STATUS OF FREE-PISTON STIRLING TECHNOLOGY AT SUNPOWER, INC.

J. Gary WoodSunpower, Inc.

Athens, OH

ABSTRACTThis paper includes an overview of the status of free-piston Stirling machine technology at Sunpower, Inc.Sunpower has achieved record-setting performance andspecific power for a wide range of machines primarilydesigned for commercial applications. These machinesrange in power from approximately 100 We inputStirling and pulse-tubes cryocoolers to approximately1300 We output engines designed for European domes-tic cogeneration. An overview of the performance andspecific powers of these current machines is presented.Also included is a description of Sunpower’s designtopology that uses gas bearings to ensure long lifethrough the use of a compliant connection withmechanical springs. All Sunpower machines alsoinclude low cost and high specific power linearalternator/motors that make maximum use of alternatormaterials. The major portion of the paper describes thedesign and current status of an advanced 35 Weconvertor (engine) that promises even higher projectedperformance. This convertor is currently underdevelopment with NASA Phase II SBIR funding.Because of its extremely high conversion efficiency,this small convertor is attractive for use in a Stirlingradioisotope space power system as a possiblereplacement for Radioisotope ThermoelectricGenerators (RTGs). Achieving the projected highefficiency would allow a reduction in the amount ofradioisotope fuel by approximately a factor of four.The small convertor is projected to have an efficiencyof greater than 50 percent of Carnot efficiency whenoperating at a temperature ratio of 2.6. Projectedspecific power of the convertor, which is designed torun at 100 Hertz, is approximately 90 W/kg. Thisengine also has significant terrestrial applications as afuel-fired battery replacement. Liquid fuels haveapproximately 300 times the energy density as Ni-Cadbatteries and 150 times that of Lithium-ion batteries.Thus such an engine coupled with a liquid fossil fuelburner is an attractive product.

CURRENT AREAS OF DEVELOPMENTSunpower was founded in 1974 by William Beale, theinventor of the free-piston Stirling engine. Today,nearly 30 years later, Sunpower works on a widevariety of free-piston machines ranging from Stirlingcycle engines and cryocoolers, to linear compressors forpulse tube, Joule-Thomson, and Rankine cycle cooling.

Sunpower’s overriding goal is the commercialization ofthese efficient and mechanically simple machines. Commercialization efforts have resulted in several free-piston machines now in production under license toSunpower. LG Electronics is currently marketing itsfree-piston based (Rankine cycle) refrigerator DIOSmodel in Korea, to be introduced worldwide in the nearfuture. CryoGen, a California based company, isselling a free-piston compressor-based JT cryosurgicaldevice. ISL, a French company, markets an aircraftfuel gel-point tester based on a free-piston Stirlingcooler. Superconductor Technologies Inc (STI) has1000s of telecommunication cryocoolers in the field.All of these machines are built under license toSunpower.

Several other machines are currently in field trials or inspecialized applications. Among these is a nominally1000 We engine (based on the Sunpower EG-1000engine) which is in field trials in Europe as part of adomestic cogeneration system. NASA-Goddard hasalso used Sunpower cryocoolers in several applications.Most notable of these is the RHESSI satellite whichwas launched in February 2002 to study solar flares.RHESSI relies on a Sunpower M77 cryocooler to coolits sensors, and minimum mission science objectiveshave already been achieved. Sunpower has also built acryocooler manufacturing facility with the capability ofmanufacturing tens of thousands of cryocoolers peryear.

In addition to the commercial effort, severalgovernment-funded programs on advanced free-pistonconcepts as well as basic research are currently beinginvestigated. Current government programs include oursecond Phase II SBIR pulse-tube cryocooler for NASA-Goddard. The objective of this project is a 3-stagepulse-tube cryocooler achieving <10 K with an inputpower of 200 We.

Another current Phase II SBIR is funded by NASA-Glenn for a highly efficient and lightweight 35 Weconvertor (engine) for possible use in an advancedStirling radioisotope space power system. Stirlingradioisotope power systems are being developed byDOE and NASA as a possible replacement forRadioisotope Thermoelectric Generators (RTGs),promising to reduce the amount of required plutonium

Copyright © 2003 by Sunpower Inc. 1st International Energy Conversion Engineering Conference, 17-21 August 2003, Portsumouth, Virginia

85

1

by a factor of four. This 35 We convertor is describedlater in this paper.

Sunpower also operates a unique in-house oscillatingflow regenerator heat transfer test rig. This rig wasdesigned by Sunpower for NASA-Lewis (now Glenn)in the 1980’s and currently is on loan to Sunpower.

PERFORMANCE AND SPECIFIC POWER While the majority of Sunpower machines are designedfor low manufacturing cost, the efficiency and specificpower of these machines typically exceed that ofmachines where low production cost was not a majordesign constraint. The high thermodynamicperformance of Sunpower’s cryocoolers and engines ispresented in the following two figures.

Sunpower machines also have exceptionally highspecific power. This is largely due to an efficient andmechanically simple linear alternator/motor design.Typically the alternator/motor and the surroundingvessel make up the major mass of the machine.

It is important to consider operating frequency whencomparing specific power of machines. Alternator massat a given electrical efficiency is inversely proportional

to frequency. Frequency in many cases is specified bythe intended end use. For example the Sunpower EG-1000 engine is designed for European grid-connectedcogeneration which requires operation at 50 hertz.

It is thus desirable to compare machines on a“frequency adjusted specific power” basis where theunits are of the form W/(kg*Hz). Figure 3 presentsrepresentative frequency adjusted specific powers ofSunpower’s current machines.

It must be noted that Figure 3 is intended only to be anapproximate representation of specific power. For agiven alternator/motor configuration, electricalconversion efficiency can be increased (withdiminishing returns) by increasing alternator mass.This is the primary reason for the difference in the twopoints labeled “SBIR I Engine” and “SBIR II Engine”in Figure 3, which is further explained in reference 15.Also the rejection temperature of the machineinfluences alternator efficiency, primarily because thecoil resistance increases with temperature.

Figure 3 includes machines with largely differentdesign constraints and operating temperatures. As suchit only presents an approximate representation ofachievable specific powers.

Figure 1: Comparative Performance of Cryocoolers with Motor (from references 1-14)

0

5

10

15

20

25

0 20 40 60 80 100 120Cold-End Temperature (K)

Frac

tion

of C

arno

t (%

) Sunpower Pulse Tube1

Sunpower M873

Sunpower M772

Mixed Gas Joule-Thomson6

Creare Brayton8

Hymatic Stirling7

Ball 2-Stage Stirling SB23012 APD Joule-Thomson

Cryotiger9

Ball 1-Stage Stirling SB16011

TRW 1-Stage Miniature Stirling15

Ricor K535 1-Stage Stirling13

Leybold Coolpower 150

1-Stage Gifford-McMahon14

Lockheed Martin Pulse

Tube10

Sunpower MiniTel5

Sunpower CryoTel4

2

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1.0 1.5 2.0 2.5 3.0 3.5

Temperature Ratio (Theater / Trejector)

Engi

ne-o

nly

Effic

ienc

y (fr

actio

n of

Car

not)

Sunpower EG-1000 (2000)

Sunpower RE-1000 (1979)

Projected EG-1000 with improvements

TDC 55 W (2000)

MTI SPRE (1990)

MTI CPTC (1993)

MTI Mod 2 kinematic hydrogen (1987)

Curzon-Ahlborn Efficiency (for reference only)

Predicted 35 W SBIR engine

Figure 2: Stirling Engine-Only Efficiencies (PV power/ Heat into Head)From reference 15

Lightweight TDC19

(55 W, 80 Hz, 1.7 kg)

TDC19

(55 W, 80Hz 4.0 kg)

SBIR I Engine M-87 CryocoolerSBIR II Engine

EG-1000 Engine with Flanges

MTI SPRE18

MTI CPTC (projected light weight)17

EG-1000 without flanges

0.00.20.40.60.81.01.21.41.61.82.0

10 100 1000 10000 100000Electrical Power watts (scaled to 100 Hz)

Freq

uenc

y A

djus

ted

Spec

ific

Pow

er (W

/(kg*

Hz)

Figure 3: Representative Frequency Adjusted Specific Power of Free-Piston Machines

3

SUNPOWER MACHINE DESIGN TOPOLOGYFigure 4 is from US patent number 5,525,845 of a free-piston compressor. In this figure, the labels 35-38identify the gas bearings that provide non-contactoperation between the piston (31) and the cylinder (32).

The planar springs (46) provide mechanical springingonly. The compliant link (47) is designed for flexibilitywithout buckling under the dynamic loads. Thiscompliant link largely reduces costs because highprecision is not required of the planar springs nor of thearea of the mounting to the casing (48). Thiscompliance allows greatly reduced cost of themanufactured parts, simple assembly because ofreduced alignment constraints, and a very compactoverall assembly.

Nearly all Sunpower machines incorporate thefollowing design features, which are illustrated inFigure 4.:

1. Self-pumped gas bearings for contact freeoperation

2. Low cost planar mechanical springs used forspringing only

3. Patented compliant (flexible) method of interfacing1 and 2

4. Hermetic sealing5. Patented low mass, high efficiency and

mechanically simple alternator, which makesmaximum use of materials

The figure also illustrates the basic configuration ofSunpower’s linear alternators and motors. A singleannular ring of magnets (42) oscillates within the stator(40-41) which produces AC power in the coil (44).This alternator is both highly efficient and has lowspecific mass. Some of the advantages are that the entire copper coil is within the reversing flux, and thatthe magnets never leave the stator (therefore reducingeddy current losses in the surrounding structure).

Figure 4

CURRENT ENGINE RESEARCH

EG-1000 EngineAmong Sunpower current engine development work isthe EG-1000 engine which is nominally 1000 We butproduces up to 1300 We. The research form of thismachine, which includes flanges, is shown in Figure 5.This engine is currently in field trials in England as partof a residential cogeneration package. This engineincluding alternator is 29% efficient (electrical out toheat into the head) when operating at a temperatureratio of 2.6 (550 C hot end and 50 C reject). Althoughthis machine is designed for low cost, it has the highestefficiency (as a percent of Carnot) of any free-pistonengine to date as is illustrated in Figure 2. Meanwhilethe low cost constraint has pushed this design to have arelatively high specific power as shown in Figure 3.This machine has a mass of 35 kg and is approximately230mm in diameter and 440mm in length.

Figure 5. EG-1000 Engine

35 Watt ConvertorAnother active engine project at Sunpower is a 35 Weconvertor being developed under NASA Phase II SBIRfunding. Figure 6 shows both the current research formand a model of the final configuration with alightweight vessel. More specific details of thismachine can be found in reference 15.

The first run of this machine is scheduled for midyear2003. Projections for this machine, which operates at100 Hz, is a power output between 35 and 40 watts with

4

a final configuration mass of 434 gm. The projectedefficiency of this machine is expected to fall between30.5 to 35% when operating at a temperature ratio of2.6. The projected specific power of this machine isincluded in Fig. 3, labeled SBIR II engine.

Figure 6: NASA-funded 35 We Convertor (in testvessel at left, model of final configuration on right)shown with Ohio state quarter for size scale

This convertor is designed to be the smallest high-efficiency Stirling engine developed to date, and assuch represents a significant technical challenge.However, the machine is similar in physical size andinternal construction to Sunpower’s line of cryocoolers.Figure 7 compares the size of this convertor toSunpower’s M87 (150 We @ 60 Hz) and Cryotel (100We @ 60 Hz) cryocoolers.

This 35 We convertor is being developed for possibleuse in an advanced Stirling radioisotope space powersystem. Stirling radioisotope power systems are ofinterest to NASA as a possible replacement for muchlower efficiency Radioisotope ThermoelectricGenerators (RTGs). RTGs have been used as the powersource on many NASA missions (Voyager, Cassini,etc.). Stirling radioisotope power systems offer asignificant reduction (~4) in the required amount ofPlutonium 238, as RTGs have a very low conversionefficiency (~5-7%)

RTGs are very reliable devices, with proven longlifetime. Typical life requirements set by NASA are100,000+ hours of operation. However, Stirling

Figure 7: Size Comparison. M87 Cryocooler (left),Cryotel (center), and NASA 35 watt convertor (right)

machines are also demonstrating their long life in manycircumstances. Both Stirling Technology Company andSunpower have single unit life demonstrations that haveachieved more than 70,000 hours of operation and arestill accumulating hours. Thus, it is becoming evidentthat Stirling convertors can meet the life requirements.Confidence in reliability is also greatly increased asmore and more free-piston devices are placed in thefield as described earlier under Sunpower’scommercialization efforts.

Terrestrial Applications for the 35 We ConvertorThere is also significant potential for the 35 Weconvertor in commercial terrestrial applications as afuel-fired battery replacement. Batteries haveextremely low energy density when compared to liquidfuels. Roughly, liquid fuels have 300 times the energydensity of NiCad batteries and 150 times that ofLithium-ion batteries. Even when considering theconversion efficiency of a small engine plus burner(roughly between 25 and 33%), a small fuel-firedStirling engine would have extremely high energydensity compared to battery packs, approaching 50times that of Lithium-ion batteries.

SUMMARYThis paper presents an overview of current free-pistonStirling machine status at Sunpower. In accordancewith the most recent data available, we compare theperformance of Sunpower machines to other currentand past machines. Through a combination of manyyears of government-funded research andcommercialization efforts, free-piston machines aredemonstrating their long promised high efficiency andreliability.

5

REFERENCES1. Sunpower Pulse Tube, 4.8 W @ 77 K, Internal Sunpower Test 2001.2. Sunpower M77, 5 W @ 77 K, 313 K Reject, Internal Sunpower Test 1997.3. Sunpower M87, 7.5 W @ 77 K, 323 K Reject, Internal Sunpower Test 1999.4. Mixed Gas Joule-Thomson, 100 W @ 100 K, 245 K

Reject, “Mixed Gas J-T Cryocooler with Precooling Stage,” 10th International Cryocooler Conference,

1998, Monterey CA. 5. Hymatic Miniature Stirling, 0.3 W @ 80 K, 293 K Reject, “Miniature Long Life Tactical Stirling Cryocoolers,” 9th International Cryocooler

Conference, 1996, Waterville Valley NH.6. Creare Brayton 1-Stage, 5 W @ 65 K, 310 K

Reject, “A Single Stage Reverse BraytonCryocooler: Performance and Endurance Tests onthe Engineering Model,” 9th InternationalCryocooler Conference, 1996, Waterville ValleyNH.

7. APD Joule-Thomson Cryotiger, 3 W @ 80 K, 308K Reject, “A Throttle Cycle Cryocooler Operatingwith Mixed Gas Refrigerants in 70 K to 120 KTemperature Range,” 9th International CryocoolerConference, 1996, Waterville Valley NH.

8. Lockheed-Martin Pulse Tube, 2 W @ 60 K, 295 KReject, “Development of a 2 W at 60 K PulseTube Cryocooler for Spaceborne Operation,” 10th

International Cryocooler Conference, 1998,Monterey CA.

9. Sunpower HTSC Filter Cooler, 12 W @ 87K Internal Sunpower Test 2001.10. Ball 1-Stage Stirling SB160, 1.6 W @ 60 K, 298 K

Reject, Ball Aerospace test,www.ball.com/aerospace/crysb160.html, lastupdated 1999.

11. Ball 2-Stage Stirling SB230, 0.45 W @ 30 K, 298 KReject, Ball Aerospace test,www.ball.com/aerospace/crysb230.html, last

updated 1999.12. Ricor K535 1-Stage Stirling, 4 W @ 65 K, 318 K

Reject, Ricor test, www.ricor.com/k535.htm.13. Leybold Coolpower 150 1-Stage Gifford McMahon,

150 W @ 77 K, 311 K Reject, Leybold test,www.leyboldcryogenics.com/coolpower150.html.

14. TRW 1-Stage Miniature Stirling, 0.25 W @ 65 K,290 K Reject, “Miniature Long-Life Space-Qualified Pulse Tube and Stirling Cryocoolers,” 8th

International Cryocooler Conference, 1994, VailCO.

15. Wood, J.G. and Lane, NW, “Advanced 35 WattFree-Piston Stirling Engine for Space PowerApplications,” STAIF February 2003 AlbuquerqueNew Mexico.

16. Sunpower Inc. test data, using electrically heatedhead. P-V power / heat into head. P-V powercalculated from electric output assuming thealternator is 85% efficient.

17. Dhar, M., Stirling Space Power Program, Volumes1 & 2, Final Report, 1997, NASA/CR-1999-209164/ Vol. 1. Point shown is average of heat-to-water and heat-to-head means of calculatingefficiency (from fig. 79). Heat-to-head efficiency is45% of Carnot efficiency and heat-to-waterefficiency is about 54 % of Carnot efficiency

18. Dochat, G., SPDE / SPRE Final Summary Report,1993, NASA/CR-187086, data based on heat-to-water from plot on pg. 92, max efficiency (upperpoint) occurs at 35 % of design power (lower point6

Data from NASA web site 19. White, M. Technology Development of a Free-

Piston Stirling Advanced Radioisotope Space PowerSystem, STAIF February 1999 Albuquerque NewMexico

6