technology focus electronics/computersbetter. optionally, the system can easily be connected to a...

Technology Focus

Electronics/Computers

Software

Materials

Mechanics

Machinery/Automation

Manufacturing & Prototyping

Bio-Medical

Physical Sciences

Information Sciences

Books and Reports

08-06 August 2006

https://ntrs.nasa.gov/search.jsp?R=20110012943 2020-01-28T12:12:31+00:00Z

NASA Tech Briefs, August 2006 1

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 Innovative Partnerships Program (IPP), its documents, and services is also available at the same facility oron the World Wide Web at http://ipp.nasa.gov.

Innovative Partnerships Offices are located at NASA field centers to provide technology-transfer access toindustrial users. Inquiries can be made by contacting NASA field centers listed below.

Ames Research CenterLisa L. Lockyer(650) [email protected]

Dryden Flight Research CenterGregory Poteat(661) [email protected]

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

Jet Propulsion LaboratoryKen Wolfenbarger(818) [email protected]

Johnson Space CenterMichele Brekke(281) [email protected]

Kennedy Space CenterJim Aliberti(321) [email protected]

Langley Research CenterMartin Waszak(757) [email protected]

Glenn Research CenterRobert Lawrence(216) [email protected]

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

Stennis Space CenterJohn Bailey(228) 688-1660 [email protected]

Carl Ray, Program ExecutiveSmall Business Innovation Research (SBIR) & Small Business Technology Transfer (STTR) Programs(202) [email protected]

Merle McKenzieInnovative Partnerships Program Office(202) [email protected]

NASA Field Centers and Program Offices

5 Technology Focus: DataAcquisition

5 Measurement and Controls Data AcquisitionSystem

5 IMU/GPS System Provides Position and AttitudeData

5 Using Artificial Intelligence to Inform Pilots ofWeather

6 Fast Lossless Compression of Multispectral-ImageData

8 Developing Signal-Pattern-Recognition Programs

9 Implementing Access to Data Distributed onMany Processors

11 Electronics/Computers11 Compact, Efficient Drive Circuit for a Piezoelectric

Pump

12 Dual Common Planes for Time Multiplexing ofDual-Color QWIPs

13 MMIC Power Amplifier Puts Out 40 mW From 75to 110 GHz

15 Software15 2D/3D Visual Tracker for Rover Mast

15 Adding Hierarchical Objects to RelationalDatabase General-Purpose XML-BasedInformation Managements

17 Materials17 Vaporizable Scaffolds for Fabricating

Thermoelectric Modules

18 Producing Quantum Dots by Spray Pyrolysis

19 Machinery/Automation19 Mobile Robot for Exploring Cold Liquid/Solid

Environments

20 System Would Acquire Core and Powder Samplesof Rocks

21 Manufacturing & Prototyping21 Improved Fabrication of Lithium Films Having

Micron Features

22 Manufacture of Regularly Shaped Sol-Gel Pellets

23 Bio-Medical23 Regulating Glucose and pH, and Monitoring

Oxygen in a Bioreactor

25 Physical Sciences25 Satellite Multiangle Spectropolarimetric Imaging

of Aerosols

26 Interferometric System for Measuring Thicknessof Sea Ice

27 Microscale Regenerative Heat Exchanger

29 Information Sciences29 Protocols for Handling Messages Between

Simulation Computers

30 Statistical Detection of Atypical Aircraft Flights

30 NASA’s Aviation Safety and Modeling Project

33 Books & Reports33 Multimode-Guided-Wave Ultrasonic Scanning of

Materials

33 Algorithms for Maneuvering Spacecraft AroundSmall Bodies

33 Improved Solar-Radiation-Pressure Models forGPS Satellites

33 Measuring Attitude of a Large, Flexible, OrbitingStructure

08-06 August 2006


This document was prepared under the sponsorship of the National Aeronautics and Space Administration. Neither the United States Govern-ment nor any person acting on behalf of the United States Government assumes any liability resulting from the use of the information containedin this document, or warrants that such use will be free from privately owned rights.


Measurement and Controls Data Acquisition SystemMarshall Space Flight Center, Alabama

Measurement and Controls Data Ac-quisition System (MCDAS) is an applica-tion program that integrates the func-tions of two stand-alone programs: onefor acquisition of data, the other forcontrols. MCDAS facilitates and im-proves testing of complex engineeringsystems by helping to perform calibra-tion and setup of test systems and acqui-sition, dissemination, and processing ofdata. Features of MCDAS include an in-tuitive, user-friendly graphical user in-terface, a capability for acquiring data atrates greater than previously possible,

cooperation between the data-acquisi-tion software subsystem and alarm-checking and analytical components ofthe control software subsystem, and acapability for dissemination of datathrough fiber optics and virtual andwide-area networks, including networksthat contain hand-held display units.The integration of the data acquisitionand control software offers a safety ad-vantage by making alarm informationavailable to the control software in amore timely manner. By enabling theuse of hand-held devices, MCDAS re-

duces the time spent by technicians ask-ing for screen updates to determine ef-fects of setup actions. Previouslyrecorded data can be processed withoutinterruption to current acquisition ofdata. Analysts can continue to view testparameters while test-data files arebeing generated.

This program was written by Rick Hall ofMarshall Space Flight Center, AliceDaniel formerly of UNITes (NTSI), andFrank E. Batts, Sr. of Langley Research Cen-ter. Further information is contained in aTSP (see page 1). MFS-32014-1

A special navigation system is being de-veloped to provide high-quality informa-tion on the position and attitude of amoving platform (an aircraft or space-craft), for use in pointing and stabiliza-tion of a hyperspectral remote-sensingsystem carried aboard the platform. Thesystem also serves to enable synchroniza-tion and interpretation of readouts of allonboard sensors. The heart of the systemis a commercially available unit, smallenough to be held in one hand, that con-tains an integral combination of an iner-tial measurement unit (IMU) of the mi-croelectromechanical systems (MEMS)type, Global Positioning System (GPS)

receivers, a differential GPS subsystem,and ancillary data-processing subsystems.The system utilizes GPS carrier-phasemeasurements to generate time data plushighly accurate and continuous data onthe position, attitude, rotation, and ac-celeration of the platform. Relative toprior navigation systems based on IMUand GPS subsystems, this system issmaller, is less expensive, and performsbetter. Optionally, the system can easilybe connected to a laptop computer fordemonstration and evaluation. In addi-tion to airborne and spaceborne remote-sensing applications, there are numer-ous potential terrestrial sensing,

measurement, and navigation applica-tions in diverse endeavors that includeforestry, environmental monitoring, agri-culture, mining, and robotics.

This work was done by Ching Fang Linof American GNC Corp. for Stennis SpaceCenter.

Inquiries concerning rights for the commer-cial use of this invention should be addressed to:

American GNC Corporation888 Easy StreetSimi Valley, CA 93065E-mail: [email protected] to SSC-00225, volume and number

of this NASA Tech Briefs issue, and thepage number.

IMU/GPS System Provides Position and Attitude DataStennis Space Center, Mississippi

Using Artificial Intelligence to Inform Pilots of WeatherWeather awareness is increased; workload is reduced.Ames Research Center, Moffett Field, California

An automated system to assist a Gen-eral Aviation (GA) pilot in improvingsituational awareness of weather inflight is now undergoing development.This development is prompted by theobservation that most fatal GA accidentsare attributable to loss of weather aware-ness. Loss of weather awareness, inturn, has been attributed to the diffi-

culty of interpreting traditional pre-flight weather briefings and the diffi-culty of both obtaining and interpret-ing traditional in-flight weather briefings.The developmental automated systemnot only improves weather awarenessbut also substantially reduces the time apilot must spend in acquiring and main-taining weather awareness.

The automated system includes com-puter hardware and software, a speech-based hardware/software user interface,and hardware interfaces between thecomputer and aircraft radio-communi-cation equipment. The heart of the sys-tem consists of artificial-intelligence soft-ware, called Aviation WeatherEnvironment (AWE), that implements a

Technology Focus: Data Acquisition

human-centered methodology orientedtowards providing the weather informa-tion (1) that the pilot needs and/orwants, (2) at the appropriate time, and(3) in the appropriate format.

AWE can be characterized as a con-text-aware, domain-and-task knowledge-able, personalized, adaptive assistant.

AWE automatically monitors weather re-ports for the pilot’s flight route andwarns the pilot of any weather condi-tions outside the limits of acceptableweather conditions that the pilot hasspecified in advance. AWE provides tex-tual and/or graphical representationsof important weather elements overlaid

on a navigation map (see figure). Therepresentations depict current and fore-cast conditions in an easy-to-interpretmanner and are geographically posi-tioned next to each applicable airportto enable the pilot to visualize condi-tions along the route. In addition to au-tomatic warnings, the system enablesthe pilot to verbally request (via thespeech-based user interface) weatherand airport information.

AWE is context-aware in the follow-ing sense: From the location of the air-craft (as determined by a Global Posi-tioning System receiver) and the routeas specified by the pilot, AWE deter-mines the phase of flight. In determin-ing the timing of warnings and themanner in which warnings are issued,AWE takes account of the phase offlight, the pilot’s definition of accept-able weather conditions, and the pilot’spreferences for automatic notification.By noting the pilot’s verbal requests forinformation during the various phasesof flight, the system learns to providethe information, without explicit re-quests, at the corresponding times onsubsequent flights under similar condi-tions.

This work was done by Lilly Spirkovska ofAmes Research Center and Suresh K.Lodha of the University of California. Fur-ther information is contained in a TSP (seepage 1).

Inquiries concerning rights for the com-mercial use of this invention should be ad-dressed to the Patent Counsel, Ames Re-search Center, (650) 604-5104. Refer toARC-14970-1.

6 NASA Tech Briefs, August 2006

Current Weather and Winds Aloft are shown alongside a pilot-selected route. Wind velocity at thepilot-selected altitude is depicted graphically with black arrows. The current weather is shown usingsymbolic and textual representations.

ARC-14970-1 ABPI

12-01-04 EM

Fast Lossless Compression of Multispectral-Image DataA low-complexity adaptive-filtering algorithm is used.NASA’s Jet Propulsion Laboratory, Pasadena, California

An algorithm that effects fast losslesscompression of multispectral-imagedata is based on low-complexity, provenadaptive-filtering algorithms. This algo-rithm is intended for use in compress-ing multispectral-image data aboardspacecraft for transmission to Earth sta-tions. Variants of this algorithm couldbe useful for lossless compression ofthree-dimensional medical imageryand, perhaps, for compressing imagedata in general.

The main adaptive-filtering algorithmon which the present algorithm is basedis the sign algorithm (also known as the

sign-error algorithm and as the binaryreinforcement algorithm). The sign al-gorithm is related to the least-mean-square (LMS) algorithm. Both algo-rithms are briefly described in thefollowing two paragraphs.

Consider a sequence of image data(or any other data) that one seeks tocompress. The sequence is specified interms of a sequentially increasing index(k) and the value (dk) of the kth sample.An estimated value of the kth sample, k,is calculated by the equation

k = wTkuk,

where wk is a filter-weight vector at

index k and uk is an input vector thatcan be defined in any of a number ofdifferent ways, depending on the spe-cific application. Once the estimate k

has been calculated, the error betweenthe estimate and the exact value is calcu-lated as

ek = k – dk.When the LMS algorithm or sign algo-rithm is used as part of a predictive com-pression scheme, the sequence of ek val-ues is encoded in the compressedbitstream.

The error value is also used to updatethe filter weights in either of two ways, de-

d

d

d

d


pending on which algorithm is in use. Inthe LMS algorithm, the update equation is

wk+1 = wk – µukek.In the sign algorithm, the update equation is

wk+1 = wk – µuksgn(ek).In both update equations, µ is a pos-

itive, scalar step-size parameter thatcontrols the trade-off between conver-gence speed and average steady-stateerror. A smaller value of µ results inbetter steady-state performance butslower convergence. In some variantsof these algorithms, the value of µchanges over time.

In the present algorithm, the index kis taken as an abstract representation ofthree indices (x,y,z) that are the coordi-nates of the sample in the multispectraldataset. Specifically, x and y are the spa-tial coordinates and z denotes the spec-tral band. The signal level (equivalently,the sample value) for that location isrepresented by dk ≡ s(x,y,z).

For purposes of compression, an imagerepresented by a stream of data to be com-pressed is partitioned spatially into conve-

niently sized, fixed regions. The data arecompressed in the order in which they arereceived, maintaining separate statistics foreach band and switching among the bandsas necessary. The data from each region arecompressed independently of those fromother regions. Performing independentcompression calculations for each regionlimits the adverse effect of loss of data.

The input vector uk chosen for this al-gorithm contains values from a six-sam-ple prediction neighborhood of a sam-ple of interest: three values fromadjacent samples in the same spectralband and one sample each from thesame location in each of three preced-ing spectral bands. Specifically,

where (x,y,z) is a mean value of pre-vious samples in the vicinity of x,y inspectral band z. The stream of ek val-ues calculated by use of this uk is fur-ther compressed by use of Golombcodes.

In tests, the compression effective-ness of this algorithm was shown tobe competitive with that of the bestpreviously reported data-compres-sion algorithms of similar complex-ity. The table presents results fromone series of tests performed onmultispectral imagery acquired byNASA’s airborne visible/infraredimaging spectrometer (AVIRIS).

This work was done by Matthew Klimeshof Caltech for NASA’s Jet Propulsion Lab-oratory. Further information is contained ina TSP (see page 1).

The software used in this innovation isavailable for commercial licensing. Pleasecontact Karina Edmonds of the CaliforniaInstitute of Technology at (626) 395-2322.Refer to NPO-42517.

%s

uk

s x y z s x y z

s x y z s x y

=

− −− − −

( , , ) ( , , )

( , , ) ( , ,

1

1 1

%% zz

s x y z s x y z

s x y z s x y z

)

( , , ) ( , , )

( , , ) ( , ,

− −− − −1

1

%% 11

2 2

3

)

( , , ) ( , , )

( , , ) ( , ,

s x y z s x y z

s x y z s x y

− − −− −

%% zz −

3)

Data From AVIRIS Images of various scenes were compressed by the present algorithm and by a number of other algorithms. The numerical entries are thenumbers of bits per sample in the compressed data streams.

Cuprite 1

Cuprite 2

Cuprite 3

Cuprite 4

Jasper Ridge 1

Jasper Ridge 2

Jasper Ridge 3

Jasper Ridge 4

Jasper Ridge 5

Low Altitude 1

Low Altitude 2

Low Altitude 3

Low Altitude 4

Low Altitude 4

Low Altitude 6

Low Altitude 7

Lunar Lake 1

Lunar Lake 2

Moffett Field 1

Moffett Field 2

Moffett Field 3

Average

4.89

5.02

4.92

4.98

5.04

5.02

5.07

5.07

5.02

5.37

5.42

5.30

5.32

5.37

5.29

5.29

4.99

4.94

5.12

5.11

4.98

5.12

5.14

5.34

5.16

5.21

5.41

5.37

5.47

5.47

5.39

5.70

5.76

5.58

5.58

5.63

5.56

5.60

5.19

5.14

5.48

5.40

5.12

5.41

7.13

7.50

7.16

7.16

7.72

7.67

7.90

7.87

7.75

7.81

7.95

7.57

7.53

7.60

7.52

7.64

6.98

6.96

7.78

7.57

7.03

7.51

6.00

6.13

6.00

6.05

6.17

6.12

6.19

6.22

6.14

6.53

6.58

6.42

6.42

6.47

6.42

6.43

6.02

5.97

6.24

6.20

5.96

6.22

5.44

5.58

5.45

5.51

5.62

5.59

5.67

5.67

5.60

5.97

6.02

5.88

5.89

5.91

5.85

5.88

5.49

5.44

5.70

5.60

5.41

5.68

5.03

5.09

5.06

5.10

5.06

5.05

5.10

5.11

5.06

5.38

5.40

5.33

5.37

5.40

5.34

5.34

5.12

5.07

5.15

5.08

4.96

5.17

4.90

4.97

4.92

4.96

4.95

4.94

4.99

5.00

4.94

5.30

5.33

5.23

5.26

5.30

5.24

5.24

4.99

4.93

5.03

4.98

4.86

5.06

SLSQ-OPT(Rizzo et al.)*

SLSQ(Rizzo et al.)*

DifferentialJPEG-LS

Rice/USESMultispectral

JPEG-LS(2-D)ICER-3D

FastLosslessScene

PresentAlgorithm

OtherAlgorithms

* F. Rizzo, B. Carpentieri, G. Motta, and J.A. Storer, “Low-complexity lossless compression of hyperspectral imagery via linear prediction,” IEEE Signal Processing Letters, 12(2):138-141, February 2005.


Developing Signal-Pattern-Recognition ProgramsSoftware system aids development of application programs that analyze signals.Lyndon B. Johnson Space Center, Houston, Texas

Pattern Interpretation and RecognitionApplication Toolkit Environment (PI-RATE) is a block-oriented software systemthat aids the development of applicationprograms that analyze signals in real timein order to recognize signal patterns thatare indicative of conditions or events of in-terest. PIRATE was originally intended foruse in writing application programs to rec-ognize patterns in space-shuttle telemetrysignals received at Johnson Space Center’sMission Control Center: application pro-grams were sought to (1) monitor electriccurrents on shuttle ac power busses to rec-ognize activations of specific power-con-suming devices, (2) monitor various pres-sures and infer the states of affectedsystems by applying a Kalman filter to thepressure signals, (3) determine fuel-leakrates from sensor data, (4) detect faults ingyroscopes through analysis of systemmeasurements in the frequency domain,and (5) determine drift rates in inertialmeasurement units by regressing measure-ments against time. PIRATE can also beused to develop signal-pattern-recognitionsoftware for different purposes — for ex-ample, to monitor and control manufac-turing processes.

PIRATE was preceded by a customstripchart-analysis program that took along time to develop and offered littleopportunity for reuse. Also availableprior to the development of PIRATEwere commercial block-oriented devel-opment software systems that were usefulfor prototyping but exhibited significantlimitations: for example, they could notbe used to produce real-time applicationprograms, could not be used to developsoftware compatible with the hardwareand the other software of the MissionControl Center, could not be used to de-velop application programs that couldfunction in the face of the communica-tion difficulties (especially, intermittencyand errors) inherent in monitoring re-mote equipment, and could not providepattern-recognition capabilities. PIRATEovercomes these deficiencies to a largeextent, and goes beyond that by includ-ing a C-language interface that providesunprecedented flexibility.

PIRATE includes the following com-ponents:• The PIRATE data-flow language. An ap-

plication program is specified by use ofthe PIRATE data-flow language. An

application-program specification de-fines which data processing moduleswill be used in the program and estab-lishes the data flowing among themodules. Similarly, for building amodule, one specifies the flow of datainto and out of a module by use of thePIRATE language.

• The PIRATE predefined modules. PIRATEcontains several predefined modules,including ones for data communication,signal processing, and data filtering.Among these are software tools to filterout the highly non-Gaussian errors thatare typical of the communicationprocess while leaving the nonerroneousdata intact. (Most other signal-process-ing software filters that can remove non-Gaussian errors also undesirably modifythe underlying signals.) Also among thepredefined modules are a Bayesian clas-sifier and other software tools for inter-preting the contents of signals.

• The PIRATE code generator. The PIRATEcode generator translates an applica-tion-program-specification file and theassociated module-configuration filesinto a standard C-language file. Thisfile contains the main routine for theapplication program. A C compiler canthen compile and link this file to pro-duce an efficient real-time pattern-recognition application program.

• The PIRATE object library. The gener-ated code makes calls to several PI-RATE infrastructure routines. The PI-RATE object library contains theobject code for these infrastructureroutines and for the predefined mod-ule routines.

• The PIRATE “imake” facility. The imakeis a programming software tool thatwas developed to address issues ofportability pertaining to the X Win-dow System and to provide a high-level view of the software-buildingprocess. However, the standard imakesuite explicitly targets the construc-tion of X Window System software.The PIRATE imake facility takes ad-vantage of the standard imake suitewhere practical, but targets the con-struction of PIRATE and PIRATE ap-plication programs rather than X Win-dow System software.

• An architecture for the development of mod-ules by the user. PIRATE is intentionallyan open-ended software tool. While

simple application programs can beconstructed by use of the predefinedmodules, it is likely that a useful pat-tern-recognition application programscould not. Instead, domain-specificlogic can be expected to be necessary.PIRATE enables the implementationof domain-specific knowledge in thewidely used C programming language.The architecture for user-developedmodules specifies how such domainknowledge can be used in a PIRATEapplication program.PIRATE is used both in building a pat-

tern-recognition application programand in the real-time execution of thatprogram. To build an application pro-gram, one constructs an application-specification file, application-specificmodules, and application imake file.The imake file identifies the compo-nents that form the executable applica-tion program and directs constructionof these components from the sourcefiles developed by the user.

During execution of the pattern-recognition application program, thesource module of the program acquiresthe incoming data and uses the PIRATEinfrastructure to feed the data to down-stream modules. The infrastructuresends the appropriate data to the appro-priate modules, which operate on thedata. Each module uses the PIRATE in-frastructure to send its output to mod-ules downstream of it.

In PIRATE, transmission of data be-tween data-processing modules is alwaysperformed by calls to C functions thatare parts of an executable module. Inother software systems, data are oftentransmitted between modules via operat-ing systems; the computational overheadof doing so is often several orders ofmagnitude greater than that of PIRATE.In PIRATE, the processing of data withina module is performed by compiledfunctions. In other systems, the process-ing of data may be performed by inter-preters, which, again entail computa-tional overhead much greater than thatof PIRATE.

This program was written by Robert O.Shelton of Johnson Space Center andDavid Hammen of LinCom. For further in-formation, contact the Johnson TechnologyTransfer Office at (281) 483-3809.MSC-22944


A reference architecture is definedfor an object-oriented implementationof domains, arrays, and distributionswritten in the programming languageChapel. This technology primarily ad-dresses domains that contain arrays thathave regular index sets with the low-level implementation details being be-yond the scope of this discussion. Thetheoretical foundations are based uponthe work “A Semantic Framework forDomains, Arrays, and Distributions inChapel” by Hans Zima. What is definedis a complete set of object-oriented op-erators that allows one to perform datadistributions for domain arrays involv-ing regular arithmetic index sets. Whatis unique is that these operators allowfor the arbitrary regions of the arrays tobe fragmented and distributed acrossmultiple processors with a single pointof access giving the programmer the il-lusion that all the elements are collo-cated on a single processor. Today’s

massively parallel High ProductivityComputing Systems (HPCS) are charac-terized by a modular structure, with alarge number of processing and mem-ory units connected by a high-speed net-work. Locality of access as well as loadbalancing are primary concerns in thesesystems that are typically used for high-performance scientific computation.Data distributions address these issuesby providing a range of methods forspreading large data sets across thecomponents of a system. Over the pasttwo decades, many languages, systems,tools, and libraries have been developedfor the support of distributions. Sincethe performance of data parallel appli-cations is directly influenced by the dis-tribution strategy, users often resort tolow-level programming models thatallow fine-tuning of the distribution as-pects affecting performance, but, at thesame time, are tedious and error-prone.This technology presents a reusable de-

sign of a data-distribution frameworkfor data parallel high-performance ap-plications. Distributions are a means toexpress locality in systems composed oflarge numbers of processor and mem-ory components connected by a net-work. Since distributions have a great ef-fect on the performance ofapplications, it is important that the dis-tribution strategy is flexible, so its be-havior can change depending on theneeds of the application. At the sametime, high productivity concerns re-quire that the user be shielded fromerror-prone, tedious details such ascommunication and synchronization.

This program was written by Mark Jamesof Caltech for NASA’s Jet Propulsion Lab-oratory. Further information is contained ina TSP (see page 1).

This software is available for commerciallicensing. Please contact Karina Edmonds ofthe California Institute of Technology at(626) 395-2322. Refer to NPO-42506.

Implementing Access to Data Distributed on Many ProcessorsNASA’s Jet Propulsion Laboratory, Pasadena, California


Electronics/Computers

Figure 1 depicts a switching circuitthat drives a piezoelectric pump at ahigh voltage with a polarity that alter-nates at the desired mechanical pumpfrequency. This circuit offers advantagesof (1) high energy efficiency relative toconventional direct-drive circuits and(2) compactness relative to conventionalresonant drive circuits.

Conventional direct-drive circuits areinefficient because the piezoelectric ac-tuators in the pump behave electricallyas large capacitors and consume only asmall part of the energy supplied to themduring each pump cycle. Conventionalresonant drive circuits are more effi-cient, but they must include inductivecoils, which can be unacceptably large.

The present switching drive circuit in-cludes inductive coils, but they are smalland not used to resonate the actuator ca-pacitance; instead, the charging of the

actuator capacitances and the ex-changes of energy among inductive andcapacitive circuit elements are accom-plished (by design) in times muchshorter than the mechanical pumpcycle. In other words, this drive circuitrapidly charges the actuator capacitanceand turns itself off until the actuatorsreach their maximum mechanical ex-pansion or contraction. Some time afterthe actuators reach their maximum me-chanical expansion or contraction, thecircuit turns itself back on to charge theactuator capacitance in the opposite po-larity; as it does so, it replenishes the en-ergy lost since the previous charge.

When power is first turned on, the ca-pacitance of the piezoelectric actuators(C2 in Figure 1) is first charged to a poten-tial of +450 V via transistor Q2. The charg-ing current and, hence, charging time aredetermined by the characteristics of the

dc-to-dc converter and the characteristicsof the piezoelectric films. Inductor L1 isused to reduce the initial current spikeduring each recharge cycle, and C1 isused to stabilize the output voltage of theconverter as well as reduce the peak out-put current demand on the dc-to-dc con-verter. Inductor L2 is the main energy-storage inductor; it is used to exchangeenergy with C2 for switching polarity, asdescribed in more detail below.

The timing signals that govern the op-eration of this circuit are transistor/ tran-sistor-logic (TTL)-level pulses with a dura-tion of 5 ms and a repetition frequencyequal to the desired pump frequency (20Hz in the original design). The shortpulse duration is necessary in order to en-able the triac (Q2) to turn itself off whenthe current in inductor L2 reaches zero.The timing signal is generated by tim-ing/control oscillator U6, which supplies

Compact, Efficient Drive Circuit for a Piezoelectric PumpPiezoelectric actuators are charged in a time much shorter than the pump cycle.Lyndon B. Johnson Space Center, Houston, Texas

+28 Vdc

S1 Power

+IN +OUTU1

dc-to-dc Converter

3

4

1

2–IN –OUT C1

36 µF

450 Vdc

8 Vdc

8 Vdc

8 Vdc

8 VdcFB R15C6

U6A

U6B

R10100 Ω R12

200 Ω

C40.1 µF

556556

R14

1 MΩC70.02 µF

C50.1 µF

VCCCONTRGTHRESRST

CONTRGTHRESRST

OUTDIS

913

OUTDIS

GND

143624

115

1210

51

7

25

L1Q1

2 3

6

78

U2

3

R1

1 kΩ

R6100 Ω

Opto Driver

Opto IsolatorR7

100 Ω

Off

On

U3

3Ground

8 Vdc

U5ALM324

U5BLM324

S2 Start Switch

R81 kΩ

R1310 kΩ

R910 kΩ

3

2

1

11 5

6

7

1

3 Run

Start4

65

2

TD2

Replenish Control

Timing Control/Oscillator

LC Tank

11

+

–

+

–

LM7606VIN VOUT

IRFBG30

R21 MΩ

D1

L2 C2PiezoelectricActuator

2 1

R31.5 kΩ

R5

150 Ω

F8

R410 kΩ

R1147 kΩ

C30.0002 ΩF

TD1

TD2

1 2

Opto-Isolator

Q2MAC223

1N4007

D21N4000

MSC22887 Fig 1

5-25-99 bs

Figure 1. This Circuit Operates at High Efficiency in driving a piezoelectric pump at a frequency of 20 Hz. High efficiency is achieved by use of nonresonat-ing inductor L2 to recycle the charge on the piezoelectric-actuator capacitance C2.

timing drive current to opto-isolator U4,which, in turn, provides the switch-on sig-nal for the triac gate.

Figure 2 illustrates the voltage and cur-rent waveforms at the terminals of thepiezoelectric actuators. Suppose that C2has been initially charged. At the begin-ning of the first half cycle, when a timingpulse occurs, the triac is turned on andcurrent begins to flow from C2 throughthe triac and L2. When the magnitude ofthe current increases above the minimumhold current, the triac becomes latchedon. The triac then continues to conductuntil the current drops below the mini-mum hold current, which is essentiallyzero for the purpose of this application.The minimum hold current is reachedand the triac turns off when the voltageon C2 reaches a negative peak near –450

V. The reversal of polarity is provided byL2. The precise magnitude of the peak re-verse voltage is slightly less than |450 V| byan amount that depends on the energylost in the various exchanges of energy in-volved in the reversal of polarity. Whilethe triac remains off, C2 remains in itsnegatively charged state, in which onlyparasitic dielectric losses slowly reduce themagnitude of the voltage on C2.

At the beginning of the second halfcycle, when the next timing pulse arrives,the process described in the precedingparagraph is repeated, except that thecurrents and voltages applied to C2 arereversed. Again, the difference between|450 V| and the magnitude of the peakvoltage depends on the energy lost.

If the operation as described thus farwere allowed to continue, the voltages

would continue to decay and operationwould halt after a number of cycles. Toenable continuous operation, it is neces-sary to replenish the energy lost. This isaccomplished as follows: After the com-pletion of the polarity reversal of the sec-ond half cycle and before the arrival ofthe next timing pulse, the replenish-con-trol subcircuit senses the triac cutoff andturns on Q1 to recharge C2 to 450 V. Atthe chosen 20-Hz pump frequency, thetriac-off time is long enough to enableC2 to be recharged to 450 V by use ofrelatively low charging current, providedthat the energy lost is relatively low.

Because recharging is done only dur-ing the positive half cycle, a dc offset is in-duced in C2; however, this offset is typi-cally a small fraction of the peak drivevoltage and does not adversely affect theoperation of the piezoelectric actuators.The advantage of recharging only duringpositive half cycle is that the circuit canbe less complex and contain fewer com-ponents than would be needed forrecharging during both half cycles.

This work was done by Chris Matice ofStress Engineering Services Inc., and FrankE. Sager and Bill Robertson of OceaneeringSpace Systems for Johnson Space Center.

Title to this invention has been waivedunder the provisions of the National Aeronau-tics and Space Act 42 U.S.C. 2457(f) toOceaneering Space Systems. Inquiries concern-ing licenses for its commercial developmentshould be addressed to:

Oceaneering Space Systems16665 Space Center Blvd.Houston, TX 77058-2268Refer to MSC-22887, volume and number

of this NASA Tech Briefs issue, and thepage number.


25 ms

Time

5A

500 V

–500 V

0

–5A

Pot

entia

l(V

),V

olts

MSC22887 Fig 2

5-18-99 bs

I

VC

urre

nt(I

),A

mpe

res

Figure 2. These Voltage and Current Waveforms occur at the terminals of the piezoelectric actuators.

Dual Common Planes for Time Multiplexing of Dual-ColorQWIPsWith external control, commercial single-color readout integrated circuits could be used.NASA's Jet Propulsion Laboratory, Pasadena, California

A proposed improved method of exter-nally controlled time multiplexing of thereadouts of focal-plane arrays of pairs ofstacked quantum-well infrared photodetec-tors (QWIPs) that operate in differentwavelength bands is based on a dual-detec-tor-common-plane circuit configuration.The method would be implemented in aQWIP integrated-circuit chip hybridizedwith a readout integrated-circuit (ROIC)chip.

There are alternative methods of mul-ticolor readout, but they involve moreon-chip circuitry and the attendant dis-advantages of greater size, power dissipa-tion, number of control signals, andcomplexity of circuitry on the QWIP andROIC chips. To minimize size of, andpower dissipation in, a multicolor pixel,one must minimize the number of tran-sistors and control signals; this can beachieved only by time multiplexing of

colors in each pixel. The time multiplex-ing can be controlled by signals from ex-ternal or internal circuitry. The pro-posed reliance on external control fortime multiplexing would make it possi-ble to use a commercial single-colorROIC with only minor modifications inthe form of extra metallization for an ad-ditional detector common plane and forclocking bias potentials. (If, instead, onewere to rely on internally controlled


multiplexing, it would be necessary toradically redesign the ROIC, at consider-ably greater development cost.) Otheradvantages of the external-control ap-proach over the internal-control ap-proach are greater flexibility and thepossibility of using a technique, knownas “skimming,” for subtracting unwanteddark-, background-, and noise-currentcontributions from readout signals

The figure schematically depicts thecircuitry in one pixel according to an in-ternally controlled multiplexing schemeand according to the proposed exter-nally controlled multiplexing scheme.In both schemes, the time multiplexingwould be accomplished by switching

(clocking or ramping) the biases ap-plied to the QWIPs via detector com-mon planes. In the internal-control case,a bias signal would be applied via a sin-gle detector common plane and two in-ternal electronic switches. In the pro-posed external-control case, bias signalswould be applied via two detector com-mon planes, without internal electronicswitches. A previously unmentioned ad-vantage of the external-control schemeshown in the figure is the need for onlyone indium bump (instead of two) ineach pixel.

This work was done by Sir B. Rafol, SarathGunapala, Sumith Bandara, John Liu, andJason Mumolo of Caltech for NASA's Jet

Propulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).

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

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

Refer to NPO-30523, volume and number ofthis NASA Tech Briefs issue, and the pagenumber.

MMIC Power Amplifier Puts Out 40 mW From 75 to 110 GHzThis amplifier operates over the full frequency band of the WR-10 waveguide.NASA’s Jet Propulsion Laboratory, Pasadena, California

A three-stage monolithic microwaveintegrated circuit (MMIC) W-band am-plifier has been constructed and testedin a continuing effort to develop ampli-fiers as well as oscillators, frequencymultipliers, and mixers capable of oper-ating over wide frequency bands thatextend above 100 GHz. There are nu-merous potential uses for MMICs likethese in scientific instruments, radarsystems, communication systems, andtest equipment operating in this fre-quency range.

This amplifier can be characterized,in part, as a lower-frequency, narrower-band, higher-gain version of the onedescribed in “Power Amplifier With 9 to13 dB of Gain from 65 to 146 GHz”(NPO-20880), NASA Tech Briefs, Vol. 25,No. 1 (January 2001), page 44. This am-plifier includes four InP high-electron-mobility transistors (HEMTs), each hav-ing a gate periphery of 148 µm. In thethird amplifier stage, two of the HEMTsare combined in parallel to maximizethe output power. The amplifier draws

a current of 250 mA at a supply poten-tial of 2.5 V.

In a test, this amplifier was driven by abackward-wave oscillator set to providean input power of 2 mW. The outputpower of the amplifier was measured bya power meter equipped with a WR-10waveguide sensor. As shown in the fig-ure, the amplifier put out a power be-tween 40 and 50 mW over the frequency

range of 75 to 110 GHz, which is the en-tire frequency band of the WR-10 wave-guide. At the stated power levels, thisamplifier offers a power-added efficiencyof slightly more than 6 percent.

This work was done by Lorene Samoska ofCaltech for NASA’s Jet Propulsion Labo-ratory. Further information is contained ina TSP (see page 1). NPO-30577

Frequency, GHz

Out

putP

ower

,mW

750

10

20

30

40

50

60

80 85 90 95 100 105 110

N

The Amplifier Shown in the Photograph was tested at a supply potential of 2.5 V, a gate bias poten-tial of –0.05 V, and an input excitation power of 2 mW at frequencies from 75 to 110 GHz. The plot ofoutput power vs. frequency can be characterized by a large-signal gain of about 13.5±0.5 dB. (Note:The amplifier is actually about 2 mm long.)


2D/3D Visual Tracker forRover Mast

A visual-tracker computer program con-trols an articulated mast on a Mars rover tokeep a designated feature (a target) inview while the rover drives toward the tar-get, avoiding obstacles. Several prior vi-sual-tracker programs have been tested onrover platforms; most require very smalland well-estimated motion between con-secutive image frames — a requirementthat is not realistic for a rover on roughterrain. The present visual-tracker pro-gram is designed to handle large imagemotions that lead to significant changes infeature geometry and photometry be-tween frames. When a point is selected inone of the images acquired from stereo-scopic cameras on the mast, a stereo trian-gulation algorithm computes a three-di-mensional (3D) location for the target. Asthe rover moves, its body-mounted cam-eras feed images to a visual-odometry algo-rithm, which tracks two-dimensional (2D)corner features and computes their oldand new 3D locations. The algorithm re-jects points, the 3D motions of which areinconsistent with a rigid-world constraint,and then computes the apparent changein the rover pose (i.e., translation and ro-tation). The mast pan and tilt anglesneeded to keep the target centered in thefield-of-view of the cameras (thereby mini-mizing the area over which the 2D-track-ing algorithm must operate) are com-puted from the estimated change in therover pose, the 3D position of the targetfeature, and a model of kinematics of the

mast. If the motion between the consecu-tive frames is still large (i.e., 3D trackingwas unsuccessful), an adaptive view-basedmatching technique is applied to the newimage. This technique uses correlation-based template matching, in which a fea-ture template is scaled by the ratio be-tween the depth in the original templateand the depth of pixels in the new image.This is repeated over the entire search win-dow and the best correlation results indi-cate the appropriate match. The programcould be a core for building applicationprograms for systems that require coordi-nation of vision and robotic motion.

This program was written by Max Ba-jracharya, Richard W. Madison, and Issa A.Nesnas of Caltech for NASA’s Jet PropulsionLaboratory, and Esfandiar Bandari, ClaytonKunz, Matt Deans, and Maria Bualat of AmesResearch Center. Further information is con-tained in a TSP (see page 1).

This software is available for commerciallicensing. Please contact Karina Edmonds ofthe California Institute of Technology at(626) 395-2322. Refer to NPO-40696.

Adding Hierarchical Objectsto Relational Database General-Purpose XML-BasedInformation Managements

NETMARK is a flexible, high-through-put software system for managing, storing,and rapid searching of unstructured andsemi-structured documents. NETMARKtransforms such documents from theiroriginal highly complex, constantly chang-ing, heterogeneous data formats into well-

structured, common data formats in usingHypertext Markup Language (HTML)and/or Extensible Markup Language(XML). The software implements an ob-ject-relational database system that com-bines the best practices of the relationalmodel utilizing Structured Query Lan-guage (SQL) with those of the object-ori-ented, semantic database model for creat-ing complex data. In particular,NETMARK takes advantage of the Oracle8i object-relational database model usingphysical-address data types for very effi-cient keyword searches of records acrossboth context and content. NETMARK alsosupports multiple international standardssuch as WEBDAV for drag-and-drop filemanagement and SOAP for integrated in-formation management using Web services. The document-organization and -searching capabilities afforded byNETMARK are likely to make this softwareattractive for use in disciplines as diverse asscience, auditing, and law enforcement.

This program was written by Shu-ChunLin and Chris Knight of Ames ResearchCenter; Tracy La of Computer Science Corpo-ration; David Maluf and David Bell of Uni-versities Space Research; Khai Peter Tran ofQSS Group, Inc.; and Yuri Gawdiak of NASAHeadquarters. Further information is con-tained in a TSP (see page 1).

This invention has been patented by NASA(U.S. Patent No. 6,968,338). Inquiries con-cerning rights for the commercial use of thisinvention should be addressed to the AmesTechnology Partnerships Division at (650)604-2954. Refer to ARC-14662-1.

Software

A process for fabricating thermoelec-tric modules with vacuum gaps separat-ing the thermoelectric legs has been con-ceived, and the feasibility of someessential parts of the process has beendemonstrated. The vacuum gaps areneeded to electrically insulate the legsfrom each other. The process involves theuse of scaffolding in the form of sheets ofa polymer to temporarily separate thelegs by the desired distance, which is typ-ically about 0.5 mm. During a bondingsubprocess that would take place in a par-tial vacuum at an elevated temperature,the polymer would be vaporized, therebycreating the vacuum gaps. If desired, thegaps could later be filled with an aerogelfor thermal insulation and to suppresssublimation of thermoelectric material,as described in “Aerogels for Thermal In-sulation of Thermoelectric Devices”(NPO-40630), NASA Tech Briefs, Vol. 30,No. 7 (July, 2006), page 50.

A simple thermoelectric modulewould typically include thermoelectriclegs stacked perpendicularly betweenmetal contact pads on two ceramic sub-strates (see figure). As the design of thethermoelectric module and the fabrica-tion process are now envisioned, themetal contact pads on the ceramic sub-strates would be coated with a suitablebonding metal (most likely, titanium),and the thermoelectric legs would beterminated in a possibly different bond-ing metal (most likely, molybdenum).Prior to stacking of the thermoelectricpads between the metal pads on the ce-ramic substrates, the polymer sheetswould be bonded to the appropriatesides of the thermoelectric legs. Afterstacking, the resulting sandwich struc-ture would be subjected to uniaxial pres-sure during heating in a partial vacuumto a temperature greater than 700 °C.The heating would bond the metal padson the legs to the metal pads on the sub-strates and would vaporize the polymersheets. The uniaxial pressure wouldhold the legs in place until bonding andvaporization were complete.

Ideally, the polymer chosen for use inthis process should be sufficiently rigid

to enforce dimensional stability of thegaps and should vaporize at a tempera-ture low enough that it does not un-dergo pyrolysis. (Pyrolysis would createan undesired electrically conductive car-bonaceous residue.) Poly(α-methyl-styrene) [PAMS] has been selected as apromising candidate. PAMS is consid-ered to be rigid, and, in a partial vacuumof 10–6 torr (≈1.3 × 10–4 Pa), it vaporizesin the temperature range of 250 to 400 °C,without pyrolizing.

Initially, the primary concern raisedby vaporization of polymer scaffoldingwas that polymer-vapor residue might in-terfere with the bonding of the thermo-electric legs to the metal pads on the ce-ramic substrates. In an experiment toinvestigate the likelihood of such inter-ference, a mockup comprising twomolybdenum legs separated by a PAMSsheet in contact with a titanium platewas placed under uniaxial pressure andheated to a temperature of 950 °C in apartial vacuum of 10–6 torr. Strong, uni-

form bonds were made between themolybdenum legs and the titaniumplate, demonstrating that the PAMSvapor did not interfere with bonding.

This work was done by Jeffrey Sakamoto,Shiao-pin Yen, Jean-Pierre Fleurial, and Jong-Ah Paik of Caltech for NASA’s Jet Propul-sion Laboratory. Further information iscontained in a TSP (see page 1).



Refer to NPO-41248 volume and number ofthis NASA Tech Briefs issue, and the pagenumber.


Materials

Uniaxial Pressure and Heating in Vacuum

Thermoelectric Legs

Polymer Sheets

Gaps Left by Vaporization of Polymer Sheets

Ceramic Substrate

Metal Bands Between Thermoelectric Legs and Contact Pads on Ceramic Substrates

BEFORE

AFTER

Ceramic Substrate

NPO41248 ABPI

11-22-04 bs

A Single Thermoelectric Module containing a small number of legs is shown here for simplicity to fa-cilitate understanding of the process described in the text. In an economically practical version of theprocess, multiple modules would be stacked in a more-complex, checkerboardlike arrangement dur-ing bonding and vaporization of the polymer sheets.

Vaporizable Scaffolds for Fabricating Thermoelectric ModulesThermoelectric legs would be separated by precise gaps.NASA’s Jet Propulsion Laboratory, Pasadena, California


Producing Quantum Dots by Spray Pyrolysis Sizes of quantum dots are determined by sizes of sprayed drops. John H. Glenn Research Center, Cleveland, Ohio

An improved process for makingnanocrystallites, commonly denotedquantum dots (QDs), is based on spraypyrolysis. Unlike the process used hereto-fore, the improved process is amenableto mass production of either passivatedor non-passivated QDs, with computercontrol to ensure near uniformity of size.

The extraordinary optical propertiesthat make QDs useful can be tailored viathe chemical compositions and sizes ofthe QDs. Heretofore, QDs have beenmade by in situ pyrolysis of molecularreagents that contain all the elemental in-gredients. In order to make all QDs in abatch have the same or nearly the samesize, it is necessary to prevent nucleationwith adjacent QDs by passivating the sur-faces of the QDs soon after they haveformed. Passivation is effected by use of apassivating coordinating solvent/ligandcompound as one of the ingredients. Thiscompound both serves as a medium forthe formation of the QDs via pyrolysisand readily coordinates to the surfacesof the QDs, thereby preventing furthernucleation. Covering of the surfaces ofthe QDs with the coordinating molecu-lar groups is known in the art as cap-ping. The capping groups can be organic or inorganic. The sol-vent/ligand can be formulated to tailorthe surface properties of the QDs for aspecific application. For example, theycan be made hydrophilic or hydropho-bic. Unfortunately, this in situ pyrolysisprocess cannot readily be scaled up tomass production, and the QDs pro-duced exhibit excessive nonuniformity.

The improved process also includes py-

rolysis, but differs from the prior processin the pyrolysis conditions and in themanner in which the ingredients are pre-pared for pyrolysis. In addition, one hasan option to form QDs of controlled sizewithout or with capping. Unlike in theprior process, one does not rely on for-mulation and pyrolysis process conditionsto cause the QDs to form spontaneouslyin the desired size range. Instead, oneforces the QDs to form in a desired rela-tively narrow size range by spraying a solu-tion of QD-precursor material through anultrasonic nozzle to form drops of con-trolled size. The average size of the dropscan be tailored via the surface tensionand density of the solution. Even moreconveniently, the average size can be tai-lored via the frequency of ultrasonic agi-tation. Experiments have shown that for agiven formulation, the average size is in-versely proportional to frequency to the2/3 power.

The basic concept of the improvedprocess admits of variations. For exam-ple, in a simple embodiment of theprocess for making non-capped QDs, aprecursor solution is sprayed into a cold-wall reactor that contains a substrate thatis electrically heated to the desired reac-tion temperature (see figure). The dropsfalling on the substrate become pyrolizedinto QDs. In a slightly more complex ex-ample, capped QDs are made by similarspraying of a precursor solution into ahot passivating solvent.

This work was done by Kulbinder Banger,Michael H. Jin, and Aloysius Hepp of GlennResearch Center. Further information iscontained in a TSP (see page 1).

Inquiries concerning rights for the commer-cial use of this invention should be addressedto NASA Glenn Research Center, CommercialTechnology Office, Attn: Steve Fedor, Mail Stop4–8, 21000 Brookpark Road, Cleveland, Ohio44135. Refer to LEW-17444.

LEW-17444ABPI

11-18-03 es

Argon Supply

Syringe Pump Containing

Precursor SolutionUltrasonic

Nozzle

Outlet to Cold Traps

APPARATUS FOR MAKING UNCAPPED QUANTUM DOTS

Argon Supply

Syringe Pump Containing

Precursor SolutionUltrasonic

Nozzle

APPARATUS FOR MAKING CAPPED QUANTUM DOTS

To Bubbler and Traps

Hot Passivating Solution

These Two Apparatuses serve to illustrate the basic principles of two variants of the spray-pyrolysisprocess for making uncapped or capped QDs of controlled size.


Machinery/Automation

The Planetary Autonomous Am-phibious Robotic Vehicle (PAARV), nowat the prototype stage of development,was originally intended for use in acquir-ing and analyzing samples of solid, liq-uid, and gaseous materials in cold envi-ronments on the shores and surfaces,and at shallow depths below the sur-faces, of lakes and oceans on remoteplanets. The PAARV also could beadapted for use on Earth in similar ex-ploration of cold environments in andnear Arctic and Antarctic oceans andglacial and sub-glacial lakes.

The PAARV design is based partly onthe design of prior ice-penetrating ex-ploratory robots and partly on the de-signs of drop sondes heretofore used onEarth for scientific and military purposes.Like a sonde, the PAARV is designed tobe connected to a carrier vehicle (e.g., aballoon, aircraft, or vessel in a terrestrialsetting) by a tether, through which thecarrier vehicle would provide powerand would relay data communicationsbetween the PAARV and an external orremote control station. Like a sonde, thePAARV could be lowered from the carriervehicle into an ocean or lake environ-ment to be explored. Unlike a sonde, thePAARV would be capable of swimmingand of crawling along the bottom, andcrawling out of the ocean or lake andmoving to a designated site of scientificinterest on the shore.

As now envisioned, the fully devel-oped PAARV (see figure) would in-clude an upper sonde segment and alower sonde segment. Protruding fromthe lower sonde segment would be twoassemblies containing both lifting sur-faces (for control of attitude during de-scent and swimming) and crawlertracks. These assemblies could be ro-tated to align them parallel, perpendi-cular, or at an oblique angle with re-spect to the longitudinal axis of thelower sonde segment. Also protrudingfrom the lower sonde segment, at apoint above the center of gravity, wouldbe a thruster for swimming.

The upper sonde segment would becylindrical, approximately 30 cm in diam-

eter and 20 cm high. Thissegment would contain atether-management-and-ac-tuation subsystem; buoy-ancy-control chambers; asubsystem of pumps, actua-tors, and valves; a chamberholding compressed gas;and a chamber containing aheater. The upper sondesegment would be attachedto the lower sonde segmentvia a single-joint actuatorthat effects rotation aboutan axis perpendicular tothe longitudinal axes of thesonde segments.

The lower sonde segmentwould be approximately 60cm long and 15 cm in diam-eter. Either a general-pur-pose radioisotope heatsource or a mass of phase-change heat-storage mate-rial would be located in thenose (lower and outer end)of the lower sonde segmentto keep instruments warm inthe cold environment. Thesonde would have a dual-walled shell with insulationto reduce the loss of heat.The heat source in the nosealso would serve as ballast tomaintain stability like that ofa traditional ocean buoy.

When the PAARV was initially low-ered from the carrier vehicle, theupper and lower sonde segments andthe lifting-surface/crawler-track assem-blies would be aligned collinearly sothat the PAARV would float in a nomi-nal vertical orientation like a buoy.Once the buoyancy chambers startedto fill, the tether would be paid outfrom the top sonde segment as thePAARV descended. Upon arrival at adepth designated for swimming, thethruster would be activated and relativealignments of the upper and lowersonde segments and the lifting-sur-face/crawler-track assemblies varied asneeded for steering. Upon contact with

the bottom, the lower sonde segmentand the lifting-surface/crawler-track as-semblies would be turned to a nomi-nally horizontal orientation with theupper sonde segment in a nominallyvertical orientation, and the crawlertracks would then be activated.

A sampling needle could be extendedfrom the lower side of the lower sondesegment into the bottom of the ocean orlake, where it would adsorb bottom ma-terial. The needle would then be re-tracted, then heated to desorb the mate-rial for analysis by instruments in thelower sonde segment. The data from theanalyses would be relayed to the exter-nal control station via the tether.

Mobile Robot for Exploring Cold Liquid/Solid EnvironmentsThis tethered robot could float, swim, crawl, and sample environmental materials. NASA’s Jet Propulsion Laboratory, Pasadena, California

UpperSonde

Segment

LowerSonde

Segment

Nose ofLower Sonde Segment,

ContainingHeat Source and

BallastThruster

Lifting-Surface/Crawler-Track

Assembly

CRAWLING

SWIMMING

Thruster

NPO-40731ABPI

3-28-04 es

The PAARV is shown here in two of its many possible crawlingand swimming configurations.


Once the bottom sampling was com-plete, the PAARV would increase itsbuoyancy by displacing liquid from thebuoyancy-control chambers and wouldreel the tether back in. An onboardguidance, navigation, and control sys-tem coupled with acoustic range sensorswould enable the vehicle to move slowly

toward shore as it ascended. Upon con-tact with ascending slope, the crawlertracks would be rotated to the angle ofthe slope and the crawler tracks wouldbe activated. Once out of the water, thePAARV would crawl to a location of in-terest designated by coordinates pro-vided by cameras on the carrier vehicle

or an aircraft overhead. The samplingprocess would be repeated at the loca-tion of interest.

This work was done by Charles Bergh andWayne Zimmerman of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1)NPO-40731

System Would Acquire Core and Powder Samples of RocksA sampling system would be built around an ultrasonic/sonic drill corer.NASA’s Jet Propulsion Laboratory, Pasadena, California

A system for automated sampling ofrocks, ice, and similar hard materials atand immediately below the surface of theground is undergoing development. Thesystem, denoted a sample preparation, ac-quisition, handling, and delivery(SPAHD) device, would be mounted on arobotic exploratory vehicle that would tra-verse the terrain of interest on the Earthor on a remote planet. The SPAHD devicewould probe the ground to obtain datafor optimization of sampling, prepare thesurface, acquire samples in the form(s) ofcores and/or powdered cuttings, and de-liver the samples to a selected location foranalysis and/or storage.

The SPAHD device would be builtaround an ultrasonic/sonic drill corer(USDC) — an apparatus that was re-ported in “Ultrasonic/Sonic Drill/ Cor-ers With Integrated Sensors” (NPO-20856), NASA Tech Briefs, Vol. 25, No. 1(January 2001), page 38. To recapitu-late: A USDC includes a hollow drill bitor corer, in which combinations of ultra-sonic and sonic vibrations are excited byan electronically driven piezoelectric ac-tuator. The corer can be instrumentedwith a variety of sensors (and/or thedrill bit or corer can be used as anacoustic-impedance sensor) for bothprobing the drilled material and acquir-ing feedback for control of the excita-tion. The USDC advances into the mate-rial of interest by means of ahammering action and a resulting chis-eling action at the tip of the corer. Thehammering and chiseling actions are soeffective that unlike in conventionaltwist drilling, a negligible amount ofaxial force is needed to make the USDCadvance into the material. Also unlike aconventional twist drill, the USDC oper-ates without need for torsional restraint,lubricant, or a sharp bit.

In addition to a USDC, the SPAHDdevice (see Figure 1) would includesensor, control, and communicationsubsystems; a subsystem for positioningthe USDC at the desired position andorientation on the ground; a set of in-terchangeable USDC bits; a tool rack tostore the bits; and mechanisms for ma-nipulating and delivering samples. The

bits would be attached to, and detachedfrom, a resonator horn of the piezoelec-tric actuator by means of simple-to-op-erate snap-on/snap-off mechanisms.The set of bits would include a probingbit, bits for cutting cores and collectingpowdered cuttings, bits for extractingthe cores after they have been cut (seeFigure 2), and an ultrasonic rock abra-sion tool (URAT) bit [described in “Ul-trasoni-cally Actuated Tools for Abrad-ing Rock Surfaces” (NPO-30403), NASATech Briefs, Vol. 30, No. 7 (July, 2006),page 58.

This work was done by Yoseph Bar-Cohen,James Randolph, Xiaoqi Bao, Stewart Sher-rit, Chuck Ritz, and Greg Cook of Caltech forNASA’s Jet Propulsion Laboratory. Fur-ther information is contained in a TSP (seepage 1).NPO-30640

Platform(Robotic Exploratory Vehicle)

Electrical and MechanicalInterfaces With the Subsystems

Positioning Subsystem

USDC

Tooling for Interchange of Bits

Tool Bit

Sample-Handling-and-DeliverySubsystem

CommunicationSubsystem

ControlSubsystem

PowerSubsystem

SensorSubsystem

NPO30640 Fig 1

8-13-02 bs

Figure 1. A SPAHD Device would be a highly in-tegrated system containing specially designedmechanisms and electronic circuits working to-gether to perform multiple functions, includingprobing, preparation of surfaces, acquisition ofcore and powder samples, and manipulationand delivery of the samples. The relative posi-tions in this block diagram indicate approximatemechanical and electrical relationships amongsubsystems and components.

Wedge

Tube Spring

NPO30640 Fig 2

8-19-02 bs

Figure 2. An Extraction Bit is one example of spe-cial-purpose bits that would be included in aSPAHD device. This bit would be inserted arounda recently cut core. The wedge in the bit wouldintroduce a transverse force that would causethe core to break off somewhere near its root.The springs in the bit would then retain the coreso that the core could be lifted out of the hole.


Improved Fabrication of Lithium Films Having Micron FeaturesDry chemicals and a dry process are used to prevent undesired reactions.NASA’s Jet Propulsion Laboratory, Pasadena, California

An improved method has been de-vised for fabricating micron-dimensionLi features. This approach is intendedfor application in the fabrication oflithium-based microelectrochemical de-vices — particularly solid-state thin-filmlithium microbatteries.

The need for this special process arisesbecause lithium engages in undesiredchemical reactions with water and withmost of the chemicals commonly used inthe microfabrication of silicon and othermaterials. The method described below in-volves the use of only water-free “dry”chemicals that are compatible withlithium. The figure illustrates the pertinentsteps in the fabrication of a microelectro-chemical cell containing a lithium anode.

The following steps are performed insequence:1. The solid-electrolyte layer is covered

with AZ 1518 (or equivalent) photore-sist, which is then exposed to ultravio-let light in a pattern that defines theareas from which the solid electrolyte isto be removed.

2. The photoresist developer dissolvesboth the photoresist and the underly-ing solid electrolyte from the afore-mentioned areas, leaving developedphotoresist and underlying solid elec-trolyte only in a defined area.

3. The developed photoresist is removedby use of dry acetone, leaving thesolid electrolyte in the defined area.

4. A layer of parylene 2 to 4 µm thick isdeposited over the entire workpiece.

5. NR 5-8000 (or equivalent) photoresist isapplied, exposed, and developed. Theopening between areas covered by devel-oped photoresist defines the anode areaon which lithium is to be deposited.

6. The exposed parylene is etched awayin oxygen plasma.

7. A thin film of lithium is evaporativelydeposited on the workpiece.

8. Adhesive tape (Kapton, or equivalent)is pressed onto the top of the work-piece, then pulled off. The lithium de-posited on the photoresist-coveredareas adheres more weakly than doesthe lithium deposited on the anode

area, to such a degree that pulling offthe tape removes all lithium from cov-ering the photoresist while leaving allthe lithium in the desired anode area.

9. The photoresist and the remainingparylene are removed by use of dryacetone and subsequent etching inoxygen plasma.This work was done by Jay Whitacre and

William West of Caltech for NASA’s Jet Propul-sion Laboratory. Further information is con-tained in a TSP (see page 1).



Refer to NPO-30725, volume and number ofthis NASA Tech Briefs issue, and the pagenumber.

Deposition andExposure ofPhotoresist

Development ofPhotoresist andDissolution of

Solid Electrolyte

Removal ofPhotoresist

Deposition ofParylene

ParyleneParyleneParylene

Deposition, Exposure, andDevelopment of Photoresist

Etching of ExposedParylene inO2 Plasma

Photoresist

Deposition of Li

Li

Lift-off of Excess Liand Removal ofPhotoresist and

Parylene

Si

Anode Current Collector

CathodeCurrent Collector

Solid Electrolyte(LiPON)

Cathode Layer

SixNy

Photoresist

NPO3072510-30-02 bs

Manufacturing & Prototyping

A Lithium Anode in a Microelectrochemical Cell is fabricated in a dry process.


Manufacture of Regularly Shaped Sol-Gel PelletsFor mass production, an extrusion process is superior to a spray process. John H. Glenn Research Center, Cleveland, Ohio

An extrusion batch process for manufac-turing regularly shaped sol-gel pellets hasbeen devised as an improved alternative toa spray process that yields irregularlyshaped pellets. The aspect ratio of regularlyshaped pellets can be controlled more eas-ily, while regularly shaped pellets packmore efficiently. In the extrusion process, awet gel is pushed out of a mold andchopped repetitively into short, cylindricalpieces as it emerges from the mold. Thepieces are collected and can be either (1)dried at ambient pressure to xerogel, (2)solvent exchanged and dried under ambi-ent pressure to ambigels, or (3) supercriti-cally dried to aerogel. Advantageously, theextruded pellets can be dropped directly ina cross-linking bath, where they develop aconformal polymer coating around theskeletal framework of the wet gel via reac-tion with the cross linker. These pellets canbe dried to mechanically robust X-Aerogel.

The extrusion process begins with thepreparation of a suitable sol. This is ac-complished online by mixing tow solu-tions as they flow from two separate con-tainers towards the mold. A typical solthat yields a gel that can be cross-linkedeasily is made by mixing two cold solu-

tions: One solution consists of 4.5 vol-ume parts of acetonitrile, 0.482 volumeparts of aminopropyltriethoxysilane,and 1.45 volume parts of tetra-methoxysilane. The other solution con-sists of 4.5 volume parts of acetonitrileand 1.5 volume parts of water.

The cold sol is equipped with a plunger.As soon as the sol gels in the mold, thenewly formed gel is pushed out of themold and is chopped to the desiredlength. When the gel mold is exhausted,the plunger is retracted and the processrepeated. An important consideration inthe process is that the sol needs to haveformed a gel, which is stable enough toslice before the extrusion process isbegun. The process can be used withnearly any sol-gel formulations, but highthroughout requires a fast gelling sol suchas the one obtained by using aminopropy-ltriethoxysilane (APTES)/tetramethoxysi-lane (TMOS) mixtures. In this formula-tion, control of the gelation rate requiresthat the precursor solutions must be re-frigerated prior to filling the mold. In anextension of the basic process describedhere, multiple molds with plungers canbe connected to a liquid manifold con-

nected to a sol-preparation apparatus andmounted together to facilitate repeatedfilling with sol and automatic extrusion.

The most benefit from this process isrealized when it is combined with chemi-cal cross linking as pellets emerge fromthe mold. A suitable cross linker is a diiso-cynanate (e.g., Desmodour N3200 fromBayer Corporations, or equivalent),which, owing to the small size of the pel-lets, diffuses and reacts quickly with theinternal surfaces of the porous nanoparti-cle framework of the wet gels. The wet-pellets are subsequently dried from super-critical carbon dioxide, yieldingmechanically robust colorless, slightlytranslucent pellets of X-Aerogel. Othercross-linker systems such as epoxies andpolyolefins have been also investigated.

This work was done by Nicholas Leventis,James C. Johnston, and James D. Kinder ofGlenn Research Center. Further informa-tion is contained in a TSP (see page 1)

Inquiries concerning rights for the commer-cial use of this invention should be addressedto NASA Glenn Research Center, InnovativePartnerships Office, Attn: Steve Fedor, MailStop 4–8, 21000 Brookpark Road, Cleveland,Ohio 44135. Refer to LEW-17787-1.


Regulating Glucose and pH, and Monitoring Oxygen in a BioreactorGlucose and oxygen concentrations are monitored, and glucose concentration and pH areadjusted as needed.Lyndon B. Johnson Space Center, Houston, Texas

Figure 1 is a simplified schematic dia-gram of a system that automatically reg-ulates the concentration of glucose orpH in a liquid culture medium that iscirculated through a rotating-wall per-fused bioreactor. Another system, shownin Figure 2, monitors the concentrationof oxygen in the culture medium.

The glucose-regulating system in-cludes an electrochemical sensor thatmeasures the concentration of glucosein the medium flowing out of the biore-actor, a reservoir containing a concen-trated glucose stock solution, and a peri-staltic pump. The sensor reading isdigitized and monitored by a computer,which generates commands to open andclose valves, as described below, to adjustthe concentration of glucose. This al-lows the system to maintain an optimalconcentration of glucose without usinglarge volumes of cell culture medium.The glucose control system operates inone of the following three modes (seeFigure 1): perfusion, where the mediumis simply circulated without adding or re-moving any liquid; infusion, where fresh

cell medium is introduced into circula-tion, replacing the spent medium in thebioreactor; injection, where a smallamount of concentrated glucose solu-tion is added to the bioreactor when theglucose concentration falls below 75mg/dL.

The electrochemical glucose sensorincludes a membrane that is coveredwith a thin (10 to 200 mm thick) layer ofimmobilized glucose oxidase enzymeand is coupled to a three-electrode am-perometric probe and a flow cell. Glu-cose diffuses through the membraneand, in the presence of the glucose oxi-dase, is converted to hydrogen peroxideand gluconic acid. The rate of genera-tion of hydrogen peroxide is measuredamperometrically on a platinum work-ing electrode at a potential of 0.7 V withrespect to an Ag/AgCl reference elec-trode. The sensor reading is correlatedwith the concentration of glucose by theMichaelis-Menten equation,

I=ImaxS/(K + S),where I is the sensor current, Imax is themaximum sensor current for the amper-

ometric reaction, K is a constant, and Sis the concentration of glucose. The con-stants Imax and K are determined bymeans of a two-point calibration, using acommercial glucose analyzer to deter-mine the glucose concentrations. Thissensor system has been tested in cell cul-ture experiments for over 50 days. Thecontrol of glucose in this manner re-duces the amount of culture medium re-quired to optimally grow cells.

The pH sensor, described in “Opto-electronic Instrument Monitors pH in aCulture Medium” (MSC-23107), NASATech Briefs, Vol. 28, No. 9 (September2004), page 4a, was coupled to the abovecontrol system containing a small volumeof a mixture of sodium and potassium hy-droxide and bicarbonate solution to forma pH control system. When the sensor pHoutput fell below a preset level, a smallamount of buffer was injected to thebioreactor to bring the pH to the desiredlevel. This system was tested in a 120-daycell run, where the pH was controlled at7.1±0.1 pH unit. The volume of bufferused was negligible compared to the vol-ume of cell culture medium required for thesame level of pH control. By controllingthe culture pH in this manner, the cellsproduced less biofilm.

The oxygen-monitoring system is de-signed to satisfy special requirements fornoninvasiveness, sterilizability, compact-ness, light weight, and nontoxicity to thecells nourished by the solution. The oxy-gen sensor exploits dynamic quenchingof fluorescence by oxygen molecules.The sensor includes a capillary tube of in-side diameter 2 mm, lined with a 0.1-mm-thick sensing layer of oxygen-sensitive flu-orescent dye, and titanium oxideencapsulated in a gas-permeable, ion-im-permeable silicone rubber. The sensinglayer is overlaid with a black shieldinglayer of carbon black encapsulated in sili-cone rubber. The solution of interestflows through the tube, and oxygen fromthe solution permeates the silicone-rub-ber layers. The degree of permeation

Bio-Medical

NO

Valve 1

NC

NO

NO – Normally Opened NC – Normally Closed

Valve 2

NC

NO

Valve 3

NC

Peristaltic Pump

Oxygenator

Oxygenator Bioreactor

Sensor

Computer

Used Nutrient

Out

Glucose Solution

In

Fresh Nutrient Medium

In

MSC-23473 fig1

2-13-03 es

Figure 1. Valves Are Momentarily Switched between open and closed positions, to inject concentratedglucose solution and drain used nutrient medium, when the measured concentration of glucose in thecirculating nutrient medium falls to a preset minimum level.


(and, hence, the rate of quenching of flu-orescence) is reversible; that is, it variesalong with the concentration of dissolvedoxygen.

The dye is tris(4,7-diphenyl-1,10-phenanthroline)ruthenium(ll)chloride[Ru(dpp)3Cl2], which exhibits a highquantum yield (30 percent) of red(wavelength 626 nm) fluorescence witha lifetime of 5.34 ms after irradiationwith blue light. A pulsed light-emittingdiode (LED) of wavelength 465 nm isused as the radiation source. Operationin pulse mode minimizes the dye-bleach-

ing effect that could occur if the LEDwere to irradiate the dye continuouslyduring long-term monitoring. A long-wavelength-pass filter is used to removethe blue LED light reflected and scat-tered by the irradiated capillary tubeand its contents. The remaining light —predominantly the red fluorescence —is detected by a photodiode.

The pulsed output of the photodiodeis digitized and processed to determinethe concentration of dissolved oxygen interms of the partial pressure of oxygen(pO2). This determination is made by

use of the Stern-Volmer equation for theintensity and fluorescence lifetime ofoxygen-quenched fluorescence of a lu-minescent dye:

I0/I= τ0/τ = 1 + Ksv(pO2),where I0 and I are the intensities of fluo-rescence in the absence and presence ofoxygen, respectively; τ0 and τ are the fluo-rescence-decay lifetimes in the absenceand presence of oxygen, respectively; andKsv (the Stern-Volmer constant) dependson details of the sensor design and con-struction. The sensor is calibrated by useof a phosphate-buffered saline solutioncontaining a known elevated concentra-tion of oxygen. The oxygen sensor wastested for over 180 days in a perfused ro-tating-wall bioreactor, using a single, ini-tial calibration for the entire experiment.

This work was done by Melody M. An-derson and Neat R. Pellis of JohnsonSpace Center, and Antony S. Jeevarajan,Thomas D. Taylor, Yuanhang Xu, FrankGao of Wyle Laboratories. For further infor-mation, contact the Johnson InnovativeTechnology Partnerships 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-23473/504/13/54.

LED

Irradiation at Wavelength of 475 nm

Black Shielding Layer

Sensing Layer Containing Fluorescent Dye

Glass Capillary Tube

Long-Pass Filter, Cutoff Wavelength 575 nm

Fluorescence at Wavelength of 626 nm

Photodiode

MSC23473 fig 2

Figure 2. Blue Light From the LED excites red fluorescence in the dye encapsulated in the silicone rub-ber. The intensity of the fluorescence, measured by use of the photodiode, varies with the concentra-tion of oxygen dissolved in the solution flowing through the tube.


Physical Sciences

Satellite Multiangle Spectropolarimetric Imaging of AerosolsOne instrument would implement a synergistic combination of multispectral, multiangle, andpolarimetric techniques.NASA’s Jet Propulsion Laboratory, Pasadena, California

A proposed remote-sensing instru-ment, to be carried aboard a spacecraftin orbit around the Earth, would gatherdata on the spatial distribution and ra-diative characteristics of troposphericaerosols. These data are needed for bet-ter understanding of the natural andanthropogenic origins of aerosols, andof the effects of aerosols on climate andatmospheric chemistry.

The instrument would implement asynergistic combination of multispec-tral, multiangle, and polarimetric meas-urement techniques to increase the ac-curacies of aerosol-optical-depth andaerosol-particle-property characteriza-tions beyond what is achievable by use ofeach technique by itself. Additional ben-efits expected to be realized by the spe-cific novel combination of differentmeasurement techniques in one instru-ment include the following:• The instrument could make simul-

taneous measurements (describedbelow) that are essential for deter-mining the mesoscale variability ofaerosols;

• The instrument would have a wide-swath, high-resolution imaging capa-bility for discerning clouds and for fre-quent global sampling; and

• The cost of building this instrumentwould be less than the cost of buildingseparate instruments for the variousmeasurements.Features of the design and perform-

ance of the instrument would includespectral coverage from near ultravioletto short-wave infrared, global spatial cov-erage within a few days, simultaneous in-tensity and polarimetric imaging at mul-tiple view angles, kilometer tosub-kilometer spatial resolution, andmeasurement of the degree of linear po-larization in one visible and one short-wave infrared spectral band.

The instrument (see block diagram)would acquire data in push-broomfashion in nine cross-ground-trackswaths, each aimed at a differentalong-ground-track look angle. A sepa-rate camera would be used for each

look angle, but all nine cameras couldbe of the same design. Reflective tele-scope and camera optics would beused because they offer significant ad-vantages over refractive optics, includ-ing high transmittance over the broadwavelength range of interest, absenceof chromatic aberration, the possibilityof achieving the desired effects by useof fewer optical elements, less suscepti-bility to formation of ghost images,and better and more stable polariza-tion performance. The instrumentwould measure intensity at wave-lengths of 380, 412, 446, 558, 866,1,375, and 2,130 nm. Polarization plusintensity would be measured at 670and 1,630 nm.

The design of the instrument wouldbe dominated by the requirement forpolarimetric accuracy. Of several polari-metric approaches that were consid-ered, the one selected to satisfy this re-quirement involves adaptation ofadvanced techniques of high-precisionimaging polarimetry used in ground-based solar astronomy. The approach

involves the use of (1) a photoelasticmodulator to rapidly rotate the plane ofpolarization, (2) one or two analyzersand corresponding linear arrays of pho-todetectors, and (3) demodulation ofthe photodetector outputs to obtaindata from which one can compute theStokes parameters Q and U (pertainingto the predominance of horizontal oververtical linear polarization and the pre-dominance of 45° over 135° polariza-tion, respectively). In order to imple-ment the specific detection scheme, itwould be necessary to design and buildflight-qualified modulators, as well as acustom complementary metaloxide/semiconductor integrated circuitfor the low-noise, rapid-readout pho-todetector array.

This work was done by David Diner,Steven Macenka, Lawrence Scherr, andSuresh Seshadri of Caltech; Russell Chipmanof the University of Arizona; and ChristophKeller of the National Solar Observatory forNASA’s Jet Propulsion Laboratory. Fur-ther information is contained in a TSP (seepage 1). NPO-40936

CalibrationCircuits

ModulatorCircuitry

Instrument-ControlCircuits

Control andData-

HandlingCircuits

PowerCircuits

Telescope RetardanceModulator

Focal-PlaneAssembly

CameraElectronics

ThermalControl

9 Along-Track Look Angles

NPO-40936ABPI

6-30-04 es

This Simplified Block Diagram shows the functional relationships among basic components of the pro-posed instrument.


Microscale Regenerative Heat Exchanger Materials and dimensions are chosen to optimize performance at microscale. John H. Glenn Research Center, Cleveland, Ohio

The device illustrated in Figure 1 isdesigned primarily for use as a regenera-tive heat exchanger in a miniature Stir-ling engine or Stirling-cycle heat pump.A regenerative heat exchanger (some-times called, simply, a “regenerator” in

the Stirling-engine art) is basically athermal capacitor: Its role in the Stirlingcycle is to alternately accept heat from,then deliver heat to, an oscillating flowof a working fluid between compressionand expansion volumes, without intro-

ducing an excessive pressure drop.These volumes are at different tempera-tures, and conduction of heat betweenthese volumes is undesirable because itreduces the energy-conversion efficiencyof the Stirling cycle. Hence, among the

Interferometric System for Measuring Thickness of Sea Ice Frequency- and spatial-domain VHF radar interferometry are used.NASA’s Jet Propulsion Laboratory, Pasadena, California

The cryospheric advanced sensor(CAS) is a developmental airborne(and, potentially, spaceborne) radar-based instrumentation system for meas-uring and mapping the thickness of seaice. A planned future version of the sys-tem would also provide data on thethickness of snow covering sea ice. Fre-quent measurements of the thickness ofpolar ocean sea ice and its snow coveron a synoptic scale are critical to under-standing global climate change andocean circulation.

The CAS system includes two rela-tively-narrow-band chirped radar subsys-tems that operate about two differentnominal middle frequencies in the very-high-frequency (VHF) range (e.g., 137and 162 MHz). The radar targets thesame surface area from two slightly dif-ferent directions (see figure). The radarbackscatter signals are processed to ex-tract angular- and frequency-domain cor-relation functions (ACF/FCF) so thatthe system acts, in effect, as a combinedspatial- and frequency-domain interfero-metric radar system.

The phase of the ACF/FCF varieswith the thickness of the sea ice. To en-able the utilization of the phase infor-mation to compute this thickness, theinteractions between the radar wavesand the seawater, ice, and snow coverare represented by a mathematicalmodel: The snow, sea ice (includingair bubbles and brine inclusions), andseawater are represented as layers,each characterized by an assumedthickness and a known or assumedcomplex-number index of refraction.Each interface (air/snow, snow/ice,and ice/water) is modeled as deviatingfrom a plane by a surface roughness

characterized by a Gaussian spectrum.The scattering of the radar waves fromthe interfaces is computed by use ofsmall-perturbation and Kirchhoffrough-surface submodels. The scatter-ing from within the layers is computedby a Rayleigh volume scattering model.The ACF/FCF is computed from thescattered signals.

Assuming that the ACF/FCF obeys themodel, the interferometric phase infor-mation can be inverted by use of a suit-able computational inversion technique(e.g., a genetic algorithm or gradient de-scent or other least-squares technique)to obtain the thickness of the sea ice. Inessence, the inversion amounts to seek-

ing whichever value of sea-ice thicknessused in the model yields the best matchbetween (1) the ACF/FCF interferomet-ric phase computed from the model and(2) the ACF/FCF measured interfero-metric phase.

This work was done by Ziad Hussein,Rolando Jordan, Kyle McDonald, Ben-jamin Holt, and John Huang of Caltech;Yasuo Kugo, Akira Ishimaru, and SermsakJaruwatanadilok of the University of Wash-ington, Seattle; and Torry Akins andPrasad Gogineni of the University ofKansas, Lawrence, for NASA’s JetPropulsion Laboratory. Further infor-mation is contained in a TSP (see page 1).NPO-42391

Two Chirped VHF Radar Subsystems operating about different nominal middle frequencies targetsthe same surface area from two slightly different angles. Frequency- and spatial-domain interferomet-ric computations performed on the radar backscatter signals enable determination of the thickness ofthe sea ice.

θ1

θ2

Width of Illuminated Swath

Width ofResolutionCell

Snow

Sea Ice

Seawater

Bubbles and

Brine Inclusions

Radar Subsystem 2: 162 MHz

Radar Subsystem 1: 137 MHz

NPO-42391ABPI

9-21-05 es

desired characteristics of a regenerativeheat exchanger are low pressure dropand low thermal conductivity along theflow axis.

As shown in the enlarged views in Fig-ure 1, the device features a multilayergrating structure in which each layer isoffset from the adjacent layer by half acell opening along both axes perpendi-cular to the flow axis. In addition, eachgrating layer is a composite of a high-thermal conductivity (in this case,nickel) sublayer (see Figure 2) and alow-thermal-conductivity (in this case,photoresist) sublayer. As a hot fluidflows through from, say, top to bottom,heat from the fluid is transferred to,and stored in, the cell walls of this de-

vice. Next, when cold fluid flows frombottom to top, heat is transferred fromthe cell walls to the fluid.

Axial thermal conduction in such adevice can be minimized by constructingit of many layers containing low-thermal-conductivity sublayers. The offset of ad-jacent layers creates reduced-size, par-tially isolated cross sections for thermalconduction, thereby further reducingthe overall axial thermal conductance ofthe device.

Once the device has been installedin its intended operational setting(e.g., as a regenerative heat exchangerin a Stirling machine), some delamina-tion of the layers is permissible, pro-vided that the half-cell offset between

adjacent layers is maintained and nodebris is produced. Indeed, delamina-tion enhances performance by reduc-ing axial thermal conduction.

The reasonably high porosity of thedevice helps to keep the pressure droplow. The layer-to-layer offset disrupts theformation of boundary layers in the flow,thereby contributing to the mainte-nance of low pressure drop. Disruptionof boundary layers also increases the co-efficient of transfer of heat between thefluid and the cell walls.

This work was done by Matthew E. Moranof Glenn Research Center, and StephanStelter and Manfred Stelter of Polar ThermalTechnologies. Further information is con-tained in a TSP (see page 1).

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-17526-1.


Figure 1. The Configuration of This Device Is Typical of a microscale heat exchanger as described in thetext. The materials and dimensions shown here are representative rather than exclusive: they were cho-sen in a design trade study in which account was taken of capabilities for fabrication and of require-ments for low axial thermal conduction and low pressure drop under a specific set of assumptions.

Flow

Flow

MULTILAYER DEVICEMAGNIFIED VIEW OF AFEW CELLS

Nickel

Photoresist

LEW-17526 fig 1ABPI

12-22-03 es

Figure 2. A Microphotograph of a portion of onelayer of the fabricated microscale heat ex-changer shows the grating structure and surfacefeatures of the nickel sublayer.

LEW-17526 fig 2ABPI

rev 08-31-04 LE


Information Sciences

Practical Simulator Network (PSimNet)is a set of data-communication protocolsdesigned especially for use in handlingmessages between computers that are en-gaging cooperatively in real-time or nearly-real-time training simulations. In a typicalapplication, computers that provide indi-vidualized training at widely dispersed lo-cations would communicate, by use ofPSimNet, with a central host computerthat would provide a common computa-tional-simulation environment and com-mon data. Originally intended for use insupporting interfaces between trainingcomputers and computers that simulatethe responses of spacecraft scientific pay-loads, PSimNet could be especially wellsuited for a variety of other applications —for example, group automobile-drivertraining in a classroom. Another potentialapplication might lie in networking of au-tomobile-diagnostic computers at repairfacilities to a central computer that wouldcompile the expertise of numerous techni-cians and engineers and act as an expertconsulting technician.

Heretofore, a message transported in adata-communication network has been ofone of two types: delivery-critical or time-

critical. Networks that transport delivery-critical messages need protocols that as-sure the sending computers that delivery-critical messages are in fact delivered tothe proper recipient computers. Networksthat transport time-critical messages needprotocols that deliver messages to the in-tended recipient computers as quickly aspossible because the value of a time-criti-cal message diminishes over time. (Typi-cally, it is better to send an updated time-critical message than to time out andresend a stale time-critical message.) Priorto the conception of PSimNet, there wasno available set of protocols that would en-able a network to handle both time-criticaland delivery-critical messages.

PSimNet is built on the TransmissionControl Protocol/ Internet Protocol(TCP/IP) suite of protocols, which in-cludes the TCP and the user datagramprotocol (UDP). The TCP/IP and theUDP protocols offer offsetting advantagesand disadvantages with respect to the reli-ability needed for transport of delivery-critical messages and the speed needed fortransport of time-critical messages:TCP/IP provides some assurances ofproper delivery of a delivery-critical mes-

sage, but at the cost of some overhead.UDP does not offer such assurances, butoffers greater efficiency, which is an advan-tage for delivering time-critical messages.

The figure depicts the relationshipsamong the hardware and software subsys-tems and a message in a data-communica-tion system that uses PSimNet. The overallfunction of the system is to convey a mes-sage from the first computer to a secondcomputer. The message begins with aheader that, among other things, specifiesa delivery characteristic (time-critical or de-livery-critical). First, the message is re-ceived by a message receiver, then its deliv-ery characteristic is analyzed by amessage-type determinator. On the basis ofthe analyzed delivery characteristic, a mes-sage-delivery selector then selects the pro-tocol to be used in transporting the mes-sage. In the present version of PSimNet,the preferred protocol for time-criticalmessages is UDP, while that for delivery-critical messages is a modified version ofUDP that incorporates additions to make itmore reliable.

A message transporter then transportsthe message to the second computer byuse of the selected protocol. The mes-sage transporter can send the messageto multiple computers if instructed to doso by the first computer.

This work was done by John P. Balcerowskiand Milton Dunnam of Hughes ElectronicsCorp. (now Raytheon Corp. Aerospace Engi-neering Services Division) for JohnsonSpace Center. Further information is con-tained in a TSP (see page 1).

Title to this invention, covered by U.S.Patent No. 6,101,545, has been waivedunder the provisions of the National Aero-nautics and Space Act 42 U.S.C. 2457 (f).Inquiries concerning licenses for its commer-cial development should be addressed to:

Hughes Electronics Corp.200 N. Sepulveda Blvd.P.O. Box 956El Segundo, CA 90245-0956

Refer to MSC-23147, volume and number ofthis NASA Tech Briefs issue, and the pagenumber.

Protocols for Handling Messages Between SimulationComputersBoth time-critical and delivery-critical characteristics are accommodated.Lyndon B. Johnson Space Center, Houston, Texas

Message Timers

Acknowledgements

Sequence Numbers

First Computer

MessageDelivery Type

DeliveryCharacteristic

TimeCritical

DeliveryCritical

Message

MessageReceiver

Message-TypeDeterminator

Message-DeliverySelector

MessageTransporter

SecondComputer

MultipleComputers

Time-CriticalMessageProtocol

EnhancedUDP

UDP

Delivery-CriticalMessageProtocol

MSC23147

1-19-01 bs

A Message Is Sent From One Computer to Another in a data-communication network that utilizes thePSimNet set of protocols. PSimNet makes it possible to handle both time-critical and delivery-criticalmessages in the same network.


A computational method and soft-ware to implement the method havebeen developed to sift through vastquantities of digital flight data to alerthuman analysts to aircraft flights thatare statistically atypical in ways that sig-nify that safety may be adversely af-fected. On a typical day, there are tensof thousands of flights in the UnitedStates and several times that numberthroughout the world. Depending onthe specific aircraft design, the volumeof data collected by sensors and flightrecorders can range from a few dozento several thousand parameters per sec-ond during a flight. Whereas these datahave long been utilized in investigatingcrashes, the present method is orientedtoward helping to prevent crashes byenabling routine monitoring of flightoperations to identify portions of flightsthat may be of interest with respect tosafety issues.

Experience has taught that statisti-cally atypical flights often pose safety is-sues. Conventional methods of findinganomalous flights in bodies of digitalflight data require users to pre-definethe operational patterns that constituteunwanted performances. Typically, forexample, a program examines flightdata for exceedences (e.g., instances ofexcessive speed, excessive acceleration,and other parameters outside normalranges). In other words, a conventionalflight-data-analysis computer programfinds only the patterns it is told to seek

in flight data, and is blind to newlyemergent patterns that it has not beentold to seek. The present method over-comes this deficiency in that it does notrequire any pre-specification of what tolook for in bodies of flight data. Themethod is based partly on the principlethat it is necessary, not only to look forexceedences, but to go beyond excee-dences, looking for more subtle datapatterns that often cannot be pre-scribed in advance.

The method involves a series of pro-cessing steps that convert the massivequantities of raw data, collected duringroutine flight operations, into useful in-formation. The raw data are progres-sively reduced using both deterministicand statistical methods. Multivariatecluster analysis is performed to groupflights by similarity with respect to flightsignatures derived from parameter val-ues. The process includes analysis ofmultiple selected flight parameters fora selected phase of a selected flight, andfor determining when the selectedphase of the selected flight is atypical incomparison with the correspondingphases of other, similar flights for whichcorresponding data are available. Foreach flight, there is computed an atypi-cality score based partly on results fromthe cluster analysis. The distribution ofatypicality scores of all flights is used toidentify flights for examination.

The data from each day’s flights areprocessed during the night and sum-

marized in a document, called the“Morning Report,” that includes a listof the 20 percent of flights having thehighest atypicality scores, ranked inorder of descending atypicality score.For each flight, the report includes aplain-language description of whatmakes the flight atypical. With thehelp of software designed to be intuitively useable, an analyst worksthrough this list of flights to the finestlevel of detail where necessary, exam-ining the characteristics that madethem atypical, assessing their opera-tional significance, and determiningthe need for further action.

This work was done by Irving Statler,Thomas Chidester, and Michael Shafto ofAmes Research Center; Thomas Ferry-man, Brett Amidan, Paul Whitney, AmandaWhite, Alan Willse, Scott Cooley, Joseph Jay,Loren Rosenthal, Andrea Swickard, DerrickBates, Chad Scherrer, and Bobbie-Jo Webb ofBattelle Memorial Institute; Robert Lawrenceof Safe Flight; Chris Mosbrucker, GaryProthero, Adi Andrei, Tim Romanowski,Daniel Robin, and Jason Prothero ofProWorks Corporation; Robert Lynch ofFlight Safety Consultants; and Michael Loweof the U.S. Navy. Further information is con-tained in a TSP (see page 1).

This invention has been patented by NASA(U.S. Patent No. 6,937,924). Inquiries con-cerning rights for the commercial use of thisinvention should be addressed to the AmesTechnology Partnerships Division at (650)604-2954. Refer to ARC-15041-1

Statistical Detection of Atypical Aircraft FlightsA priori specification of search criteria is not necessary.Ames Research Center, Moffett Field, California

NASA’s Aviation Safety and Modeling ProjectCapabilities for automated analysis of flight data are under development.Ames Research Center, Moffett Field, California

The Aviation Safety Monitoring andModeling (ASMM) Project of NASA’sAviation Safety program is cultivatingsources of data and developing auto-mated computer hardware and softwareto facilitate efficient, comprehensive,and accurate analyses of the data col-lected from large, heterogeneous data-bases throughout the national aviationsystem. The ASMM addresses the needto provide means for increasing safety byenabling the identification and correct-

ing of predisposing conditions thatcould lead to accidents or to incidentsthat pose aviation risks.

A major component of the ASMMProject is the Aviation PerformanceMeasuring System (APMS), which is de-veloping the next generation of soft-ware tools for analyzing and interpret-ing flight data (see figure). Airlines,military units, corporate operators, andothers analyze aircraft flight data to identify contributing factors and

corrective actions for situations inwhich aircraft performance parametersexceed normal limits during phases offlight. The programs for performingsuch exceedance-based analyses are de-noted flight operations quality assur-ance (FOQA in civilian settings) pro-grams and take their inspiration fromstatistical process control. However,FOQA analysis involves use of only aportion of the data: large quantities ofdata are scanned to extract and under-

stand a small number of predefinedevents. The sets of data contain farmore information that is potentiallyhelpful for understanding and enhanc-ing the safety, reliability, and economyof flight operations.

The challenge is to find and under-stand key information from the mass ofdata generated by aircraft and collectedby data recorders. There is a need to au-tomate scanning, analysis, and report-ing to produce meaningful informationupon which human analysts can act.The APMS software is intended to satisfythis need. Beginning with workload en-hancements to exceedance-basedFOQA analyses and progressing to so-phisticated multivariate statistical analy-ses, the APMS has developed key soft-ware tools to advance the science offlight-data analysis.

Another major component of theASMM Project is the Performance DataAnalysis and Reporting System(PDARS), which is developing network-ing and analysis hardware and softwarefor application air-traffic-control(ATC) radar data. PDARS, which wasdeveloped jointly with the Federal Avi-ation Administration (FAA) Office ofSystem Capacity, provides ATC deci-sion-makers at the facility level with acomprehensive means of monitoringthe health, performance, and safety ofday-to-day ATC operations. PDARS en-ables analysis of daily operation of theNational Airspace System (NAS) atlocal and inter-facility levels. By trans-lating flight-track and flight-plan datainto useful performance information,PDARS significantly augments the abil-

ity of the FAA to adjust operationalprocedures and techniques. The netoutcomes of these adjustments arequantifiable improvements in safetyand efficiency throughout the NAS.

Future progress in aviation safety canbe expected to involve the routine inte-gration of information from APMS,PDARS, and other sources. Such inte-gration will require both greater depthof analysis of individual data sourcesand automated ability to integrate in-formation extracted from diversesources to draw sound conclusions on

causal factors and risk assessment. Italso will be necessary for users to evalu-ate such integration to determine itsusefulness. It is planned to develop ca-pabilities for such integration duringthe next few years.

This work was done by Thomas R.Chidester and Irving C. Statler of Ames Research Center. Further information iscontained 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-14362-1.


ARC14362-1ABPI

8-24-04 CC

The Aviation Performance Measuring System — a component of NASA’s Aviation Safety Monitoringand Modeling (ASMM) Project — is developing the means of automated processing of large collec-tions of flight data to help answer questions pertaining to performance and safety.


Books & Reports

Multimode-Guided-Wave Ul-trasonic Scanning of Materials

Two documents discuss a method ofcharacterizing advanced composite ma-terials by use of multimode-guided ultra-sonic waves. The method at an earlierstage of development was described in“High-Performance Scanning Acousto-Ultrasonic System” (LEW-17601-1),NASA Tech Briefs, Vol. 30, No. 3 (March2006), page 62. To recapitulate: A trans-mitting transducer excites modulated(e.g., pulsed) ultrasonic waves at one lo-cation on a surface of a plate specimen.The waves interact with microstructureand flaws as they propagate through thespecimen to a receiving transducer at adifferent location. The received signal isanalyzed to determine the total (multi-mode) ultrasonic response of the speci-men and utilize this response to evaluatemicrostructure and flaws. The analysis isperformed by software that extracts pa-rameters of signals in the time and fre-quency domains. Scanning is effected byusing computer-controlled motorizedtranslation stages to position the trans-ducers at specified pairs of locations andrepeating the measurement, data-acqui-sition, and data-analysis processes at thesuccessive locations. The instant docu-ments reiterate the prior descriptionand summarize capabilities of the hard-ware and software of the method at thepresent state of development. One docu-ment presents results of a scan of a spec-imen containing a delamination.

This work was done by Don Roth of GlennResearch Center. Further information iscontained in a TSP (see page 1).

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, Cleve-land, Ohio 44135. Refer to LEW-17527.

Algorithms for ManeuveringSpacecraft Around SmallBodies

A document describes mathematicalderivations and applications of au-tonomous guidance algorithms for ma-neuvering spacecraft in the vicinities ofsmall astronomical bodies like comets orasteroids. These algorithms compute

fuel- or energy-optimal trajectories fortypical maneuvers by solving the associ-ated optimal-control problems with rele-vant control and state constraints. In thederivations, these problems are con-verted from their original continuous(infinite-dimensional) forms to finite-di-mensional forms through (1) discretiza-tion of the time axis and (2) spectral dis-cretization of control inputs via a finitenumber of Chebyshev basis functions. Inthese doubly discretized problems, theChebyshev coefficients are the variables.These problems are, variously, eitherconvex programming problems or pro-gramming problems that can be convex-ified. The resulting discrete problemsare convex parameter-optimizationproblems; this is desirable because onecan take advantage of very efficient androbust algorithms that have been devel-oped previously and are well establishedfor solving such problems. These algo-rithms are fast, do not require initialguesses, and always converge to globaloptima. Following the derivations, thealgorithms are demonstrated by apply-ing them to numerical examples of fly-by, descent-to-hover, and ascent-from-hover maneuvers.

This work was done by A. Bechet Acikmeseand David Bayard of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).

The software used in this innovation isavailable for commercial licensing. Pleasecontact Karina Edmonds of the CaliforniaInstitute of Technology at (626) 395-2322.Refer to NPO-41322.

Improved Solar-Radiation-Pressure Models for GPSSatellites

A report describes a series of compu-tational models conceived as an im-provement over prior models for deter-mining effects of solar-radiationpressure on orbits of Global PositioningSystem (GPS) satellites. These modelsare based on fitting coefficients ofFourier functions of Sun-spacecraft-Earth angles to observed spacecraft or-bital motions. Construction of a modelin this series involves the following steps:1. Form 10-day “truth” orbit arcs from

precise daily GPS orbit data gatheredduring more than four years.

2. Construct a model of the solar-radia-

tion pressure and estimate model-pa-rameter values that make a least-squares best fit of the model-predictedtrajectory to each of the “truth” 10- dayorbit arcs.

3. Using a least-squares procedure andutilizing the full covariance informa-tion from each 10-day fit, combine theestimates from all satellite arcs into asingle set of model parameters for thetwo GPS constellations of the satellitesnow or soon to be placed in service.

4. Evaluate the model thus derived bymeans of orbit-data-fit and orbit-pre-diction tests.In evaluations performed thus far,

these models have been found to offeraccuracies significantly greater thanthose of the prior models.

This work was done by Yoaz Bar-Sever andDa Kuang of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).NPO-41395

Measuring Attitude of aLarge, Flexible, OrbitingStructure

A document summarizes a proposedmetrology subsystem for precisely meas-uring the attitude of a large and flexiblestructure in space. Two cameras wouldbe mounted at the base of the structure:(1) A star camera equipped with two

separate fields of view: (a) imagingstars in the background near thestructure tip while excluding the tipfrom view to prevent saturation fromsunlight reflected from the tip, and(b) imaging the tip and have simul-taneous stars in the background.First, in the absence of reflected sun-light and with the self-illuminatedfiducials on the structure turned off,the star camera would open bothfields of view and establish the angu-lar relationship between the twofields of view.

(2) The second camera (metrology cam-era) is too insensitive to observe starsbut sensitive enough to image anumber of bright self-illuminatedfiducials on the structure through anarrow band pass filter (even in thepresence of sunlight) at high rates.Still in the absence of sunlight, theself-illuminated fiducials at the tip ofthe structure are imaged simultane-


ously by the star and metrology cam-eras to establish the relationship be-tween the two different cameras.

During operations, the star camerawould be operated in the tip-excludingconfiguration while the metrology cam-era tracked the tip. The orientation ofthe base to tip line relative to the starswould be determined by use of informa-

tion from the metrology camera, the starcamera, inertial measurement unit(IMU), and calibration data. The advan-tage of the proposed scheme is (1) it ispossible to obtain attitude knowledge athigh rates (based on IMU and metrol-ogy camera); (2) it is possible to operatewhen the antenna is illuminated by theSun; and (3) it is possible to perform on-

orbit alignment after launch.This work was done by Carl Christian

Liebe, Randall Bartman, AlexanderAbramovici, Jacob Chapsky, Edward Litty,and Keith Coste of Caltech for NASA’s JetPropulsion Laboratory. Further informa-tion is contained in a TSP (see page 1).NPO-41411

National Aeronautics andSpace Administration

technology focus electronics/computersbetter. optionally, the system can easily be connected to a...

Documents