Back to Home Page of CD3WD Project or Back to list of CD3WD Publications

Home - English - French - German - Italian - Portuguese - Spanish
                         TP# 20: 02/85
                       SOLAR WATER PUMPS
                           C. J. Swet
                      Technical Reviewers:
                         Paul E. Dorvel
                         John D. Furber
                         Daniel Ingold
                         Published by:
               1600 Wilson Boulevard, Suite 500
                 Arlington, Virginia 22209 USA
            Tel:   703/276-1800 * Fax:   703/243-1865
ISBN #0-86619-220-4
This paper is one of a series published by Volunteers in Technical
Assistance to provide an introduction to specific state-of-the-art
technologies of interest to people in developing countries.
The papers are intended to be used as guidelines to help
people choose technologies that are suitable to their situations.
They are not intended to provide construction or implementation
details.  People are urged to contact VITA or a similar organization
for further information and technical assistance if they
find that a particular technology seems to meet their needs.
The papers in the series were written, reviewed, and illustrated
almost entirely by VITA Volunteer technical experts on a purely
voluntary basis.  Some 500 volunteers were involved in the production
of the first 100 titles issued, contributing approximately
5,000 hours of their time.  VITA staff included Maria Giannuzzi
and Leslie Gottschalk as editors, Julie Berman handling typesetting
and layout, and Margaret Crouch as project manager.
C.J. Swet, the author of this paper, has a background in engineering,
and is a consultant in solar and other "gentle" technologies,
with special emphasis on energy storage.   He has 20 years
experience in the field of solar energy, and has consulted on
solar energy and other appropriate technology projects in developing
countries.  He has published several papers on solar energy
and other energy related topics.   Reviewers Paul E. Dorvel, John
D. Furber, and Daniel Ingold are also experts in the field of
solar energy.  Paul E. Dorvel is currently Associate Principal
Engineer in the Power Systems Division of the International
Engineering Company.  He has over seven years experience in Africa
doing market research and field engineering for solar micropump
irrigation systems.  John D. Furber is President of Pleasant
Valley Software Corporation and Starlight Energy Technology.  He
frequently lectures and consults overseas on solar energy technologies.
Daniel Ingold is a biophysicist by training and a
research engineer at Appropriate Technology Corporation.
VITA is a private, nonprofit organization that supports people
working on technical problems in developing countries.   VITA offers
information and assistance aimed at helping individuals and
groups to select and implement technologies appropriate to their
situations.  VITA maintains an international Inquiry Service, a
specialized documentation center, and a computerized roster of
volunteer technical consultants; manages long-term field projects;
and publishes a variety of technical manuals and papers.
For more information about VITA services in general, or the
technology presented in this paper, contact VITA at 1815 North
Lynn Street, Suite 200, Arlington, Virginia 22209 USA.
                By VITA Volunteer C.J. Swet
This paper examines water pumping systems that use solar radiation
as a direct source of energy.   We look primarily at small-scale
rural applications in the Third World, where the potential
benefits are greatest and the near-term economics seem most
favorable.  Two generic technical approaches for solar water
pumping systems will be examined:   (1) thermodynamic (in which
the radiant energy is first converted to heat); and (2) photovoltaic
(in which it is first converted to electricity).   Since
photovoltaic technology is more mature, it is used for economic
comparisons with other methods of pumping water.   Our treatment
of this complex subject is necessarily cursory; the aim is to
provide prospective users with sufficient insight to determine
whether solar water pumping is a plausible option for their
specific situation, and to furnish a guide for further investigation.(*)
The history of thermodynamic solar water pumping technology goes
back nearly four hundred years, when Solomon deCaux in France
raised water for a fountain by the expansion of solar-heated air.
By the beginning of this century, many of the currently competing
design concepts had already been explored, and several promising
attempts at commercialization were under way when activity subsided
due to the advent of the internal combustion engine and
cheap commercial fuels.  Much of this development had been aimed
at relatively large-scale applications.   It was not until the
resurgence of interest in solar energy caused by the 1973-1974
oil embargo that greater attention began to be directed at small-scale
rural applications in developing countries.
(*) Of particular interest to the serious reader in this field is
the the definitive and comprehensive work done in conjunction
with the United Nations Development Programme/World Bank Small-Scale
Solar-powered Irrigation Pumping Systems Project by Sir
William Halcrow & Partners and the Intermediate Technology Development
Group, Ltd.  The most important reference for the prospective
solar pump user is Handbook on Solar Water Pumping (see
In contrast, the development of photovoltaic water pumps has
heavily emphasized small-scale rural applications since the
1960s.  While both of these technical approaches continue to
mature, neither can yet be judged inherently superior.   Most of
the recent efforts, however, have concentrated on photovoltaic
systems and a majority of solar pump field operating experience
has been with photovoltaic installations.
Solar water pumps may be used for irrigation, community water
supply, livestock watering, and in various industrial processes.
In principle they may be used virtually anywhere, but the most
compelling needs and opportunities are found in the fuel-poor but
sun-rich rural areas of the Third World.   Solar pumps may also be
almost any size, but most small farms, villages, and animal herds
in developing countries require hydraulic output power of less
than a kilowatt.  Many of these potential users are too far from
an electrical grid to economically tap that source of power, and
engine-driven pumping tends to be prohibitevely expensive as well
as unreliable due to the high cost of purchased fuel and insufficient
maintenance and repair capabilities.
Developing countries increasingly require less costly and more
reliable methods of pumping water that do not rely on commercial
energy supplies.  Solar water pumping is one of several potentially
satisfactory alternatives, which also include wind, water,
biomass, animal, and human power.
Greater use of irrigation is clearly needed to increase the yield
from existing cultivated land and to permit cultivation of presently
marginal or unusable land.  Nearly all this additional
irrigation will have to use pumped water, since most of the
available sources of gravity-fed water are already fully exploited.
In Third World countries, most irrigated land is in family
plots of less than four hectares, a large proportion of these
being under one hectare.  This practice can be expected to extend
to lands not yet under cultivation, since small plots have been
found to be more productive than large farming units in terms of
yield per hectare although more demanding in terms of labor
Daily demand for pumped irrigation water varies widely depending
on the season, crop, stage of growth, region, method of water
distribution, and water management effectiveness, with maximum
values ranging from about 20 to 120 cubic meters per hectare per
day.  Water sources include ground water from open (dug) wells or
boreholes that are surface water from rivers, ponds, or canals,
and typically at depths of two to 10 meters below ground level.
Irrigation water is seldom extracted from depths greater than 10
meters because the value of its benefits is seldom high enough to
justify the extra cost of deeper wells and additional pumping
energy.  To be economically feasible for agricultural applications,
the cost of water delivered must be less than the value of
the benefits obtained through use of the irrigation water, either
through improved yields or by enabling more crops to be grown per
year.  In 1982 a global norm for the cost ceiling for water delivered
to the field (not to the crop) was approximately US$0.06
per cubic meter, although clearly the actual figure in a particular
situation will depend on the crops grown, the field application
efficiency, and market prices.   If 60 percent of the pumped
water is used by the crop itself (a fairly typical condition for
earth furrows), the cost ceiling for that water would be US$0.10
per cubic meter.
Most villages in developing countries have fewer than 1,500 inhabitants,
and in many of those villages the per capita water
consumption is far less than the 40 or more liters per day deemed
desirable from a health standpoint.   A common reason for this low
consumption is that all the water is drawn from a single well,
resulting in lines, and in larger villages, the need to carry
water considerable distances.   Although multiple dispersed wells
will alleviate these problems, polluted sources become more difficult
to avoid.  Ground water is usually extracted from depths
of 30 meters or more because its value for human consumption is
much greater than that for irrigation; water sellers in developing
countries often command a price equivalent to more than
US$3.00 per cubic meter for 10 to 30 liters per day.   For this
application, the economic feasibility of solar water pumping is
much less a factor than its competitive position relative to
other methods of mechanized pumping.
For livestock-watering in remote areas, daily demand per head
varies widely depending on breed and type of forage; about 40
liters is fairly representative for dairy cattle.   In recent
years many engine-driven borehole pumps have been installed for
this purpose, pumping from depths as low as 30 meters.   Thus, a
herd of cattle and a village of the same population may have
comparable pumping power requirements.   However, for this application
it is often desirable to have multiple dispersed pumps in
order to minimize overgrazing near each watering place.   These
considerations accent the need for pumps that can operate reliably
when unattended for long periods of time, and that do not
require secure fuel stores at each watering place.
Although thermodynamic and photovoltaic solar water pumping
systems are conceptually similar in that both are powered
directly by solar radiation, their operating principles are quite
different.  The following discussion highlights the distinctive
features of these systems.(*)
All thermodynamic systems use a solar collector to convert solar
radiation to heat and a heat engine to convert the heat to mechanical
power for pumping.  In heat engines a fluid or gas absorbs
heat at a higher temperature, which causes it to expand; it
then contracts upon removal of the heat at a lower temperature.
This expansion and contraction is harnessed to move a reciprocating
piston in a cylinder, or it can expand against a turbine
Figure 1 illustrates the basic energy flows, showing qualitatively

31p05.gif (600x600)

the necessary temperature differential across the heat engine
and the unavoidable losses associated with each stage of the
process.  Of the two indicated temperatures, the lower one cannot
be lower than that of the pumped water to which the unusable
degraded heat is typically rejected, while the upper one is
largely controlled by the type of collector.   Increasing the
upper temperature (within practical limits) raises the overall
system efficiency and reduces the required collector size, but
usually at the cost of greater complexity or more expensive
materials.  This broadly characterized conceptual approach can
have many different embodiments, with various types and combinations
of collectors, working fluids, heat engine cycles, engines,
and pumps, as discussed in Section III.
These systems exploit the photovoltaic effect to convert solar
radiation to direct current electricity, which powers a motor-driven
pump.  A basic photovoltaic system layout is shown in
Figure 2.  Photovoltaic conversion occurs when light falls upon a

31p06.gif (600x600)

thin, flat material called a solar cell.   One side of the cell
becomes electrically positive, and the other electrically negative.
This is a solid-state, electronic effect.   Like a transistor,
the solar cell has no moving parts except electrons.
(*) Detailed explanations of these operating principles can be found
in the publications listed in the bibliography.
As long as light falls on the cell, the electrons flow as an
electrical current through an external circuit containing the
motor.  Individual solar cells are connected in series strings to
obtain the desired output voltage.   Series strings can be connected
in parallel to obtain the desired output current of a
module.  Several modules are then interconnected and mounted.
Photovoltaic array output current and power--to the extent voltage
is constant--vary linearly with solar irradiance.   Efficiency
and power output decrease with increasing cell temperature on the
order of 0.5 percent per [degrees] C above 28 [degrees] C.
Figure 3 shows the performance characteristics of the components

31p08.gif (600x600)

of a typical photovoltaic pumping system, illustrating the importance
of proper matching of the electical source and the hydraulic
load over a range of operating conditions.   Some optional
components and configurations of these systems are discussed in
Section III.
Most of the small-scale systems that have been developed beyond
the prototype stage use Rankine cycles similar to the one shown
schematically in Figure 4, with organic working fluids such as

31p09.gif (600x600)

Freon 11 and slow-speed reciprocating engines that directly drive
piston pumps.  Many developing regions are familiar with Rankine
systems because of experience with steam engines.   Organic working
fluids can produce higher heat-to-work conversion efficiencies
than steam at temperatures up to their stability limit of
about 150 [degrees] C, but extreme care must be taken to ensure zero leakage
since very small amounts incapacitate the system and recharging
is difficult in the field.  A reciprocating engine is virtually
the only choice,  since turbines and rotary expanders are
excessively expensive in the small sizes of interest.   Slow-speed
reciprocating (piston) pumps tend to be more efficient than
conventional high-speed centrifugal pumps at heads greater than
about 10 meters, although single stage centrifugal pumps (which
are easy to make) are well suited for very low-head irrigation.
The system depicted in Figure 4, which was designed by a Finnish
company, has a trough-type concentrating solar collector that
follows the sun by rotating about a horizontal north-south axis.
Sun following is automatic, powered by the shifting weight of
solar-heated Freon and controlled by a sun shade mounted on the
collector, but the orientation must be reset manually each day.
The collector has a 12 square meter aperture, occupies 16.5
square meters of ground space, and weighs 170 kilograms.  All of
the other above-ground components occupy a 0.4 cubic meter volume
and weigh about 50 kilograms.   Aluminum is used extensively for
both weight reduction and corrosion resistance.   At a Freon
temperature of 107 [degrees] C the reported delivery rate is about two
liters per second (roughly equivalent to 40 cubic meters per day
if the pump operates eight hours) against a total pumping head of
14 meters.  The reported rate is nearly five liters per second
(100 cubic meters per day) against a head of three meters.   The
above-ground pump location limits the use of this system to
shallow wells or surface water sources involving suction lifts no
greater than about eight meters.
A somewhat similar system from West Germany has about 40 square
meters of stationary flat plate collectors that can heat the
Freon 11 working fluid to about 90 [degrees] C.   Its pump can be located
below grade and is adaptable to wells up to 60 meters deep.
Preliminary testing in India indicates a delivery rate of 40
cubic meters per day against total pumping heads of 15 to 20
meters.  The pump is sized to permit greater output when larger
collectors are used.
Water (steam) has some important advantages as a Rankine cycle
working fluid.  It can be used at higher temperatures than are
possible with organic fluids, to achieve higher efficiencies.
Also, the consequences of leakage are far less severe.   An Indian
firm has developed a 2-kilowatt uniflow reciprocating steam
engine powered by a glass strip reflector trough collector.
However, higher temperatures require greater optical and tracking
precision, which increases the cost per unit collector area and
tends to offset the size reduction made possible by improved
efficiency.  The economic competitiveness of high-temperature
Rankine cycle solar pumps is still a under contention.
Stirling cycle heat engines offer perhaps the most promising
means of exploiting the very high temperatures (over 500 [degrees] C) that
can be obtained with point focusing collectors, such as parabolic
dish reflectors.  Sunpower Inc. in the United States has developed
a free piston Stirling engine with an integral diaphram pump,
using helium as the working fluid.   In tests by the manufacturer
with a simulated solar thermal input of 1 kilowatt (corresponding
to the output of a dish approximately 1.4 meters in diameter),
the Stirling engine delivered 2 liters per second at 560 [degrees] C
against a head of four meters.   At its present stage of development,
however, it is easily damaged, and test results have been
disappointing.  Another promising Stirling engine pump is the
"Fluidyne" liquid piston system being developed by another Indian
company, but no solar version has yet been demonstrated.
Many other technically intriguing and potentially useful systems
have been or are being developed, including:
     1.   smaller organic Rankine systems;
     2.   very small (about 25 watts) steam Rankine systems;
     3.   an organic vapor liquid piston pump;
     4.   a heated air liquid piston pump;
     5.   a fluid overbalancing rocking beam engine pump; and
     6.   various solid state systems based on "memory metals,"
         polymers, and the differential expansion of bimetal
Some of these systems have become commercially available, but it
must be emphasized that none of them (or of the other concepts
described above) is known to have successfully undergone the
extensive testing under field conditions that characterize a
mature product.
Ability to manufacture and repair such technology often depends
on the region.  The several systems presently being developed in
India presumably would be manufactured there, and an Indian
affiliate of the West German company is evaluating the possibility
of locally producing all or part of the German system.   This
does not mean, though, that these systems of Indian design could
or would be manufactured elsewhere in the Third World.   Supporting
frames, conventional heat exchangers, and some types of
collectors could be made and repaired in many developing countries,
but reciprocating engines and piston pumps of high efficiency
call for close tolerances that may not be readily achievable
with available skills and equipment.
A number of types and sizes of photovoltaic systems are available
commercially, in various stages of product development, that meet
the range of pumping needs outlined in Section I.   The significant
design variations of these systems are fewer and more easily
presented than those of the relatively immature thermodynamic
approach.  These variations center mainly on:
     1.   the choice of solar cell material;
     2.   the choice between stationary and sun-following solar
     3.   the choice between planar and concentrating solar
     4.  the type of electric motor;
     5.   the type of pump; and
     6.   the method of source/load matching.
All commercially available systems use crystalline silicon solar
cells, of either the single crystal or polycrystal type.  Other
types of solar cells, which may be less expensive, are under
development.  These use thin films of semiconductor materials,
such as amorphous silicon or cadmium sulfide.   Currently available
solar arrays produce roughly 100 watts per square meter under the
most favorable conditions.  Specific pumping needs do not influence
the choice among these competing designs.
In most systems the solar arrays have a fixed orientation; they
are tilted permanently toward the equator at an angle that maximizes
energy collection during the season of maximum demand (or
for the year if demand is fairly constant).   This is the simplest
and most affordable configuration, but not necessarily the least-expensive
in terms of cost per unit of delivered water.   At
slightly greater first cost and complexity, the orientation may
be manually adjusted several times during the day, thereby increasing
the daily output by up to 30 percent.   This tends to be
cost effective provided that manual labor is available and is
inexpensive for highly seasonal irrigation applications.  If the
system is used over most of the year, a fully automatic tracking
device may be justified.  Although such systems have not yet
demonstrated sufficient reliability under field conditions, some
recent field operating experience with gravity-driven Freon
trackers on pumping systems has been encouraging.(*)
As long as solar cells remain the dominant cost item there is an
incentive to reduce the required area not only through sun following
but also through concentrating the intercepted solar
radiation.  The cost decrease due to further cell area reduction
tends to be offset by the added cost of concentrating optics and
the need for better cooling of the cells and more precise tracking.
If solar cell prices diminish as predicted, the incentive
will become much less compelling.
Permanent magnet direct current motors are the most commonly used
pump drivers for small-scale systems.   Alternating current motors
cost less but are much less efficient in the sizes of interest.
Linear actuators have been used to drive piston pumps, but the
concept requires considerably more development.   Many of the
direct current motors in current use are of the conventional
brush type, which is efficient but poorly suited for submerged
operation and needs brush replacement after every few thousand
(*) See for example Dankoff, W., "Pumping Water,"  Solar Age, February
1984, pp. 29-35.
hours of use.  Electronically-commutated brushless direct current
motors are finding favor because they require less maintenance
and are more readily adapted to submerged operation, although
they are slightly less efficient.
Single-stage centrifugal pumps are frequently used when the total
pumping head is less than 10 meters, and are either self-priming
or (if the suction lift is too great) submerged.   With open wells
or surface water sources, these pumps and the motors can float,
thereby minimizing the suction lift.   For higher heads, either
multistage centrifugal or positive displacement (piston or progressive
cavity) types are most efficient.   If the pump is above
ground or floating, it usually is closely coupled to the motor;
if submerged, the pump may either be closely coupled to a submersible
motor or driven by a vertical shaft.   Positive displacement
pumps ordinarily are submerged except in cases where the lift is
small but the total pumping head is high.
Single-stage centrifugal pumps can be made with head-capacity
characteristics that fairly closely the solar array current-voltage
characteristics, so that the array can operate at near-peak
efficiency over a wide range of operating conditions.   This
matching cannot take place with multistage centrifugal or positive
displacement pumps.  For systems that are not inherently
compatible in this respect, it is possible to install an electronic
impedance matching device between the array and the motor
that will automatically optimize the load on the array.   These
devices, called maximum power point trackers or maximum power
controllers (MPCs), will increase daily pumped output and will
allow pumping to start under low moring irradiance.   Maximum power
controllers add to the complexity and cost of a system, in addition
to creating an approximate five percent power drain on the
array.  Indications are that MPCs are most cost-effective in
systems over about one kilowatt peak capacity.   Below this level,
it may be more cost-effective to substitute extra array capacity
for an MPC.
Local production of nearly all components except the solar cells
appears possible in many developing countries.(*) India and Brazil
have begun cell manufacture and some other countries are considering
the assembly of modules from imported cells.   Since solar
cell technology is advancing so rapidly, and crucial choices
among the candidate semiconducting materials have yet to be made,
(*) For an in-depth discussion of the potential for local production,
see Small-Scale Solar-Powered Pumping Systems:   The Technology
Its Economics and Advancement, by William Halcrow and
Partners, and Intermediate Technology Power, Ltd., and its supporting
documents concerning manufacture of solar water pumps in
the less developed countries (June 1983).
it may be prudent for most of the Third World to await the
emergence of clearly superior design approaches before investing
in production facilities.  Other than solar cells, the most
doubtful items for potential local manufacture appear to be
piston pumps, because of their close machining tolerances.
To gain widespread acceptance, small-scale water pumps must not
only deliver water at a cost below the value of that water; they
must also do so at a lower cost than alternative methods of
pumping water.  Most prominent among the currently available
alternatives are diesel or kerosene engines, wind power, animals,
and humans.  A good basis for comparing them with solar is the
cost per unit volume of water delivered under like conditions
over a like number of years.   This takes into account costs of
purchasing, financing, delivery, installation and start-up, fuel,
operating and maintenance labor, repairs, and replacements.  In
United Nations Development Programme studies, comparative costs
of delivered water have been estimated for irrigation, village
water supply, and livestock watering in Kenya, Bangladesh, and
Thailand (see bibliography).
Based on 1982 prices, some typical results are shown in Figures 5, 6, 7, and 8.

31p150.gif (600x600)

There are many qualifications to these results,
too numerous and complex for exposition here.   It must be pointed
out, though, that the solar costs are based on photovoltaic
systems in which the dominant cost item is the solar module at
its 1982 price of about US$8/ peak watt.   That price is likely to
become much lower within a few years, making the solar option
more competitive.  It should also be noted that the attractively
low wind power costs are based on average mean wind velocities
for each country; within those countries there are regions with
totally inadequate wind.  The costs for animal power are somewhat
optimistic because they do not include the cost of diverting
animals from other activities, and the "low case" diesel costs
are based on unrealistically optimistic values for fuel cost and
consumption.  With these points in mind it seems fairly clear
that within a few years solar will be quite competitive in most
sunny regions that have little wind.
                          GLOSSARY OF TERMS
Aperature.  The solar collection area.
Dish collector.  One in which the reflecting surface is a paraboloid
         of revolution that concentrates direct solar radiation
         onto an absorber at its focal point.  Usually for
         temperatures above 250 [degrees] C, with two-axis tracking.
Drawdown.  The distance the water level in a well is temporarily
         lowered by pumping.
Flat plate solar collector.  One in which the aperture is essentially
         identical to the area of the absorber surface, the
         absorbing surface is essentially planar, and no concentration
         is employed.  Usually for temperatures below 100 [degrees] C.
Hydraulic output power.  The power imparted by the pump to the
         water, proportional to the product of the flow rate and the
         total pumping head.  In watts, roughly equal to liters per
         second times meters times ten.
Irradiance (radiation intensity).   The energy flux density in the
         solar radiation, usually expressed in watts per square meter.
Static head.  The vertical distance between the water source
         level at no flow and the point of discharge.
Suction lift.  The height that water must be lifted from the
         source level to the pump.
Total pumping head.  The static head plus drawdown and flow pressure
          losses in piping.
Trough collector.  One with a cylindrical parabolic reflecting
         surface that concentrates direct solar radiation onto an
         absorber (usually a tube) at its focal line.  Usually for
         temperatures from 100 [degrees] to 250 [degrees] C, tracking about one axis.
Overall system efficiency.  The fraction of intercepted solar radiation
          that imparts pumping energy to the water, i.e., pump
          hydraulic output power per unit aperture/irradiance.
Halcrow, William and Partners, and Intermediate Technology Power,
      Ltd.   Small-Scale Solar-Powered Pumping Systems:  The Technology,
      Its Economics and Advancement (United Nations Development
      Programme Project GLO/80/003).  Washington, D.C.:  World
      Bank, June 1983.  Available through the World Bank, along
      with the following supporting documents:
      1.   Performance tests on improved photovoltaic pumping systems
      2.   Economic evaluation of solar water pumps
      3.   Potential for improvement of photovoltaic pumping systems
      4.   Review of solar thermodynamic pumping systems
      5.   Manufacture of solar water pumps in developing countries
Small-Scale Solar-Powered Irrigation Pumping Systems
      (United Nations Development Programme Project GLO/78/004,
      Phase I report).  Washington, D.C.:   World Bank, July 1981.
      See also Small-Scale Solar-Powered Irrigation Pumping System
      Technical and Economic Review (September 1981), amplifying
      this report.
Handbook on Solar Water Pumping (United Nations Development
      Programme Project GLO/80/003).  Washington, D.C.:
      World Bank, February 1984.  This handbook directly addresses
      the concrete issues and methods of selecting, evaluating,
      and specifying a solar water pumping system.
Kreider, J., and Kreith, F., eds.   Solar Energy Handbook.  New York:
      McGraw Hill, 1981.  The reader is referred to the following
      Chapter 1 for history of solar thermodynamic water pumping
      Chapter 7 for non-concentrating solar-thermal collectors
      Chapter 8 for intermediate concentration collectors
      Chapter 9 for high concentration collectors
      Chapter 22 for solar powered heat engines
      Chapter 24 for photovoltaics
McNelis, ed.  United  Kingdom Section of International Solar Energy
      Society.   Proceedings of Conference on Solar Energy for Developing
      Countries, Volume on Refrigeration and Water Pumping.
      London:   ISES, January 1982.
United Nations Development Programme; World Bank; and Philippines
      Ministry of Energy.  Proceedings of Workshops on Solar Pumping
      in Developing Countries.  Washington, D.C.:   World Bank, June
THERMODYNAMIC SYSTEMS (not necessarily mature products):
Dornier                            organic Rankine flat plate,
Postfach 1360                      approx. 500 watts output
7990 Friedrichshafen 1
Grinakers                          fluid overbalancing beam engine,
c/o A. de Beer                     flat plate, approx. 200 watts
P.O. Box 349
Rosslyn 0200
Grinakers                          fluid overbalancing beam engine,
c/o Pelegano Village Industries    flat plate, approx. 200 watts
P.O. Box  464
Wrede-Ky                           organic Rankine trough
P.O. Box 42                        approx. 300 watts output
SF-02701 Kaunianen
PHOTOVOLTAIC SYSTEMS (commercially available and fairly mature):
AEG--Telefunken Raumfahrttechnik und Neue Technologien
Industriestrasse 29
2000 Wedel, Holstein
Aerimpianti S.p.A.
Via Bergano, 21
20135 Milano
ARCO Solar, Inc.
20554 Plummer Street
Chatsworth, California 91311 USA
A.Y. McDonald Corp.
P.O. Box 508
Dubuque, Iowa 52001 USA
133 Enterprise St.
Evansville, Wisconsin 53536 USA
BP 43
37009 Tours
Grundfos Pump Corp.
2555 Clovis Ave.
Clovis, California 93612 USA
Caixa Postal 8085
Sao Paulo 01000
Intersol Power
11901 West Cedar Avenue
Lakewood, Colorado 80228 USA
Jacuzzi Brothers
11511 New Benton Hwy.
Little Rock, Arkansas 72201 USA
Mobil Solar Energy Corp.
16 Hickory Dr.
Waltham, Massachusetts 02254 USA
Pompes Gitnard Etablissements
179, Boulevard Saint Denis
92402 Courbevoie
Philips GmbH, Unternehmensbereich Licht
  und Anlagen-Energie-Systeme
Monckebergstrasse 7
2000 Hamburg 1
Solar Electric International
31 Queen Anne's Gate
London, SW1H 9BU
Solar Usage Now Inc.
Box 306
420 East Tiffin St.
Bascom Ohio USA
Solarex Corp.
1335 Piccard Dr.
Rockville, Maryland 20850 USA
Solavolt International
3646 E. Atlanta Ave.
Phoenix, Arizona 85040 USA
Solec International
12533 Chadron Avenue
Hawthorne, California 90250 USA
Tri-Solar Corp.
10 DeAngelo Dr.
Bedford, Massachusetts 10730 USA
Virden Perma-Bilt
2821 Mays Ave.
Amarillo, Texas 79109 USA
Windlight Workshop
P.O. Box 6015
Santa Fe, New Mexico 87502 USA