N92-22720

THERMOACOUSTIC REFRIGERATION

Steven L. Garrett and Thomas J. Hofler

Physics Department - Code PH/Gx
Naval Postgraduate School
Monterey, CA 93943

ABSTRACT

A new refrigerator which uses resonant high amplitude sound in inert gases to pump heat will be described
and demonstrated. The phasing of the thermoacoustic cycle is provided by thermal conduction. This "natural"
phasing allows the entire refrigerator to operate with only one moving part (the loudspeaker diaphragm). The
thermoacoustic refrigerator has no sliding seals, requires no lubrication, uses only low-tolerance machined parts, and
contains no expensive components. Because the compressor moving mass is typically small (,, 15 gin) and
oscillation frequency is high (- 400 Hz), the small amount of vibration is very easily isolated. This low vibration
and lack of sliding seals makes thermoacoustic refrigeration an excellent candidate for space applications. Since the
thermoacoustic refrigerators use no CFC's and have coefficients-of-performance which are competitive with
conventional vapor compression cycle refrigerators, thcrmoacoustics is also a good candidate for food rcfrigeration
and commercial/residential air conditioning applications. The design, fabrication, and performance of the first
practical, autonomous thermoacoustic refrigerator, which will be flown on the Space Shuttle (STS-42), will be
described and designs for terrestrial applications will be presented.

INTRODUCTION

The history of refrigeration technology during the second half of the 20th century has been singularly
uninteresting. Since the introduction of chlorofluorocarbons (CFC) as the working fluid in a vapor compression
refrigeration cycle these chemicals have become dominant in almost all small and medium scale food
refrigerator/freezer and building/residential air conditioner applications. That situation is about to change
dramatically 1 and, at this moment, unpredictably.

The End of the CFC Era

At present, it is estimated that there is $135 billion of products which uses CFC's within the United States
alone 2. The spectacular success of CFC's in refrigeration was brought about by their good thermodynamic
properties (,phase changes at modest pressure at the required temperatures) and their excellent chemical stability which
made them compatible with hydrocarbon lubricants. It was this compatibility with lubricants which allowed the
production of compressors, bathed in oil, which could operate for decades without maintenance. (When was the last
time you had to replace your home refrigerator compressor7.) It is now known that the chemical stability of the
CFC's has lead to their ultimate downfall.

Two recent events are responsible for the "new era" in refrigeration which will dawn at the beginning of the
21st cenlmT. The most significant of these is the international ban on the production of CFC's which were found to
be destroying the Earth's protective ozone layer. The ban was brought about with the signing on January 1, 1989,
of the Montreal Protocols. This international agreement, signed by the United States and thirty-one other
industrialized countries, started a stepped reduction in the production of CFC's to 20% of their present levels by
1993, one-half of their present levels by 1998, and imposes a complete ban on their production worldwide by the
year 2000.

The second event was the discovery of "high temperature" superconductors and the development of high
speed and high density electronic circuits which require active cooling. Although the immediate impact of this
merging requirement for cryocoolers is dwarfed by the revolution which will be brought about by the Montreal

397 _RECEDING PAGE BLANK NOT F!LMED-,

Protocols, much of the longer term futm¢ development of high speed electronics, electro-_tics, long-haul fiber-optic
communication, and computers, will be dictated by the availability, reliability, and efficiency of cryocooler
technology.

Alternative Refrigeration Technolo_es

The chemical companies have not abandoned the refrigeration business. As one might expect there has been
a rush to develop alternative fluids which are not as detrimental to the ozone layer. Thus far, several HCFC
compounds, which will be banned in 2010, have been developed but these have exhibited compatibility problems
with hydrocarbon lubricants 2 and some have recently been found to be ca_inngenic 3. This has lead others to
reconsider fluids which were in use before the rise of CFC (ammonia, carbon dioxide, etc.) and other refrigeration
cycles such as Stirling and Malone cycles which do not require phase changes.

It is the p_ of this paper to introduce an entirely new approach to refrigeration which was first
discovered 4 in the early 1980's which uses high intensity sound waves to pump heat using inert gases as the
working fluid.

THE THERMOACOUSTIC HEAT PUMPING CYCLE

The interaction between acoustics and thermodynamics has been recognized ever since the dispute between
Newton and Laplace over whether the speed of sound was determined by the adiabatic or isothermal comwessibility
of air. Although today there are probably many physicists who might still make the wrong choice (as Newton didl),
most physicists have at least been exposed to a lecture demonstration such as the Rij3_eTube5 or have observed
(cursed?) Taconis oscillations in liquid helium6, so they are not surprised that thermal gradients can lead to the
production of sound. The reverse Wocess - thermoaconstic heat pumping - is far less well known and was the first
intentional demonstration of a new class of intrinsically irreversible heat engines.

Traditional heat engine cycles, such as the Camct Cycle typically studied in elementary thermodynamics
courses, assume that the individual steps in the cycle are reversible. Such analyses, which invoke the First and
Second Laws of Thermodynamics, lead to the limiting values for the efficicncies of prime movers and the
coefficients-of-pedormance of refrigerators. These limiting values are never realized in practical heat engines due to
the unavoidable irreversibililies, such as thermal diffusion and viscous dissipation, which always reduce perfornunge
below the ideal Carnot values. Reversible engines also require various mechanical devices (e&., valves, cams, push-
rods, linkages, timing chains, etc.) in order to execute the _ phasing of various cyclic processes (e&.,
compressions, expansions, regeneration, etc.). In thermoacoustic engines, the irreversibility due to the imperfect
(diffusive) thermal contact between the acoustically oscillating working fluid and a stationary second themodymmic
medium provides the required phasing. This "naturalphasing" has produced heat engines which require no moving
parts other than the self-maintained oscillations of the working fluid.

A Simple. Invisid. Lat,ran_an Model of the Heat Pum_na Process

Although a complete and detailed analysis of the thermoacoustic heat pumping process is well beyond the
scope of this study, the following simple, invisid. Lngrangian representation of the cycle contains the essence of the
process. A complete analysis 7 would necessarily include the gas viscosity, fmite wavelength effects, longitudinal
thermal conduction along the stationary second thermodynamic medium and through the gas, and the ratio of the gas
and second medium dynamic heat capacities. A schematic diagram of a simple, oue-quaner wavelength, 7,/4,
thermoacoustic refrigerator is shown in Figure 1. The loudspeaker at the leftsets up the standing wave within the
gas-fdled tube. Its frequency is chosen so that the loudspeaker excites the fundamental (7,/4) resonance of the tube.
The terminatiou at the fight-hand end of the tube is rigid so the longitudinal particle velocity at the rigid end is zero
(a velocity node) and the acoustical pressure variations are maximum (a pressure antinode). At the loudspeaker end of
the tube there is an acoustic pressure node and a particle velocity antinode. To the left of the termination is a stack
of plates (the "stack") whose spacing is chosen to be a few thermal penetration depths.

398
 I
N
+

\\

_;_,-
I

_il-II
I

_
x

*.i r.i +,+,

I_1 v l

I .... i ÷11 I-
i
I

• II

I,_ <_
,,,n

\\
'
I
I

II
INN
oI_\ "
!.
I- el
I.

I P_N
I
! i
I il
P- I
\\

\\ N o
U ml

iI
,,N" .\\

---stock--'-

399
 The thermal penelration depth, _, represents the distance over which heat will diffuse during a time which is
on the order of an acoustic period, t = l/f, where f is the acoustic frequency. It is defined in terms of the thermal
conductivity of the gas, Ic, the gas density, p, and its isobaric specific heat (per unit mass), Cp.
aK = .l-x-- (I)
"Vxfpcp
This, length scale is crucial to understanding the perfccmance of the thetmoacoustic cycle since the diffusive heat
trans]3on between the gas and the "stack" is only significant within this region. It is for that reason that the stack
and the spacing between its plates are central to the thermmconstic cycle.

For this analysis we will focus our attention on a small portion of the "stack" and the adjacent gas which is
und,_'going acoustic oscillations at a distance from the solid stack material which is small enough that a substantial
amount of thermal conduction can ___keplace in an amount of lime which is on the order of the acoustic period. In
the lower half of Figure 1, a small portion of the stack has been magnified and a parcel of gas undergoing an acoustic
oscillations is shown. The four steps in the cycle are represented by the four boxes which are shown as moving in a
rect_mgularpath for clarity although in reality they, of course, simply oscillate back and forth. As the fluid
oscillates backandforth alongtheplate itundergoes changesintemperature duetotheadiabatic compression and
exl_msion resulting fromtheIXeSsurc variations whichaccompanythestanding soundwave. Thecompressions and
exlxmsions ofthegaswhichconstitute thesoundwave arcadiabatic iftheyarefarfromthesurface oftheplate. The
rela_on betweenthechangeingaspressure duetothesoundwave,Pl,relative tothemean (ambient) pressure,Pro,
and the adiabatic temperature change of the gas due, T 1, to the acoustic pressure change, relative to the mean
absolute (Kelvin) temperature, Tm, is given below in equation (2).
T---L= 7" 1 P_ (2)
Tm _ Pm
The polytropic coefficient, _, is equal to the ratio of the specific heat of the gas at constant pressure to the specific
heal at constant volume and is exacdy 5/3 for the inert gases. It is smaller for all other gases but is always greater
than one.

Although the oscillations in an acoustic heat pump are sinusoidal functions of time, Figure 2 depicts the
motion as articulated (a square wave) in order to simplify the explanation. The thermodynamic cycle can be
cousidered as consisting of two reversible adiabatic steps and two irreversible isobaric (constant lzessure) steps. The
plate is assumed to have a mean temperature, Tm, and a temperature gradient,VT, referenced to the mean position, x
= 0. The temperature of the plate at the left-most position of the gas parcels excursion is therefore Tm - xlVT, and
at the right-most excursion is T m + xlVT.

In the first step of this four-step cycle, the fluid is transported along the plate by a distance 2x 1 and is heated
by ediabatic compression from a temperature of Tm - x IVT to Tm - x 1VT + 2TI. The adiabatic gas law provides
the relationship between the change in gas lZessure, Pl, and the associated change in temperature, T 1, as described in
equation (2). Because we are considering a heat pump, work, in the form of sound, was done on the-gas parcel and it
is now a temperature which is higher than that of the plate at its present location.

In the second step, the warmer gas parcel transfers an amount of heat, dQho t, to the plate by thermal
conduction at constant pressure and its temperature decreases to that of the plate, T m + Xl VT. In the third step, the
fluid is transported back along the plate to position -x 1 and is cooled by adiabatic expansion to a temperature Tm +
x 1WI"- 2TI. This temperature is lower than the original temperature at location -Xl, so in the fourth step the gas
parcel adsorbs an amount of heat, dQcold, from the plate thereby raising its temperature back to its original value,
Tm + xlVT.

The net effect of this process is that the system has completed a cycle which has returned it to its original
state and an amount of heat, dQcold, has been transported un a temperature _tn_adient by work done in the form of
sound. It should be stressed again at this point that no mechanical devices were used to provide the proper phasing
between the mechanical motion and the thermal effects.

If we now consider the full length of the stack as shown in the upper portion of Figure 1, the overall heat
pumping process is analogous to a "bucket brigade" in which each set of gas parcels picks up heat from its neighbor
to the left at a lower temperature and hands off the heat to its neighbor to the right at a higher temperature. Heat

4OO
exchangers m'e placed at the ends of the stack to absorb the useful heat load at the left-hand (cold) end of the stack and
exhaust the heat plus work at the right-hand (hot) end of the stack. The fact that the gas parcels actually move a
distance which has typically been on the order of several millimeters means that good physical contact between the
heat exchangers and the stack is not crucial since the moving gas provides good thermal contact.

Thennmcoustic Energy Tran_sport

If there were no external (/e. usefull) heat load applied to the slack or longitudinal thermal conduction along
the stack or through the gas, then eventually the temperature gradient in the plate would approach that caused by rite
adiabatic processes in the gas. In the absence of gas viscosity, this critical temperature gradient, VTcrit, is a
function only of the gas thermophysical properties, the wavelength of the sound, Z, and the mean position of the
stack, x, within the standing wave field.
VTcrit- 2X (7-1) tan [2K--X-I (3)
7.
The ratio between the tempmature gradient in the stack and the critical gradient, V ffiVTm/VTcrit, plays an important
role in the performance of the stack as will be explained below.

The rate of heat wanspca or heat pumping power, Q2, within the stack can be expressed in a simple form if
we assume that the stack is much shorter than the wavelength of the sound and we again neglect viscosity.

Q2 =
_n4 a pl ul
(r-1) (4)
The first term in (4) is simply one quarter of the thermally effective cross-sectional area of the stack, where II is the
stack surface area per unit length and _1cis the thermal penetration depth. The heat pumping power is also
proportional to the product of the acoustic pressure, Pl, and particle velocity, u1, in the stack. If the stack were
located at a pressure or velocity node, no heat pumping would take place since either the pressure variations which
cause the adiabatic temperature changes would be absent or them would be no motion of the gas parcels. Since
pressure and particle velocity are proportional (the l_Opuftionality constant depends upon the location of the stack
within the standing wave field), a doubling of the acoustic pressure would quadruple the heat transport. This is the
origin of the subscript "2"on the heat transport symbol, Q2, which emphasizes the second order ¢_dance of the
magnitude of the thermoacoustic heat transport on the square of the acoustic field variables.

The final term in equation (4) is a measure of how close the system is to the limiting temperature gradient.
As mentioned befog'e, when F ffi1, the gas parcels "see" their adiabatic temperature span on the stack so no heat
transfer from the gas to the plate takes place. When F ffi0, the temperature of the plate is uniform and a large
quantity of heat is pumped by the oscillating gas parcels.

The stack also absorbs work, W2, at a rate proportional by (T'- 1). The following simple expression for the
work absorbed by the stack of length Ax, in the absence of viscosity, can be written if one assumes that the heat
capacity of the stack is much greater than that of the gas.

W2 = _ Ax (7- I) 2_f(pl)2 (r- I) (5)

4 _n
This work represents the acoustical energy dissipated due to irreversible thermal conduction between the gas and the
plate.

The ratio of the heat pumped, Q2, to the work done, W 2, to pump that heat is defined as the coefficient -of-
performance, COP, of the refrigerator. Since the tempemtm_ spanned by the stack, AT ffiFVTerit Ax, one can show
that the thermoacoustic COP ffiF COPCarnot, where COPCarnot ffiAT/Thot, is the ideal coefficient-of-performance
dictated by the.First and Second Laws of Thmmodynamics. Since for a heat pump, F < 1, we see that the
thermuaconstic COP is always less than that of Carnot. We also see that with the thermoacoustic heat pump, there
will be the same competition between power density and efficiency which exists in all heat engines. As we pointed
out earlier, them is no useful heat pumped when F = 1, which is the point where the efficiency is at its maximum.

401
 The pnerul results derived above from the simple, invisid picture arc essentially preserved when the viscosity
of the gas is included but the detailed mathematical descriptions are substantially more complicated. Discussion of
these; equations is well beyond the scope of this introduction to thennuacousfics. The reader is referred to the
excellont review article by Dr. Swift [7] for a detailed derivation and discussion of this more complete analysis.

THE SPACE THERMOACOUSTIC REFRIGERATOR (STAR)

STAR was the first attempt to exploit the advantages of the thermoacoustic heat pumping cycle for
cryo:ooler applications in space. It is intended to operate autonomously in low Earth orbit aboard the Space Shuttle
in a Get Away Special (GAS) canister. It derives its power from an internal battery power sutwce (700 Watt-horn's)
and was optimized for a modest temperature span (AT < 80 "K) and small heat loads (Qcold < 5 Watts). Due to
reqtg_ments of small size and light weight imposed by the GAS envelope, it operates as a frequency of about 400
Hzandis drivenbyanelectrodynamicIondspeakez
8. Itsfrequency
of operationis adjustedautomaecally9
to keepthe
system on resonance. The resonator length is approximately equal to a quarter wavelength of sound in a mixture of
heliumandargon or helium and xenon gasl0 maintained at a mean pressure of ten atmospheres (150 psia). It has a
single stack with a uniform spacing made of polyester film (Mylar TM) and fishing liuel0, I I which is spiral wound
like a "jelly roll". The stack used in STAR is 7.9 cm in length and 3.8 cm in diameter. Copper-strip, parallel plate
heat exchangers are located at either end of the stack. Figures 3 are scale drawings which show the details of the
acousticalsub-systems.

Aconstlcal Sub-Systems.

The driver housing does more than simply hold the electrodynamic driver. The considerable mass and size of
the housing is a consequence of the fact that a commerciaily available loudspeaker was modified for this application.
The ,driverhousing serves as a heat sink for the heat generated by the resistive losses in the driver voice coil and the
heat pumped away from the cold end of the resonator. It also contains the ten atmospheres of the helium/xenon gas
mixtm'¢ which is the "working fluid" within the refrigerator. The housing is bolted to the standard 12 inch bolt
circle provided on the GAS cannister lid which acts as the radiator while on orbiL

The driver voice coil is attached to an aluminum reducer cone which is in Van attached to a nickel
electmfcmned bellows. The bellows provides a means of transferring acoustic pressure to the resonator without the
need for sliding seals. A miniature accelerometer is attacheA to the surface of the reducer cone opposite the bellows
to monitor the displacement magnitude and its phase relative to the w,oustic pressure at the face of the bellows. That
acoustic pressure is monitored by a piezoelectric quartz microphone 12 followed by a MOSFET impedance converter
located within the driver housing in close proximity to the microphone.

The resonator is a modified quarterwavelength tube. The open end is terminated by a tapered "trumpet"and
sealed by a suaounding sphere. Thus, an "open" termination is simulated while still allowing the resonator to retain
the ten aUnosphere gas mixtm'e. The thennoncoustic slack and heat exchangers me located in a section of the
rescmtor designed to allow a minimum of heat conduction back to the cold end. The resonator is wrapped with
multiple layers of superinsulation to prevent heating by thermal radiation. An electrical strip heated element is
attached to the cold end of the resonator near the cold heat exchanger to permit measurement of refrigerator
pert_mnauce with a variable and quantif'mbleheat load. A thermal isolation vacuum chamber surrounds the resonator
and _w,als against the bottom surface of the driver housing with an O-ring.

Elec_cal Sub-Systems.

In order for the refrigerator to operate autonomously in space, a family of analog and digital electronic
circvits are employed to monitor the "health" of the system, keep the driver running at the proper amplitude and
f_ltg_3', and to acquire and store useful data for post-flight analysis. The design and function of each of these sub-
systuns is documented in Reference [9].

4O2
 t"tOUS | 17 0

Bel lows

FIoL I-te e I-_

E× Ctn,=Jrlger-

SLacl_.

Co[el I-teeL
E-<ctnarTg_r

Re s on,_ {_ ol-

Figure 3. Scale drawing of the STAR acoustical sub-systems showning the driver, resonator, stack, and he,at exchangers.
Shown to the right of the sphere is the official logo for the experiment which will be flown on Space Shuttle mission
STS-42 scheduled for launch on 22 January 1992.

403
 1.00 °, .... ,°.°.. ............ ; ..... .°°.,..; ............. • ........ i,.°..,.,,....,:°.

Tc_rh = 0.720 +.: 0.0393 * (_

0.95

0.90

0.85

0.80

0.75

0.70 l i • • . t . i I . t t i l J i a i , . _ J _ I • J ,, • I

0 1 2 3 4 5 6
Q (load, Watts)

Figure 4. The ratio of the cold and hot heat exchanger temperatures, Tc/T h, as a function of heat load on the cold heat
exchanger at an acoustic pressure ratio of pl/p m = 2% in a lean mixture (4% Xe) of helium and xenon.

0.20 ........... _ ............ _ ............. , ............. °°°. ......... ° ............. °

• ° ° ° ° •

0.18

0.16

0.14

0.12

¢_ 0.10
0

0.08

0.06

0.04

0.02
° °

0.00 .... i .... ; .... I,.,.; .... i .... i

0 1 2 3 4 5 6
q (load, Watts)

Figure $. Measured coefficient-of-performance relative to the ideal Carnot coefficient-of-performance (COPR) as a

function of heat load for STAR operating at an acoustic pressure ratio of Pl/Pm = 2% in a lean mixture (4% Xe) of helium
and xenon.

4O4
 The circuit which is unique to this application is the Resonance Control Board (RCB). Its function is to
maintain the system at acoustical resonance and to control the amplitude of the acoustical wessure at pre-determiued
levels dictated by the Controller Board. As the temperatuxe changes, the sound speed changes, hence, for fixed
resonator dimensions, the resonance frequency will be a function of temperature.

The RCB maintains of the resonance condition by comparing the relative phase of the microphone and
accelerometer outputs. At resonance the pressure and velocity of the gas mixture at the bellows location must be in-
phase. Therefore, the pressure and acceleration are required to be in quadrature (ie. 90" out-of-phase). The two
signals me electronically multiplied together and the dc-componcnt of their product, which is proportional to the
cosine of their phase difference, is used as an error voltage. This error voltage is electronically integrated and fed
back to a voltage controlled oscillator to close the phase accurate phase-locked-lonp circuit which maintains the
entire system at resonance.

Refrigerator Performance.

The performance of STAR can be summarized by the two graphs which are presented here as Figures 4 and 5.
Other measurements of the performance of the entire system, including the electrodynamic driver axe included in
Reference [I 1]. The STAR performance shown in Figures 4 and 5 is by no means the best which has been achieved
in a thermoacoustic refrigerator of this style. Other improvements to the design, such as the use of a stack which
has non-uniform spacing, has produced a single-stage, no-load temperature span of 118 "K even without the use of
gas mixtures to reduce viscous losses or improve the coefficient of performance. COPR's as large as 20% have also
been measured in a similar refrigerator over the same temperature span 10.

lllGH POWER TllERMOACOUSTIC REFRIGERATORS

The STAR is the first in a series of space cryocoolers now under development at the Naval Postgraduate
School. Those applications generally require relatively little heat pumping power (a few watts) but a very large
temperature span (AT = 100 to 200 "K). The requirements of residential refrigerator/freezers and air conditioners are
opposite of those for spacecraft cryocoolers. Those application require modest temperature spans (AT- 25 - 45 "K),
but much higher heat pumping powers on the order of hundreds of watts for refrigerators and thousands of watts for
air conditioners. They also axe typically powered by 110 volt alternating current rather than 28 volt direct current.
There are several design modifications which are therefore required to adapt this established space cryocc_ler
technology to residential applications, but these modifications do not present any substantial technological barriers.

Figure 6. Schematic diagram of a half-ton (2 kW) capacity thermoacoustic chiller. The half-wavelength resonator
operates at 60 Hz, is powered by a single double-acting electrodynamic driver and contains two "stacks".

4O5
 Operation at fixed frequency (60 Hz or 120 Hz) can be accomplished by selective absorption of one
component of the gas mixune as the temperatme of the working fluid changes. Half-wavelength r'_tonators,such as
_ one shown in Figure 6, can utilize both oscillating faces of a single driver which can be more massive and hence
n_3reefficient because it operates at fixed frequency. The presence of two stacks reduces the required heat pumping
capacity of each individual stack although the diameters of the stacks would necessarily be larger. Th6 design shown
in Figure 7 has stack diameters of 18 cm, operates at 60 Hz, is about 90 cm wide, 60 cm tall, and 20 cm thick. It
should be capable of pumping about 2 kW of heat across a 30 "Ctemperature span with a COP - 3-4 (about 30% of
Cm_ot perfmmance) including electroacoustic conversion efficiencies. At the present time, the greatest uncertainty
in the use of thermoacoustic engines in high power cooling applications is the design of the heat exchangers.

ACKNOWLEDGEMENTS

The authors are indebted to Dr. Greg W. Swift, Los Alamos National Labs, for several conversations
regarding thisnew technology andhisexcellent reviewarticle on thermoacoustic engineswhosecontents was
exploited extensively in the first part of this paper7. The development of the Space Thermoacoustic Refrigerator was
the work of several staff and students at the Naval Postgraduate School. Staff members include David Rignmiden
(Payload Manager), Jay Adeff (Physicist), Glenn Harrell (Machinist), Jim Homing and Ron Phelps (Programmers)
and CDR (Dr.) David Gardner, NOAh, Corp. (Group Leader). The students who worked on the Ixuject as part of
tl_dr Master's degrees were LT Michelle Fitzpatrick, USCG (Driver), LT. Michael Susalla, USN (Thermodynamics
and Gas Mixtures), CAPT David Harris, Canadian Forces (Driver), LT Richard Volkerk USN (Driver), and MAJ
Ronald Bymes, USA (Electronics). The loudspeakers used as the drivers were donated by Harmon-JBL and modified
by their engineer, Fancher Murray. This work is supported by the the Office of Naval Re.search (Physics Division),
the Offme of Naval Technology and the Naval Postgraduate School Direct Funded Research Program. Substantial
ftu_din8 and technical support for STAR was provided by the Naval Research LaboratcaT - Spacecraft Engineering
Division. The payload launch services were provided by the Air Force Space Test Program.

REFERENCES

1. S. Derra, "CFC's: No Easy Solutions," Res. & Dev. 32(5), 56-66 (1990)
2. R. Monastersky, "Decline of the CFC Empire," Sci. News 133, 234-236 (1988).
3. A.K. Naj, "CFC Substitute," Wall St. Journal, (2 July 1991).
4. J.C. Wheatley, T. Hofler, G. W. Swift, and A. Migliori, "Experiments with an Intrinsically Irreversible
Acoustic Heat Engine," Phys. Rev. Lett. 50,499 (1983); "Intrinsically irreversible heat engine," U. S.
Patent No. 4,489,553 (Dec. 25, 1984).
5. J.W. Strutt (Lord Rayleigh), The Theory of Sound, 21_ ed., Vol. II (Dover, 1945), $322j.
6. T. Yazaki, A. Tominaga, and Y. Narahara, "Experiments on thermally driven acoustic oscillations of gaseous
helium," J. Low Temp. Phys, 41, 45 (1980).
7. G.W. Swift, "Thermoacoustic Engines," J. Acoust. Soc. Am. 84(4), 1145-1180 (1988).
8. D.A. Harris and R. E. Volken, "Design and Calibration of an Electrodynamic Driver for the Space
Thermeacoustic Refrigeratv¢", Master's Thesis, Naval Postgraduate School, Monterey, CA (Dec. 1989);
DTIC Report No. AD A212 022.
9. R.B. Bymes, Jr., "Electronics for Autonomous Measureanent and Control of a Thennoacousfic Refrigerator
in a Space Environment", Master's Thesis, Naval Postgraduate School, Monterey, CA (Dec. 1989); DTIC
Report No. AD BI41 388.
I0. M.P. Snsalla, "Thermodynamic Improvements for the Space Thermoacoustic Refrigerator", Master's Thesis,
Naval Postgraduate School, Monterey, CA (June, 1988); DT[C Report No. AD A196 958.
11. J.A. Adeff, "Measurement of the Space Thermcecoustic Refrigerator Pexfcmnance", Master's Thesis, Naval
Postgraduate School, Monterey, CA (Sept. 1990).
12. T. Hofler, "Accurate Acoustic Power Measurements with a High-Intensity Driver," J. Acoust. Soc. Am. g3,
777 (1988).

406