid into a spiral flow, which gives rise to a pressure gradient that drives bubbles toward the inner cylinder.

In response to the loss of seven astro-nauts in the Space Shuttle Columbia dis-aster, large, lightweight, inflatable atmos-pheric-entry vehicles have been proposedas means of emergency descent and land-ing for persons who must abandon a

spacecraft that is about to reenter the at-mosphere and has been determined to beunable to land safely. Such a vehiclewould act as an atmospheric deceleratorat supersonic speed in the upper atmos-phere, and a smaller, central astronaut

pod could then separate at lower altitudesand parachute separately to Earth.

Astronaut-rescue systems that havebeen considered previously have beenmassive, and the cost of designing themhas exceeded the cost of fabrication of a

Inflatable Emergency Atmospheric-Entry VehiclesBallutes would act as inexpensive, lightweight atmospheric decelerator “lifeboats.”NASA’s Jet Propulsion Laboratory, Pasadena, California


space shuttle. In contrast, an inflatableemergency-landing vehicle according tothe proposal would have a mass between100 and 200 kg, could be stored in a vol-ume of approximately 0.2 to 0.4 m3, andcould likely be designed and built muchless expensively.

When fully inflated, the escape vehi-cle behaves as a large balloon para-chute, or ballute. Due to very low mass-per-surface area, a large radius, and alarge coefficient of drag, ballutes decel-erate at much higher altitudes and with

much lower heating rates than the spaceshuttle. Although the space shuttle at-mospheric reentry results in surfacetemperatures of about 1,600 oC, ballutescan be designed for maximum tempera-tures below 600 oC. This allows ballutesto be fabricated with lightweightZYLON®, or polybenzoxazole (PBO), orequivalent.

Two preliminary cocoon ballute“lifeboat” concepts are shown in the fig-ures. The cocoon portion of the vehiclewould be, more specifically, a capsule

pressurized to 1 bar (0.1 MPa — ap-proximately 1 standard atmosphere).Crewmembers would enter the cocoonpod and then zip it shut. The spacecraftwould be placed on a reentry trajectory,and the inflated cocoon with deflatedballute would be ejected.

Once the vehicle was safely away fromthe spacecraft, the entire ballute wouldbe inflated. For this inflation at high alti-tude, the ballute would be pressurized toabout 0.01 bar (1 kPa). As low as this pres-sure is, it is at least ten times the expecteddynamic pressure on the vehicle duringthe heating portion of very high atmos-pheric reentry, and hence it is sufficientto enable the ballute to retain its shape.From thermal reentry heating analysesperformed at JPL, the diameter of the in-flated ballute would be made largeenough (30 to 40 m) to limit the maxi-mum temperature to about 500 oC —safely below the 600 oC limit for PBO, orequivalent.

The spherical ballute shown in the upperfigure would have a mass of about 200 kgfor a seven-astronaut rescue mission, whilethe lens-shaped ballute in the lower figurehas been further improved by reducing theoverall mass required and increasing thecoefficient of drag. To maintain stability,the center of mass of both concepts mustbe kept low, and spin stabilization may benecessary.

This work was done by Jack Jones, JeffreyHall, and Jiunn Jeng Wu of Caltech forNASA’s Jet Propulsion Laboratory. Fur-ther information is contained in a TSP (seepage 1).NPO-40156

Flexible Insulation

Zipper EntrancesCapsule

Capsule

30 m

PBO Fabric BallutesDiameter 40 m

Aft Cover (PBO)

Forward Cover (PBO)

Radiation Shield

Torus Inflated to >1 Bar To Maintain Shape

(Note: To simplify illustration, the shapes are not to scale.)

A Spherical Ballute (upper) and a Lens-Shaped Ballute (lower) have been considered for inflatableemergency atmospheric-entry vehicles.

Lightweight Deployable Mirrors With Tensegrity SupportsExtremely lightweight, deployable structures could be made by assembling tensegrity modules.Marshall Space Flight Center, Alabama

The upper part of Figure 1 shows asmall-scale prototype of a developmentalclass of lightweight, deployable struc-tures that would support panels in pre-cise alignments. In this case, the panel ishexagonal and supports disks that rep-resent segments of a primary mirror of alarge telescope. The lower part of Figure1 shows a complete conceptual structurecontaining multiple hexagonal panelsthat hold mirror segments.

The structures of this class are of thetensegrity type, which was invented fivedecades ago by artist Kenneth Snelson. A

tensegrity structure consists of moment-free compression members (struts) andtension members (cables). The structuresof this particular developmental class areintended primarily as means to erect largesegmented primary mirrors of astronomi-cal telescopes or large radio antennas inouter space. Other classes of tensegritystructures could also be designed for ter-restrial use as towers, masts, and supportsfor general structural panels.

An important product of the presentdevelopment effort is the engineeringpractice of building a lightweight, de-

ployable structure as an assembly oftensegrity modules like the one shownin Figure 2. This module comprises twooctahedral tensegrity subunits that aremirror images of each other joined attheir plane of mirror symmetry. In thiscase, the plane of mirror symmetry isboth the upper plane of the lower sub-unit and the lower plane of the uppersubunit, and is delineated by the mid-height triangle in Figure 2. In the con-figuration assumed by the module tobalance static forces under mild loading,the upper and lower planes of each sub-


unit are rotated about 30°, relative toeach other, about the long (vertical) axisof the structure. Larger structures canbe assembled by joining multiple mod-ules like this one at their sides or ends.

When the module is compressed axi-ally (vertically), the first-order effect is

an increase in the rotation angle, but byvirtue of the mirror arrangement, thenet first-order rotation between the up-permost and lowermost planes is zero.The need to have zero net rotation be-tween these planes under all loadingconditions in a typical practical structureis what prompts the use of the mirrorconfiguration. Force and moment load-ings other than simple axial compres-sion produce only second-order defor-mations through strains in the struts andcables.

Simple algebraic expressions have beenderived to describe the deformations,under load, of multimodule platelike andmast structures, thereby making it possible

to design such structures without need forcomputers. Perhaps the most importantrules for designing a tensegrity structureare that (1) the lengths of the struts and ca-bles are critical and they determine the un-loaded shape of the structure, but that (2)the preloads (discussed in the next para-graph) in the cables and struts determinethe degree of rigidity under external load.

To make a module stowable, it is nec-essary to provide for disconnection ofthe ends of many of the struts and/ormake the struts collapsible (e.g., tele-scoping). To make a module deployable,one must provide means to reconnectthe struts if disconnected and re-extendthem if collapsed. The means of deploy-ment must include means to apply therequired preloads to the cables andstruts. In cases of manual stowage anddeployment, such means can includetoggles and turnbuckles. For automateddeployment, more sophisticated meansare needed. In the structure of Figure 2,the struts are telescoping piston/cylin-der units that are extended pneumati-cally and locked at full extension byspring-loaded mechanisms.

This work was done by Glenn W. Zeiders ofthe Sirius Group, Larry J. Bradford of CATFlight Services, and Richard C. Cleve of ElkRiver Engineering for Marshall Space FlightCenter. For further information, contact:

Glenn W. ZeidersSirius Group2803 Downing CourtHuntsville, AL 35801E-mail: [email protected]

Refer to MFS-31872.

Figure 1. A Tensegrity Structure supports a light-weight, thermally formed, hexagonal plasticpanel that, in turn, supports silicon disks thatrepresent segments of an astronomical mirror. Afully developed version would comprise a hexag-onal array of multiple hexagonal panels on asupporting tensegrity structure.

Figure 2. This Tensegrity Module comprises twosubunits that, together, undergo zero net rota-tion (to first order) under a longitudinal (verticalin this view) load.


Machinery/Automation

Centrifugal Adsorption Cartridge SystemNotable features include efficient collection of bioproducts and removal of bubbles.Lyndon B. Johnson Space Center, Houston, Texas

The centrifugal adsorption cartridgesystem (CACS) is an apparatus that re-covers one or more bioproduct(s) from adilute aqueous solution or suspensionflowing from a bioreactor. The CACS canbe used both on Earth in unit gravity andin space in low gravity. The CACS can beconnected downstream from the biore-actor; alternatively, it can be connectedinto a flow loop that includes the biore-actor so that the liquid can be recycled.

A centrifugal adsorption cartridge inthe CACS (see figure) includes two con-centric cylinders with a spiral ramp be-tween them. The volume between theinner and outer cylinders, and betweenthe turns of the spiral ramp is packedwith an adsorbent material. The innercylinder is a sieve tube covered with a gas-permeable, hydrophobic membrane.

During operation, theliquid effluent from thebioreactor is introduced atone end of the spiral ramp,which then constrains theliquid to flow along the spi-ral path through the adsor-bent material. The spiralramp also makes the flowmore nearly uniform thanit would otherwise be, andit minimizes any channel-ing other than that of thespiral flow itself.

The adsorbent material isformulated to selectively cap-ture the bioproduct(s) of in-terest. The bioproduct(s)can then be stored in boundform in the cartridge or elseeluted from the cartridge.

The centrifugal effect of the spiralflow is utilized to remove gas bubblesfrom the liquid. The centrifugal effectforces the bubbles radially inward, to-ward and through the membrane of theinner cylinder. The gas-permeable, hy-drophobic membrane allows the bub-bles to enter the inner cylinder whilekeeping the liquid out. The bubbles thatthus enter the cylinder are vented to theatmosphere. The spacing between theramps determines rate of flow along thespiral, and thereby affects the air-bubble-removal efficiency. The spacing betweenthe ramps also determines the length ofthe fluid path through the cartridge ad-sorbent, and thus affects the bioproduct-capture efficiency of the cartridge.

Depending on the application, sev-eral cartridges could be connected in a

serial or parallel flow arrangement. Aparallel arrangement can be used to in-crease product-capturing and flow ca-pacities while maintaining a low pres-sure drop. A serial arrangement can beused to obtain high product-capturingcapacity; alternatively, series-connectedcartridges can be packed with differentadsorbents to capture different bio-products simultaneously.

This work was done by Steve R. Gonda ofJohnson Space Center and Yow-Min D.Tsao and Wenshan Lee of Wyle Laboratories.

This invention is owned by NASA, and apatent application has been filed. Inquiriesconcerning nonexclusive or exclusive licensefor its commercial development should beaddressed to the Patent Counsel, JohnsonSpace Center, (281) 483-0837. Refer toMSC-22863.

Liquid From Bioreactor In

Bubbles RemovedFrom Cartridge

Inner Cylinder (Sieve TubeCovered by Membrane)

Adsorbent PackedBetween the Spirals

CENTRIFUGAL ADSORPTION CARTRIDGE

Liquid Out MULTIPLE CARTRIDGES IN SERIES

MULTIPLE CARTRIDGES IN PARALLEL

SpiralRamp

In a Centrifugal Adsorption Cartridge, the liquid effluent from a bioreactor is channeled along a spiral flow path througha selectively adsorbent material. The length of the spiral path contributes to efficient collection of a bioproduct sus-pended or dissolved in the liquid. Multiple centrifugal adsorption cartridges can be connected in series or parallel.

Ultrasonic Apparatus for Pulverizing Brittle MaterialCharacteristics include light weight, low preload, and low power demand.NASA’s Jet Propulsion Laboratory, Pasadena, California

The figure depicts an apparatus thatpulverizes brittle material by means of acombination of ultrasonic and sonic vi-bration, hammering, and abrasion. The

basic design of the apparatus could bespecialized to be a portable version foruse by a geologist in collecting pow-dered rock samples for analysis in the

field or in a laboratory. Alternatively, alarger benchtop version could be de-signed for milling and mixing of precur-sor powders for such purposes as synthe-


sis of ceramic and other polycrystallinematerials or preparing powder samplesfor x-ray diffraction or x-ray fluores-cence measurements to determine crys-talline structures and compositions.Among the most attractive characteris-tics of this apparatus are its light weight

and the ability to function without needfor a large preload or a large power sup-ply: It has been estimated that a portableversion could have a mass <0.5 kg, wouldconsume less than 1 W·h of energy inmilling a 1-cm3 volume of rock, andcould operate at a preload <10 N.

The basic design and principle of op-eration of this apparatus are similar tothose of other apparatuses described in aseries of prior NASA Tech Briefs articles,the two most relevant being “Ultra-sonic/Sonic Drill/Corers With Inte-grated Sensors” (NPO-20856), Vol. 25,No. 1 (January 2001), page 38 and “Ul-trasonic/Sonic Mechanisms for DeepDrilling and Coring” (NPO-30291), Vol.27, No. 9 (September 2003), page 65. Asbefore, vibrations are excited by meansof a piezoelectric actuator, an ultrasonichorn, and a mass that is free to move ax-ially over a limited range. As before, theultrasonic harmonic motion of the horndrives the free-mass in a combination ofultrasonic harmonic and lower-fre-quency hammering motion.

In this case, the free-mass is confinedwithin a hollow cylinder that serves as acrushing chamber, and the free-massserves as a crushing or milling tool. Thehammering of the free-mass against a ma-terial sample at the lower end of thechamber grinds the sample into powderin a relatively short time. The restrictionof the free-mass to axial motion onlymakes the grinding very efficient. Thefree-mass can be fabricated to have teethon its lower face to enhance the grindingeffect. Optionally, there can be a hole atthe bottom of the chamber covered with asieve to tailor the size distribution of thepowder leaving the crushing chamber.

This work was done by Stewart Sherrit, Xi-aoqi Bao, Yoseph Bar-Cohen, Benjamin Dol-gin, and Zensheu Chang of Caltech forNASA’s Jet Propulsion Laboratory. Fur-ther information is contained in a TSP (seepage 1).NPO-30682

Piezoelectric Stack/HornActuator

Base Cut To AllowPowder To Leave

Chamber

CrushingChamber

Rock Sample Prior to Crushing

Free-Mass/Crushing Bit

Powder Sample After Crushing

Driven by the Piezoelectric Stack/Horn Actuator, the free-mass hammers the rock sample in the crush-ing chamber. Particles of rock leave the chamber through a sieve-covered hole at the bottom.


Bio-Medical

A novel treatment for retinal degener-ative disorders involving transplantationof cells into the eye is currently underdevelopment at NASA Ames ResearchCenter and Stanford University Schoolof Medicine. The technique uses buckypaper as a support material for retinalpigment epithelial (RPE) cells, iris pig-ment epithelial (IPE) cells, and/or stemcells. This technology is envisioned as atreatment for age-related macular de-generation, which is the leading cause ofblindness in persons over age 65 in West-ern nations. Additionally, patients withother retinal degenerative disorders,such as retinitis pigmentosa, may betreated by this strategy.

Bucky paper is a mesh of carbon nan-otubes (CNTs), as shown in Figure 1,that can be made from any of the com-mercial sources of CNTs. Bucky paper isbiocompatible and capable of support-ing the growth of biological cells. Be-cause bucky paper is highly porous, nu-trients, oxygen, carbon dioxide, andwaste can readily diffuse through it.The thickness, density, and porosity ofbucky paper can be tailored in manu-facturing. For transplantation of cellsinto the retina, bucky paper serves si-multaneously as a substrate for cellgrowth and as a barrier for new bloodvessel formation, which can be a prob-lem in the exudative type of maculardegeneration.

Bucky paper is easily handled dur-ing surgical implantation into the eye.Through appropriate choice of manu-facturing processes, bucky paper can

be made relatively rigid yet able toconform to the retina when the buckypaper is implanted. Bucky paper of-fers a distinct advantage over othermaterials that have been investigatedfor retinal cell transplantation — lenscapsule and Descemet’s membrane —which are difficult to handle during

surgery because they are flimsy and donot stay flat.

In preparation for implantation, theselected cells are first cultured onto apiece of bucky paper. A retinotomy isthen performed, the cell-covered buckypaper is implanted, and the retina isreattached. Because bucky paper does

Transplanting Retinal Cells Using Bucky Paper for SupportBucky paper supports the cells before, during, and after surgery.Ames Research Center, Moffett Field, California

Figure 1. The Mesh of Carbon Nanotubes in BuckyPaper can be seen in this high magnification scan-ning electron micrograph .

a

b

Figure 2. Micrographs of RPE Cells illustrate the following: (a) human RPE cells cultured on buckypaper, as shown in this scanning electron micrograph, form a monolayer which is suitable for trans-plantation into the retina and (b) light micrograph of human RPE cells (stained blue) cultured onbucky paper (black) viewed in cross section.


not trigger an inflammatory reaction inthe eye, it can be left in place after trans-plantation to serve as a basement mem-brane patch.

The attachment of RPE (see Figure2), IPE, and/or stem cells onto thebucky paper may be enhanced by chem-ically modifying or coating the buckypaper with one or more biologically ac-tive substances. The ability to easily

make these modifications may serve asan important way of optimizing retinalcell transplantation for macular degen-eration and retinitis pigmentosa andmay facilitate other ophthalmologic ap-plications as well.

This patent pending work was performedby David J. Loftus, Martin Cinke, andMeyya Meyyappan of Ames Research Cen-ter, Center for Nanotechnology, and by Har-

vey Fishman, Ted Leng, Philip Huie, andKalayaan Bilbao of Stanford UniversitySchool of Medicine, Department of Ophthal-mology. Further information is contained in aTSP (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-14940.

Using an Ultrasonic Instrument to Size Extravascular BubblesMeasurements could be used to guide prebreathing of oxygen to reduce the risk ofdecompression sickness.Lyndon B. Johnson Space Center, Houston, Texas

In an ongoing development project,microscopic bubbles in extravascular tis-sue in a human body will be detected byuse of an enhanced version of the appa-ratus described in “Ultrasonic Bubble-Sizing Instrument” (MSC-22980), NASATech Briefs, Vol. 24, No. 10 (October2000), page 62. To recapitulate: Thephysical basis of the instrument is theuse of ultrasound to excite and measurethe resonant behavior (oscillatory ex-pansion and contraction) of bubbles.The resonant behavior is a function ofthe bubble diameter; the instrument ex-ploits the diameter dependence of theresonance frequency and the generalnonlinearity of the ultrasonic responseof bubbles to detect bubbles and poten-tially measure their diameters.

In the cited prior article, the applica-tion given most prominent mention wasthe measurement of gaseous emboli (es-

sentially, gas bubbles in blood vessels)that cause decompression sickness andcomplications associated with car-diopulmonary surgery. According to thepresent proposal, the instrument capa-bilities would be extended to measureextravascular bubbles with diameters inthe approximate range of 1 to 30 µm.

The proposed use of the instrumentcould contribute further to the under-standing and prevention of decompres-sion sickness: There is evidence that sug-gests that prebreathing oxygen greatlyreduces the risk of decompression sick-ness by reducing the number of micro-scopic extravascular bubbles. By usingthe ultrasonic bubble-sizing instrumentto detect and/or measure the sizes ofsuch bubbles, it might be possible to pre-dict the risk of decompression sickness.The instrument also has potential as atool to guide the oxygen-prebreathing

schedules of astronauts; high-altitudeaviators; individuals who undertakehigh-altitude, low-opening (HALO)parachute jumps; and others at risk ofdecompression sickness. For example,an individual at serious risk of decom-pression sickness because of high con-centrations of extravascular microscopicbubbles could be given a warning to con-tinue to prebreathe oxygen until it wassafe to decompress.

This work was done by Patrick J. Magari,Robert J. Kline-Schoder, and Marc A. Kentonof Creare, Inc., for Johnson Space Center.For further information, contact:

Creare, Inc.P.O. Box 71Hanover, NH 03755Phone: (603) 643-3800Fax: (603) 643-4657E-mail: [email protected]

Refer to MSC-23128.


Physical Sciences

Coronagraphic Notch Filter for Raman SpectroscopyDesign could be optimized for attenuating pump light and transmitting Raman-scattered light.NASA’s Jet Propulsion Laboratory, Pasadena, California

A modified coronagraph has been pro-posed as a prototype of improved notchfilters in Raman spectrometers. Corona-graphic notch filters could offer alterna-tives to both (1) the large and expensivedouble or triple monochromators in olderRaman spectrometers and (2) holo-graphic notch filters, which are less ex-pensive but are subject to environmentaldegradation as well as to limitations ofgeometry and spectral range.

Measurement of a Raman spectrum is anexercise in measuring and resolving faintspectral lines close to a bright peak: InRaman spectroscopy, a monochromaticbeam of light (the pump beam) excites asample of material that one seeks to ana-lyze. The pump beam generates a smallflux of scattered light at wavelengths slightlygreater than that of the pump beam. Theshift in wavelength of the scattered lightfrom the pump wavelength is known in theart as the Stokes shift. Typically, the flux ofscattered light is of the order of 10–7× thatof the pump beam and the Stokes shift liesin the wave-number range of 100 to 3,000cm–1. A notch filter can be used to suppressthe pump-beam spectral peak while passingthe nearby faint Raman spectral lines.

The basic principles of design and op-eration of a coronagraph offer an oppor-tunity for engineering the spectral trans-mittance of the optics in a Ramanspectrometer. A classical coronagraphmay be understood as two imaging sys-tems placed end to end, such that the firstsystem forms an intermediate real imageof a nominally infinitely distant object andthe second system forms a final real imageof the intermediate real image. If the lightincident on the first telescope is colli-mated, then the intermediate image is apoint-spread function (PSF). If an appro-priately tailored occulting spot (e.g., aGaussian-apodized spot with maximumabsorption on axis) is placed on the inter-mediate image plane, then the instru-ment inhibits transmission of light froman on-axis source. However, the PSFs ofoff-axis light sources are formed off axis— that is, away from the occulting spot —so that they become refocused onto thefinal image plane.

A properly designed coronagraph uti-lizes the diffraction from the intermediateocculting spot. In the exit-pupil plane, thisdiffraction forms a well-defined ringimage in the vicinity of the geometric

image of the exit pupil. By placing anaperture stop sized to block the passage ofthe diffracted light (such an aperture isknown in the art as a Lyot stop) in the exit-pupil plane, it is possible, in principle, toobtain an extremely high rejection ratio.

While coronagraphs are not new, re-cent developments make it possible to en-hance performance. One such develop-ment is that of the ability to write arbitraryabsorption patterns on occulting spots atsubmicron resolution by use of electron-beam lithography. Another such develop-ment is that of superpolished optics.

One characteristic of a classical corona-graph essential to the proposed notch fil-ter is that within the narrow typicalRaman spectral range associated with agiven pump laser line, the size of the PSFchanges little with wavelength. However,the position of the PSF (in particular, itsdisplacement from the occulting spot)can be made to vary considerably withwavelength by introducing a diffractiongrating or other dispersive element intothe optical train. Hence, one could obtainan extraordinarily sharp notch in thespectral transmittance of a coronagraphicfilter by designing the dispersive element

Incoming Light

Aperture

FlatSteering

Mirror

DiffractionGrating

Air-SpacedDoublet Lens

OcculterPlane

Spherical-SurfaceDoublet Lens

Final Image Plane(Detector Plane)

Lyot Stop

A Coronagraphic Filter would include a modified coronagraph equipped with a diffraction grating and other components. The modification would be suchas to optimize the functioning of the resulting instrument as a narrowband rejection (notch) filter.

An improved method of calibrating alarge directional radio antenna of the typeused in deep-space communication andradio astronomy has been developed. Thismethod involves a raster-scanning-and-measurement technique denoted on-the-fly (OTF) mapping, applied in considera-tion of the results of a systematic analysisof the entire measurement procedure.Phenomena to which particular attentionwas paid in the analysis include (1) thenoise characteristics of a total-power ra-diometer (TPR) that is used in the mea-surements and (2) tropospherically in-

duced radiometer fluctuations. Themethod also involves the use of recentlydeveloped techniques for acquisition andreduction of data. In comparison withprior methods used to calibrate such an-tennas, this method yields an order-of-magnitude improvement in the precisionof determinations of antenna aperture ef-ficiency, and improvement by a factor offive or more in the precision of determi-nation of pointing error and beam width.

Prerequisite to a meaningful descriptionof the present method is some backgroundinformation concerning three aspects of

the problem of calibrating an antenna ofthe type in question:• In OTF mapping measurements in

which a TPR is used, the desired data arethe peak temperature corresponding to aradio source, the pointing offset whenthe antenna is commanded to point to-ward the source, and the shape of themain lobe of the antenna beam, all asfunctions of the antenna beam elevationand azimuth angles. These data enableone to calculate the (1) antenna aper-ture efficiency by comparing the mea-sured peak temperature with that ex-pected for a 100-percent-efficientantenna, (2) the mechanical pointingerror resulting from small misalignmentsof various parts of the antenna structure,and (3) misalignments of the antennasubreflector and other mirrors.

• For practical reasons having to do with ob-taining adequate angular resolution andall-sky coverage, it is necessary to performazimuth and elevation scans fairly rapidly.

• Many natural radio sources used in cali-brating antennas are only approximatelypointlike: some sources subtend anglesgreater than the beam width of a givenantenna. In such a case, the antenna par-tially resolves the source structure anddoes not collect all of the radiation emit-ted by the source. This makes it neces-sary to estimate how much of the totalknown radiation from the source wouldactually be collected by the antenna if itwere 100-percent efficient. The resultingestimate, leading to a source-size correc-tion factor, introduces another degree ofuncertainty to the measurements. OTFmapping can remove this uncertainty.The key to using OTF mapping to solve

all three aspects of the calibration problem


and the other coronagraphic optics sothat at the pump wavelength, the PSF iscentered on the occulting spot.

The figure shows the optical layout ac-cording to one possible design of the pro-posed coronagraphic filter for a pumpwavelength of 550 nm. The dispersive el-ement would be a 500-line-per-millimeterdiffraction grating, of which the first-order diffraction would be utilized. Afterpassing through an aperture, the incom-ing light would strike the grating, fol-lowed by a flat steering mirror. An air-

spaced doublet lens incorporating an as-pherical element would generate a PSF atthe occulter (intermediate-image) plane.A spherical-surface doublet lens wouldreimage the light onto a detector plane.On its way to the detector plane, the lightwould pass though a Lyot stop. In princi-ple, a linear array of photodetectorscould be placed in the final image planeto measure the Raman spectrum. Thedepth of the notch at the pump wave-length, as well as other parameters of theperformance of the coronagraphic filter,

could be tailored through the choice ofthe parameters of the optical compo-nents, including especially the dispersionof the grating; the aperture diameter,focal length, and aberrations of the firstdoublet lens; the length of the occultingspot along the axis of dispersion; and thediameter of the Lyot stop.

This work was done by David Cohen andRobert Stirbl of Caltech for NASA’s Jet Propul-sion Laboratory. Further information is con-tained in a TSP (see page 1).NPO-30504

156

148

140

132

0.125°Cross-Elevation

Angle

0.125°Cross-Elevation

Angle

Venus Observed by DSS-13 at 58° Elevation13.8 - GHz, 33 x 33 Array

Noi

se T

empe

ratu

re, K

elvi

ns

On-the-Fly Mapping for Calibrating Directional AntennasSource-size corrections are not necessary in this method.NASA’s Jet Propulsion Laboratory, Pasadena, California

The TPR Readout Data plotted here were acquired in a test raster scan of a portion of the sky, near anelevation angle of 58°, that contained the planet Venus when it was relatively close to Earth. The hor-izontal axes on this plot correspond to elevation and cross-elevation angles. The vertical axis repre-sents noise temperature in Kelvins.


is to maintain a constant, known angularvelocity when scanning the antenna alonga given direction. To ensure alignment ofthe individual subscans within the fullraster, the angular position of the first datapoint of each subscan is determined fromreadings of azimuth- and elevation-angleencoders, while the angular positions ofthe rest of the subscan data points are de-termined by timing at the constant angularvelocity. Hence, if the TPR reading is sam-pled at a constant known rate, then the rel-ative angular position at which each datumis taken is known with high accuracy, andantenna-settling time is no longer an issue.The data-acquisition algorithms used in

OTF mapping provide for computation ofthe angular positions of radio sources,such that at any given time, the position ofthe antenna relative to a source is known.

The acquisition of data in the OTFmode necessarily entails attenuation ofhigh-frequency information as a conse-quence of the integration that occurs dur-ing the sampling intervals. The high-fre-quency information can be recovered inan inverse-filtering computation.

Even though the antenna beam does notsample all of the radiation from an ex-tended radio source at a given instant, thecompleted raster scan does cover the entiresolid angle subtended by the source and,

hence contains a sampling of all the radia-tion from that source. Consequently, nosource-size correction is necessary in OTFmapping. The resulting set of data regis-tered on a two-dimensional field of sam-pling points (see figure) can be used to de-termine a least-squares-best-fit main beampattern. The calibration parameters canthen be determined from the main beampattern.

This work was done by David Rochblatt, PaulRichter, and Philip Withington of Caltech forNASA’s Jet Propulsion Laboratory. Furtherinformation is contained in a TSP (see page 1).NPO-30648

Working Fluids for Increasing Capacities of Heat PipesFluids are formulated to make surface tensions increase with temperature.John H. Glenn Research Center, Cleveland, Ohio

A theoretical and experimental investi-gation has shown that the capacities ofheat pipes can be increased through suit-able reformulation of their working flu-ids. The surface tensions of all of theworking fluids heretofore used in heatpipes decrease with temperature. As ex-plained in more detail below, the limitson the performance of a heat pipe are as-sociated with the decrease in the surfacetension of the working fluid with temper-ature, and so one can enhance perfor-mance by reformulating the workingfluid so that its surface tension increaseswith temperature. This improvement isapplicable to almost any kind of heat pipein almost any environment.

The heat-transfer capacity of a heatpipe in its normal operating-temperaturerange is subject to a capillary limit and aboiling limit. Both of these limits are as-sociated with the temperature depen-dence of surface tension of the working

fluid. In the case of a traditional workingfluid, the decrease in surface tension withtemperature causes a body of the liquidphase of the working fluid to move to-ward a region of lower temperature, thuspreventing the desired spreading of theliquid in the heated portion of the heatpipe. As a result, the available capillary-pressure pumping head decreases as thetemperature of the evaporator end of theheat pipe increases, and operation be-comes unstable.

Water has widely been used as a workingfluid in heat pipes. Because the surfacetension of water decreases with increasingtemperature, the heat loads and other as-pects of performance of heat pipes thatcontain water are limited. Dilute aqueoussolutions of long-chain alcohols haveshown promise as substitutes for water thatcan offer improved performance, becausethese solutions exhibit unusual surface-tension characteristics: Experiments have

shown that in the cases of an aqueous so-lution of an alcohol, the molecules ofwhich contain chains of more than fourcarbon atoms, the surface tension in-creases with temperature when the tem-perature exceeds a certain value.

There are also other liquids that havesurface tensions that increase with temper-ature and could be used as working fluidsin heat pipes. For example, as a substitutefor ammonia, which is the working fluid insome heat pipes, one could use a solutionof ammonia and an ionic surfactant.

This work was done by David F. Chao ofGlenn Research Center and Nengli Zhangof Ohio Aerospace Institute. Further informa-tion is contained in a TSP (see page 1).

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-17270.


Computationally-Efficient Minimum-Time Aircraft Routes in the Presence of WindsMinimum-time routes are achieved using 10 times less computational effort.Ames Research Center, Moffett Field, California

A computationally efficient algo-rithm for minimizing the flight time ofan aircraft in a variable wind field hasbeen invented. The algorithm, re-ferred to as Neighboring OptimalWind Routing (NOWR), is based uponneighboring-optimal-control (NOC)concepts and achieves minimum-timepaths by adjusting aircraft heading ac-cording to wind conditions at an arbi-trary number of wind measurementpoints along the flight route. TheNOWR algorithm may either be usedin a fast-time mode to compute mini-mum-time routes prior to flight, ormay be used in a feedback mode to ad-just aircraft heading in real-time. Bytraveling minimum-time routes insteadof direct great-circle (direct) routes,flights across the United States cansave an average of about 7 minutes,and as much as one hour of flight timeduring periods of strong jet-streamwinds. The neighboring optimalroutes computed via the NOWR tech-nique have been shown to be within1.5 percent of the absolute minimum-time routes for flights across the conti-nental United States. On a typical 450-MHz Sun Ultra workstation, theNOWR algorithm produces completeminimum-time routes in less than 40milliseconds. This corresponds to arate of 25 optimal routes per second.The closest comparable optimizationtechnique runs approximately 10times slower.

Airlines currently use various trial-and-error search techniques to deter-mine which of a set of commonly trav-eled routes will minimize flight time.These algorithms are too computation-ally expensive for use in real-time sys-tems, or in systems where many optimalroutes need to be computed in a shortamount of time. Instead of operating inreal-time, airlines will typically plan atrajectory several hours in advanceusing wind forecasts. If winds changesignificantly from forecasts, the result-ing flights will no longer be minimum-time. The need for a computationally

efficient wind-optimal routing algo-rithm is even greater in the case of newair-traffic-control automation concepts.For air-traffic-control automation, thou-sands of wind-optimal routes may needto be computed and checked for con-flicts in just a few minutes. These fac-tors motivated the need for a more effi-cient wind-optimal routing algorithm.

The NOWR algorithm is a special typeof perturbation feedback control asshown in the figure. The nominal windsare modeled as system states, and areconsidered to be zero magnitude so thatthe nominal trajectory solution is simplya great-circle route between origin anddestination. The actual winds are inputas perturbations to the nominal winds,multiplied by time-varying NOWR gains,and then fed back as heading commandperturbations to achieve a minimum-

time trajectory solution. The NOWRgains are computed using techniquesfrom the calculus of variations. Becausethe nominal route is a great circle, andbecause the nominal winds are zeromagnitude, the NOWR feedback gainsmay be normalized and applied toflights at any airspeed between any ori-gin and destination using coordinate ro-tations and simple time and distancescaling. This one-solution-fits-all aspectof NOWR makes it a very powerful andpractical technique.

The implementation procedure forNOWR is to either run a fast-time simu-lation to compute an optimal windroute, which can then be used as thebasis for a filed flight plan, or one mayuse NOWR in real-time to achieve wind-optimal routes in a future free flight en-vironment.

NominalSolution NOC Gain

MatrixDynamicsModule

∑

unom

xnom ∆x ∆u u x+

++

–Sum Module

Feedforward and Feedback quantities are computed in an NOC approach to adjustment of controlvariables for optimization of the route taken by an aircraft.

Information Sciences


The optimization process begins byfirst rotating coordinates so that theorigin and destination points appear tobe located along the equator of the ro-tated coordinate system. The distanceunit is then scaled such that the dis-tance from origin to destination isunity. The time unit is similarly scaledsuch that the airspeed of the aircraft isalso unity. Because of this scaling, anyflight between any two points appearsto be mathematically similar. Thewinds along the flight route are thenobtained from the Rapid Update Cycle(RUC), a national gridded wind systemdeveloped at the National Oceanic and

Atmospheric Administration. TheNOWR algorithm may be adapted touse as many wind measurement pointsalong the route as are desired, but inpractice, about 10 or 15 points are suf-ficient. At each of these points, the pre-dicted winds are multiplied by theNOWR gains to determine the optimalheading angles the aircraft should useto fly a minimum-time route. Once theoptimal heading and the coordinatesof the minimum-time route have beencomputed in the normalized coordi-nate system, the solution may be trans-formed back to real Earth coordinatesthrough the inverse rotation equations.

The reason why this perturbationscheme is so efficient is that it is a sim-ple linear feedback algorithm involv-ing just a few algebraic steps. The ex-cellent performance of NOWR inpractice is achieved because winds typ-ically vary in a smooth manner and donot contain many sharp nonlinearitiesor discontinuities.

This work was done by Matthew R. Jardinof Ames Research Center. Further infor-mation 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-14554.


Books & Reports

Liquid-Metal-Fed PulsedPlasma Thrusters

A short document proposes liquid-metal-fed pulsed plasma thrusters for smallspacecraft. The propellant liquid for such athruster would be a low-melting-tempera-ture metal that would be stored molten inan unpressurized, heated reservoir andwould be pumped to the thruster by a mag-netohydrodynamic coupler. The liquidwould enter the thruster via a metal tubeinside an electrically insulating ceramictube. A capacitor would be connected be-tween the outlet of the metal tube and theouter electrode of the thruster. The pump-ing would cause a drop of liquid to form atthe outlet, eventually growing largeenough to make contact with the outerelectrode. Contact would close the circuitthrough the capacitor, causing the capaci-tor to discharge through the drop. The ca-pacitor would have been charged withenough energy that the discharge wouldvaporize, ionize, and electromagneticallyaccelerate the contents of the metal drop.The resulting plasma would be ejected at aspeed of about 50 km/s. The vaporizationof the drop would reopen the circuitthrough the capacitor, enabling recharg-ing of the capacitor. As pumping contin-ued, a new drop would grow and theprocess would repeat.

This work was done by Thomas Markusicof Marshall Space Flight Center. For fur-ther information, contact Brian Johnson,MSFC Commercialization Assistance Lead, at(256) 544-3518, [email protected],or access the Technical Support Package(TSP) (see page 1).MFS-31962

Personal Radiation Protection System

A report describes the personal radi-ation protection system (PRPS), whichhas been invented for use on the Inter-national Space Station and other space-craft. The PRPS comprises walls thatcan be erected inside spacecraft, whereand when needed, to reduce theamount of radiation to which person-nel are exposed. The basic structuralmodules of the PRPS are pairs of 1-in.(2.54-cm)-thick plates of high-densitypolyethylene equipped with fasteners.The plates of each module are assem-bled with a lap joint. The modules aredenoted bricks because they are de-signed to be stacked with overlaps, in amanner reminiscent of bricks, to build2-in. (5.08-cm)-thick walls of variouslengths and widths. The bricks are oftwo varieties: one for flat wall areas andone for corners. The corner bricks arespecialized adaptations of the flat-areabricks that make it possible to join wallsperpendicular to each other. Bricks areattached to spacecraft structures and toeach other by use of straps that can betightened to increase the strengths andstiffnesses of joints.

This work was done by Mark McDonald ofJohnson Space Center and Victoria Vinciof Johnson Engineering Corp. For further in-formation, contact:

Johnson Engineering Corp.18100 Upper Bay Road, Suite 220Houston, TX 77058-3547Telephone No.: (281) 333-9729

Refer to MSC-23330.

Attitude Control for a Solar-Sail Spacecraft

A report discusses the attitude-controlsystem of a proposed spacecraft that wouldderive at least part of its propulsion from asolar sail. The spacecraft would include abus module containing three or more re-action wheels, a boom attached at one endto the bus module and attached at itsother end to a two-degree-of-freedom(DOF) gimbal at the nominal center ofmass of a sail module. Each DOF of thegimbal could be independently lockedagainst rotation or allowed to rotate freely.By using the reaction wheels to rotate thebus when at least one gimbal DOF was inthe free state, the center of mass (CM) ofthe spacecraft could be shifted relative tothe center of pressure (CP) on the solarsail. The resulting offset between the CMand CP would result in a solar torque,which could be used to change the atti-tude of the spacecraft. The report dis-cusses numerous aspects of the dynamicsand kinematics of the spacecraft, alongwith the relationships between these as-pects and the designs of such attitude-con-trol-system components as sensors, mo-tors, brakes, clutches, and gimbals.

This work was done by Edward Mettler andScott Ploen of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).NPO-30522

National Aeronautics andSpace Administration

Documents

Technology Focus Computers/Electronics · (256) 544-2615 [email protected] Stennis Space Center Robert Bruce (228) 688-1929 [email protected] Carl Ray Small Business