its first cost, owing to the peculiar conditions of service.
The principal object in testing a stam engine is to deter-
mine the cost of power or the effect of such conditions as super-
heating, jacketing, etc., upon the' economy of the engine. We
must therefore, in the first case, measure the cost of fuel, and in
THE STEAM ENGINE.
the second the actual heat used. In either case we must calculate
the power of the engine.
The indicated power is determined in the familiar way by
means of the indicator, and the actual power by means of a dyna-
mometer or friction brake. (For further particulars of these in-
struments and apparatus see " Steam Engine Indicators.") To
determine the cost of power in terms of coal it is necessary to con-
duct a careful boiler test, usually of twenty-four hours duration.
When the cost is expressed in terms of steam per horse-
power per hour, we may follow either of two methods. We may
condense and weigh the exhaust steam, or we may weigh the feed
water supplied to the boiler. An hour under favorable condi-
tions is usually sufficient for such tests. Steam used for any
purpose other than running the engine must be determined sepa-
rately and allowed for.
Probably the most accurate terms in which to state the per-
formance of an engine is in B. T. U. per horse-power per minute.
When the cost is expressed thus, it is necessary to measure the
steam pressure, amount of moisture in the steam, and temperature
of condensed steam when it leaves the condenser. Jacket steam
must be accounted for separately. Important engines with their
boilers, etc., are usually built under contract to give a certain
efficiency, and their fulfilment of this contract can be determined
only by a complete test of the entire plant. Before beginning
the test, the engine should be run for a short time in order to
limber it up and get it thoroughly warmed. It is of the utmost
importance that all conditions of the test should remain constant,
especially the boiler pressure and the load. All instruments used
in the test should be tested themselves before being used, in order
to determine the effect of any errors to which they may be subject.
Thermometers. All important temperatures, such as feed
water, injection water, condensed steam, etc., must be taken by
reliable thermometers, the errors of which have been previously
determined and allowed for. Good thermometers sold by reliable
dealers are usually satisfactory. Thermometers with detachable
scales are subject to serious errors, and should be used only for
very crude work. Cheap thermometers are of little value in an
THE STEAM ENGINE. 129
Indicators. The most important and least satisfactory in-
strument used in the test is the indicator. It is subject to an
errqr of 2 per cent or 3 per cent at low speed, and this may easily
be two or three times as much at high speed. It does not work sat-
isfactorily at more than 400 revolutions per minute. If the indi-
cator is carefully tested under conditions similar to those on the
engine, the errors may.be reduced to a minimum, but there will
always be some uncertainty. The principal errors to which the
indicator is subject have been mentioned in the instruction paper
on " Steam Engine Indicators." It may, however, be well to add
that for accurate work we should always use two indicators, as
the long piping and joints necessary for only one causes a con-
siderable loss of pressure and much condensation. For marine
work it is customary to use only one indicator, with a three-way
cock, the lower end of the cylinder usually being inaccessible.
Scales. Weighing should be done on standard platform
scales. The water may be weighed in barrels provided with large
drain valves which will allow the water to run out quickly. It is
seldom possible to drain barrels completely, and so it is best to
let out what will run freely, then shut the valve and weigh
the barrel. This we call " empty " weight, and deducted from
the weight "full," evidently gives us the true weight of water.
If not convenient to weigh the water, it may be measured in
tanks or receptacles of known capacity, and the temperature
taken, allowing the proper weight per cubic foot for water at that
temperature ; or it may be determined by meters.
Meters. Water meters are of two kinds, those that record
the amount of water by displacement of a piston, and those in
which the flow is recorded by means of a rotating disc. Piston
water meters can be made very accurate, and if working under
fair conditions of service they may be relied upon to a close de-
gree. The chief error in a meter arises from the air that may be
in the water. To reduce this error to a minimum, the meter
should be vented, to allow the air to escape without passing
through the meter. Rotary meters are good enough for very
rough work, but are seldom sufficiently accurate for a careful en-
gine test. So far as possible, weirs should not be used in engine
work. They may be fairly accurate under certain conditions,
THE STEAM ENGINE. 131
but a very little oil in the water may affect them seriously. They
may sometimes be used to measure the discharge from a jet con-
denser, for then the volume is so large that the actual error is
G-ayes. Pressures should be measured on good gages that
have been recently tested by comparison with a mercury column.
The atmospheric pressure should be read from the barometer, and
for accurate work this pressure should be used. For ordinary
work, 30 inches, or 14.7 pounds, will do.
Calorimeter. When using superheated steam it is sufficient
to take the temperature and pressure in the steam pipe, but if
saturated steam is used, we must determine the amount of mois-
ture it contains. This is done by means of a calorimeter such as
described in ''Boiler Accessories."
Brake. Any of the forms of friction brake described in
" Steam Engine Indicators " will answer the purpose. For smooth
and continuous running it is essential that the brake and its band
be cooled by a continuous stream of water. The water may either
circulate in the rim of the wheel or around the brake band, but it
must not come in contact with the rubbing surfaces.
If the load is steady, seven or eight observations at equal in-
tervals will usually be sufficient. If possible, the cards should be
taken simultaneously, and then all the data averaged for the final
result. If the load fluctuates, the cards must be taken oftener,
and a greater number of observations wall be required. The great-
est care and accuracy must be used in all this work. In conduct-
ing a test, a careful log should be kept of the data; the outline
given on page 56 being a suggestion.
THE STEAfl TURBINE.
A description of the steam turbine and the general principles
of the engine were given in Part I of " The Steam Engine." Now
we shall discuss its efficiency. The turbine is such a comparatively
new engine that there have been but few tests made, and conse-
quently we know very much less about its possibilities than is the
case with the ordinary reciprocating engine.
Probably the greatest loss in the reciprocating engine is that
due to condensation and the subsequent re-evaporation at exhaust
THE STEAM ENGINE.
The exhaust cools the cylinder, so that the incoming steam meeta
cool walls, while in the turbine no such conditions exist. The
admission takes place at one end of the engine and the exhaust
occurs at the other end. The temperature gradually falls from
admission to exhaust and the expanded steam never comes in
contact with the part in which the higher-pressure steam works.
Thus Watt's principle seems to be fulfilled, namely, " The cylinder
should always be as hot as the steam that enters it." Of course
there is considerable loss from radiation, but there should be very
much less condensation than in the reciprocating type.
Superheating may be made use of. with considerable gain in
economy. There are no rubbing surfaces, no lubricants to decom-
pose and no glands to burn out, as is the case in the reciprocating
engine. The steam may be superheated to 60 or more to advan-
tage. There being no internal lubrication, there is of course no oil
to get into the condenser, and so the feed water may be used with-
out fear of getting grease into the boiler.
Another advantage seems to be in the more complete expan-
sion of the steam. There is little gain in the reciprocating engine
by expanding the steam beyond a certain limit, because of the in-
creased condensation. The boiler pressure cannot be increased
indefinitely, neither can the expansion be carried out to the limit.
From these considerations it would seem as if the turbine
ought to show much better efficiency than the reciprocating engine,
and were it not for the friction of steam against the vanes of the
turbine the advantage would doubtless be in its favor. Tests have
shown a consumption of about 16 pounds of steam per B. H. P.
per hour. Assuming an efficiency of 85 per cent, this would give
about 14 pounds per I. H. P. per hour. Tests of the best modern
triple expansion pumping engines have shown a steam consump-
tion of a little over 11 pounds per I. H. P., and numerous tests of
ordinary triple expansion engines have been made which show a
consumption of 12 to 13 pounds.
The most recent test of which we have accurate knowledge
was made on a Westinghouse-P arsons engine of 600 H. P. when
running at 3,600 revolutions per minute. With ordinary Bteam
having 3 per cent of priming, the turbine used 15.5 pounds of
steam per B. H. P. per hour with a vacuum of 25 J- inches in the
THE STEAM ENGINE.
condenser. With 30 of superheating and 26| inches of vacuum,
the steam consumption fell to 14.2 pounds. Assuming 85 per
cent mechanical efficiency as before would give a relative steam
consumption of 13.17 pounds per I. II. P. in the first case, and
12.07 pounds in the second case. It will be instructive to com-
pare these figures with those given for reciprocating engines oa
STEAM ENGINE INDICATORS
A most important question concerning a steam engine is,
"What is its horse-power? " or "How much' work will it do in a
given time ? "
Work is defined as pressure, force, or resistance multiplied by
the distance through which it acts.
Power is work done in a specified time.
In the steam engine, steam is the agent by means of which
heat is transformed into mechanical work. It is the heat in the
steam that does the work, not the steam itself.
Work is obtained from the heat in steam by confining it in a
closed cylinder which is fitted with a piston and a piston-rod.
Steam is admitted at one side of the piston while the other is open
to the atmosphere or in communication with a condenser. The
pressure of steam, usually 75 to 150 pounds per square inch, forces
the piston to the other end of the cylinder, driving out the low-
pressure steam in front of it. When it arrives at the other end,
steam is admitted to that end and the piston is driven back.
The piston moves because the pressure on one side is greater
than that on the other.
In order to move the piston, work must be performed. The
amount of work is easily found, since work equals total pressure
multiplied by the distance through which the piston moves.
Suppose a piston is 2 square feet in area and steam at a
pressure of 64.7 pounds per square inch acts on it during the
entire stroke of 4 feet ; the other side of the piston being in com-
munication with the atmosphere. The total pressure is then
2 X 144 X 64.7 = 18,633.6 pounds. If this pressure acts
through 4 feet it is evident from the definition that the work done
per stroke will be,
18,633.6 X 4 = 74,534.4 foot-pounds.
Another method is as follows: The pressure on the above
piston is 64.7 X 144 = 9316.8 pounds per square foot. The
volume swept by the piston during one stroke is 2 X 4 = 8 cubic
feet. If we multiply the pressure per square foot by the volume
or 9316,8 by 8, \ve get 74534.4 foot pounds, the same result as
before. Thus we see that work equals unit pressure multiplied
Let P - pressure on the piston in pounds per square foot.
p pressure on the piston in pounds per square inch.
A rr: area of piston in square inches.
L = length of stroke in feet.
V = volume swept by piston in one stroke in cubic feet.
W = work done in foot-pounds.
Then from the above example,
Work unit pressure multiplied by volume.
or, W = P X V
It is evident that P = 144 p, and V = A . X L.
Then we have these expressions for work,
W = P X A 7 = 144^ X V = 144 p X X L = p L A.
Suppose steam is admitted to the cylinder during the whole
stroke, as in the above example, that is, one end of the-cylinder is
in communication with the boiler. The other end is open to the
atmosphere. If we draw two lines at right angles to each other,
as O Y and O X in Fig. 1, the volume of steam for any position
may be represented by some distance measured on the line O X.
Similarly the pressure of the steam at any position of the piston
may be represented. by the length of a vertical line parallel to the
line O Y.
In the above example, the area of the piston was 2 square
feet, the length of stroke 4 feet and the pressure by gage 50
pounds. Then we let O A = the atmospheric pressure =14.7
pounds. At the beginning of the stroke the pressure (absolute)
is 14.7 -f- 50 = 64.7 pounds, represented by the distance O B, or
A B = 50 pounds pressure. When the piston has passed through
| of the stroke it is represented as the point 1, or B 1 is the
volume swept through when the piston has completed J of the
stroke. At this point the pressure is also 64. T pounds as repre-
sented at 1'. Similarly, when the piston is at 2, 3, and 4 the cor-
responding pressure is 2', 3 f , 4'. Since the pressure is constant
the line B D is parallel to O X. We see from the above that 50
pounds is the net pressure acting on the piston during the stroke,
and is represented by A B and lines parallel to it. The volumes
are represented by the horizontal lino A C. Then since "W P
X V it also equals O B X O X which is evidently the area of the
rectangle O B D X. The area of the rectangle O B D X is pro-
portional to the work done by the steam.
In Fig. 1, one inch on the line O Y = 40 pounds, then O B
is 1.6175 inches - long y
since it represents 64.7
pounds. Similarly O A
must be .3675 inch since
it represents 14.7
pounds. The line A C
is 2 inches long; then
referring to the preeed-
ing example, one inch
in length = | = 4 cubic
Since the rectangle
O B D X is 1.6175 by
But one inch in height
2 inches, the area is 3.235 square inches,
equals 40 pounds pressure and one inch in length equals 2 cubic
feet. Then p V = 40 X 3.235 X 4 = 517. G foot-pounds and,
W = 144 p V ~ 517.6 X 144 = 74,534.4 foot-pounds.
In the above cylinder the pressure acting on one side of the
piston was 64.7 pounds per square inch. There was also a press-
ure of 14.7 pounds per square inch (the atmospheric pressure) act-
ing in the opposite direction. Then the work done against the
steam pressure is represented by the area O A C X and is equal
to 144 p V - 144 X 14.7 X 8 = 16934.4 foot-pounds. Then
since O B D X represents the total work done on one side of the
piston and O A C X represents the work done against the piston
the difference A B D C represents the net work. This net work
is represented by the shaded area. Also if the amount of work done
on the piston is 74,534.4 foot-pounds and the work done against
the piston is 16,934.4 foot-pounds, the net work is the difference,
or 57,600 foot-pounds.
In this theoretical discussion the same result may be obtained
by subtracting the atmospheric pressure or back pressure from the
absolute initial pressure and using the difference as the value of
p. This value of p is called the mean effective pressure.
Then 64.7 14.7 = 50 and
W = 144 p V = 144 X 50 X 8 = 57,600 foot-pounds.
The area is proportional to the work done whatever the
shape may be ; provided the line B D represents the relation
between pressures and volumes on the steam side of the piston
and the lower line A C represents the relation between pressures
and volumes on the exhaust side. If the engine is of the con-
densing type the line A C will be nearer O X, which is the line
representing absolute vacuum.
Whatever the shape of the diagram, the area is equal to the
area of a rectangle of the same length and a height equal to the
mean height, or mean ordinate as it is called. The mean ordinate
represents the mean or average net pressure on the steam side of
the piston. Then we can follow these rules in finding the work
of the steam from the diagram.
Multiply the area in square inches by the scale of pressures,
by the scale of volumes and by 144, or ;
Multiply the length of the mean ordinate by the scale of press-
ures , by the length of stroke, and this product by the area of the
piston in square inches.
Example : The area of a diagram A B D C like that of Fig.
1 is G.3 square inches and its length is 3 inches. The scale of
pressure is 30 pounds per inch and the scale of volumes is 1.99985
cubic feet to the inch. If the piston is 20 inches in diameter and
the length of stroke 2|- feet, what is the work done per stroke ?
W = area of diagram X scale of pressures X scale of volumes X 144.
= 6.3 X 30 X 1.99985 X 144 = 54,428 foot-pounds.
\V = mean ordinate X scale of pressures X area of piston X
length of stroke.
=~ 2.1 X 30 X 314.159 X 2f 54,428 foot-pounds-
Thus we see that we get the same result by both rules. The
latter is the more common method because the mean ordinate is
easily found and the scale of volumes seldom considered.
In our consideration of Fig. 1, steam was admitted to the
cylinder during the entire stroke. In modern engines this method
is rarely used ; instead, steam is admitted during part of the
stroke then the communication to the boiler is cut off, and the
3team in the cylinder allowed to expand, as the piston moves for-
ward, until it fills the entire volume of the cylinder. This is rep-
resented graphically in Fig. 2.
Steam is admitted to the cylinder until the piston reaches
the point 2 which repre-
sents one-half the volume
of the cylinder. Then the
cylinder is half full of
steam, that is, it contains
|=4 cubic feet. The
four cubic feet of steam
expand until they fill the
cylinder. Since there is
the same weight of steam
present at every point in
the stroke and the volume
continued to increase, the
pressure must diminish.
This is shown in Fig. 2. The line B 2' is horizontal because the
pressure remains constant to the point of cut off. Then the
pressure begins to fall as is represented by the curved line 2' E.
This curve is nearly an equilateral hyperbola.
From Fig, 1 we know that the area 15 2' 2 A is proportional
to the work done while the piston moves from A to 2 or during
the first half of the. stroke. If we use the same data as we did in
Fig. 1, the work done must be one-half the work done in the first
case, or 576 ^ = 28800 foot-pounds. Also the area B 2' 2 A is
easily found since it is a rectangle. The area 2' E C 2 is found
by dividing it up into smalJ sections, by calculus or by the use of
It is easily seen that the area of the second case Fig. 2, is
less than that of Fig. 1. Therefore the work done is less ; but
the amount of steam admitted is only one-half as much as in the
In the first case, Fig. 1, 8 cubic feet of steam at 50 pounds
pressure were admitted per stroke and the wcrk done was found
to be 57600 foot-pounds. In the second case only half as much
steam is admitted and the work done is _ -f- the amount
represented by the area 2 f E C 2. Thus we see that there is a
considerable gain by expanding the steam.
Watt's Diagram of Work. Fig. 3 illustrates the method
adopted by James Watt to show the action of steam in the cylin-
der. The horizontal line A G called the abscissa represents the
length of the stroke and is divided into ten equal parts. The ver-
tical line A B called the ordinate indicates the pressure of steam.
When the piston ha? moved to the point E steam is cut off,
that is, a volume of steam equal to ^ the volume of the cylinder
expands until it fills the entire cylinder. The area may be found
by adding the several pressures (shown by the dotted lines), divid-
ing by the number of divisions, and multiplying by the length.
If by some arrangement of steam tight pistons working in cyl-
inders and having pencils fastened to them, we could get a dia-
gram like that shown in Fig. 3 it would be of great use but too
large' for convenience.
To obtain the same diagram on a small scale an indicator is
used. The value of such a diagram has already been shown when
finding the work done in the cylinder. The indicator has enabled
engineers to bring the engine of today to its present state of excel-
lence. A correct idea of the action of steam in the cylinder can
be obtained only by means of an indicator. It shows whether or
not the valves are set properly and how the condenser is working.
It also shows the engineer which end of the cylinder is doing the
most work. By comparing the expansion iine with an equilateral
hyperbola, with a curve of constant steam
weight, or with an adiabatic curve for
steam, the cylinder condensation is
James Watt was the first to see the
need of accurate knowledge of the action of
steam in the cylinder. He invented the in-
dicator. The improved form consisted of a
steam cylinder S, about one inch in dia'm-
eter and six inches long, in which a solid
piston P, is accurately fitted. A spiral
spring A, is attached to this piston, and
controls the motion of a pencil a, which
is also attached to the piston. This
pencil can operate on a sheet of paper
fastened to a sliding board, B. This board ,
moves back and forth by means of a weight at one end and a cord
at the other which is connected to some reciprocating part of the
engine. The indicator cylinder S, may be put in communication
with the engine cylinder by means of the cock C. With this
instrument a complete diagram can be taken.
Watt's first indicator had no lateral motion, therefore all it
showed was the pressure of steam in the cylinder and the perfec-
tion of the vacuum.
The diagram, or card as it is often called, obtained by the use
of an indicator is the result of two motions. The horizontal move-
ment of the paper corresponds exactly to the movement of the
piston, and the vertical movement of the pencil is an exact ratio
to that of the pressure of steam in the cylinder. The diagram
represents by its length the stroke of the engine and by its height
the steam pressure on the piston at the corresponding point of the
stroke. The diagram shows the action of steam on one side of
the piston only; to obtain the same information in regard to the
other side it is necessary to take another diagram from the other
end of the cylinder.
The essential features of an indicator are found in the instru-
ment invented by James Watt. Since his time, however, the
many improvements have made the indicator light, compact, dura-
ble, and accurate. Watt's diagram was traced on paper stretched
on a sliding board but now a revolving drum is used. The height
also of Watt's diagram was equal to the movement of the spring,
and the pencil arrangement was a simple contrivance. In the
indicators of the present day, the spring has a slight movement,
the height of the card being obtained by a multiplying arrange-
ment of levers. This method requires a parallel motion to obtain
accuracy in the vertical lines ; for if a lever is pivoted at one end
and power applied near the pivot the lever tends to rise and the
free end will describe an arc of a circle, not a straight vertical
THE THOHPSON INDICATOR.
Two views of the American Thompson Indicator, the outside
and the inside, are shown in Figs. 5 and 6. The form of spring
is shown in Fig. 7. The indicator consists of a cylinder in which
a piston is fitted, a spring, multiplying lever and parallel motion
for the pencil and a cylinder or drum for the paper. The piston,
which is .798 inch in diameter = | square inch in area, is fitted
accurately to the cylinder and has a travel of about one-half inch.