xOo .<> ^'t.^ ^ %- * '- s o ^ X Cj '^ ~o ,<^- '*A >"* ^ --^ . ^^^ ^^ -^^^ -^r .V* >^ % -^ O. "oo^ A ^0 '^ /- ■ ■ - ^J< ■-. ..^^ ..\^ r -^z. O v^" ^^ s^^'l: ' . K-^ ^\ ^^ l> ■ vO ^0 \ 1 « <3^ '-'^\a'' c^. A O^ ^.. v^' { ^P^^ ^^'"^.'^ ' -f^ ■=^0 0^ . -p -^ ^ f. , 'J . " ^ v*^^ vO ;\ 1/ ■ \ o5 -Tt, ^ W <> » 1 \ *■ \ k ■" ^ '-z.jyt- „ - , -* A ^ 'J N O ^/. v^^ -^^ •^^ f r -:, a"* ^ - J^" --' "^ r ^^ - ^ - A^ ^ c. ^^ -"^ i^ % .<>^'^' r A^^' ■''/' ■/■'. . --^ VALVE-GEARS FOR STEAM-ENGINES. BY CECIL H. PEABODY, ASSOCIATE PROFESSOR OF STEAM-ENGINEERING AT THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY. NEW YORK: JOHN WILEY & SONS, 53 East Tenth Street. 1892. ^^ Copyright, i8q2. BV CECIL H. PEABODY. ■Robert DuuMHOin>, Electrotyper, iAi &i46 Pearl Street, New York. Febsis Bros., Printers, S26 Peaxl Streefi»» New York, PREFACE. This book is intended to give engineering students instruc- tion in the theory and practice of designing valve-gears for steam-engines. With the vast number of valves and gears in use at the present time, an exhaustive treatment in a text- book appears out of place ; the author's aim is rather to give the learner a firm, grasp of the principles and some facility in their application. Each type discussed is illustrated by one or more examples selected from good practice. In the presentation of the elementary principles, geometri- cal or analytical methods are used as necessity or convenience may suggest ; but in the application, geometrical methods are used exclusively, in conformity with the usual and preferable habit of laying out valve-gears by construction. Zeuner's valve-diagram is used because it is widely and favorably known and appears to the author to be at least as good as any other circular diagram. In the discussion of radial valve-gears, ^he underlying Drinciples found in all such gears are pointed out, and a few prominent forms are illustrated. All such gears have neces- sarily or designedly large irregularities in their motions, so that analytical methods are useless if not misleading, and general methods of treatment are of small value. Facility in design is to be obtained through experience only. ill IV PREFA CE. Drop cut-off gears are represented by a few examples chosen to illustrate the principles and show a variety of treatment ; especial attention is given to the Corliss gear. Advantage is taken in this connection of the opportunity to illustrate the use of cam-gears. Common and well-known methods and processes have been used in most cases, and novelty has been rather avoided than sought. Some things, however, are believed to be here pre- sented for the first time ; for example, the combination of a skeleton model with construction for laying out link-motions and other irregular or complicated gears, and several examples of double valve-gears ; the latter are introduced mainly to show the scope of the methods used. So much of the material used is the accumulation of common practice, and so many of the forms and methods are known by the names of the originators, that references to authorities or formal acknowledgments ap- pear superfluous. C. H. P. Massachusetts Institute of Technology, March, 1892. \ TABLE OF CONTENTS. PAGE CHAPTER I. Plain Slide-valve, . i CHAPTER II. Shifting Eccentrics, 33 CHAPTER III. Link-motions, 39 CHAPTER IV, Radial Valve-gears, 88 CHAPTER V. Double Valve-gears, 97 CHAPTER VI. Drop Cut-off Valve-gears, 115 V VALVE-GEARS FOR STEAM-ENGINES. CHAPTER I. PLAIN SLIDE-VALVE. The valve-gear of a steam-engine consists of the valve, or valves, for adnnitting steam to, and exhausting steam from, the cylinder of the engine, together with the mechanism for giving motion to the valve, or valves. The discussion of valve-gears is therefore a part of kinematics or mechanism ; the extent and importance of the subject make a separate presentation of it desirable. The larger part of valve-gears derive their motion from one or more eccentrics ; of such gears, the plain slide-valve is the simplest. Other valve-gears are best studied after an exami- nation of the plain slide-valve, since they accomplish the same results, and by analogous methods. Slide-valve Engine. — The working parts of a plain slide- valve engine are shown by Fig. i, PL I ; the frame is omitted in order that those parts may be the more readily understood. The centre of the main shaft is at 6^ ; the crank-pin is at C; the centre of the eccentric is at £ ; Z is the connecting-rod, joining the crank-pin to the cross-head at //; / is the eccentric- rod, joining the eccentric to the head /i of the valve-spindle. Both cross-head and valve-spindle head are constrained by 2 VALVE-GEARS FOR STEAM-ENGINES. guides to move in a line parallel to the axis of the cylinder. The steam-chest is represented with the cover off, to show the valve and its seat ; the upper part of the valve is cut away to show the steam-ports and the exhaust-space. Fig. 2 gives a section of the valve and a half-section of the cylinder, with the piston P near the beginning of the stroke, and with the valve F partially open. The section being through the centre of the valve-spindle, the form of the exhaust-cavity is partially ob- scured. Fig. 3 gives a clearer idea of the valve and its seat. Crank and Connecting-rod. — The crank, connecting-rod, and cross-head, in its guides, form what is called a slider-crank train, more clearly represented by OCH, Fig. 4. Let the length of the connecting-rod be represented by Z, the length of the crank by R, and the angle which the crank makes with the centre line XX', by 6 ; then the displacement HA of the cross-head from the beginning of its stroke is ; D=OA-Oc -{HC - CcJ ; D =^ L-\- R- RcosO ^ {U - R' sin' 6)^ ; In designing valve-gears, it is more convenient, and suffi- ciently accurate, to find the displacement of the piston for a given crank-angle by the construction shown by Fig. 4. The converse of this, i.e. to find the crank-angle corresponding to a given piston-position, is found with equal facility by such a construction, while the calculation by aid of equation (i) is troublesome. The principal use of that equation is in study- ing the nature of the irregularity introduced by the connect- ing-rod into the motion of the cross-head, and for that purpose it is convenient to expand the expression containing L by the PLAIN SLIDE-VALVE. 3 binomial theorem, rejecting terms having the higher powers of L in the denominator ; whence D=R{\-co^e)-\-L\ I - (i - T\ r^f a\ \ ^ sin' d n = R{i- COS 6)+ ^^ (2) The ratio of crank to connecting-rod in stationary-engine practice varies from i : 5 to i : 7^^. In marine engineering the ratio I : 4, or even more, sometimes obtains. The maximum value of the term containing L occurs at ^ = 90° ; for the ratios of crank to connecting-rod just given, the term containing L has then the values -^^R, -^jR, ^R, respectively. It is appar- ent that the difference between the motion of the cross-head and piston, and harmonic motion represented by the equation D ^ R {1 - cos 6), (3) is always notable, and may be large. Eccentric and Eccentric-rod. — The eccentric is derived from the crank, by the expansion of the crank-pin till it in- cludes the shaft and obliterates the crank. Consequently the eccentric and eccentric-rod form a slider-crank train. The dis- placement of the valve is always reckoned from the middle position, and in Fig. 4 is ^ = ^^2: ; it may be calculated as fol- lows : e = Oe+Oa-{£k' - £7)^; ,'.e = r cos{90° -{0 + ^)\+^- U'- r'sm'{go°- {6 +d)}]^ ; • /zi , ^x , /f i r'cos»(^+<^)H1 , X ,-. ^ = rsm(^+d) + /|^i -| I _^_^|J. . . (4) 4 VALVE-GEARS FOR STEAM-ENGINES. This equation differs from equation (i) only in that the eccentric-angle and the valve-displacement are reckoned from different points. Expanding by the binomial theorem and rejecting terms containing the higher powers of / in the denom- inator gives • (f^\ ,, , r^cos''(^ + pOn — r cos (i8o° — 90° — 6 — d) = e\ .*. e = rsm{d -\- 6\ as given by equation (6). But the triangles Opn and OPN are equal, since they are right-angled triangles with the sides Op and OP equal, and the angles PON and pOn are each equal to 180° - 90° - 6/ - d. The use of the diagram is shown by Fig. i, PL III. Two valve-circles are drawn, OP for the forward and OP' for the return stroke. The circles oo'o" and ii'i'', drawn with the out- side and inside laps as radii, are called the lap-circles. At the beginning of the stroke the valve-displacement is On = Oo -\- on, and the lead is on ; the valve is in the position shown by Fig. 3, PI. II. As the crank moves forward the valve opens, at first rapidly, and then more slowly, till the maximum dis- placement is attained, when the crank coincides with 0P\ the position of the valve is then shown by Fig. 6, PI. II. As the crank moves beyond the last position the valve returns, till at ORc the displacement is equal to the lap and cut-off occurs ; the position of the valve is then shown by Fig. 4, PI. II, and it is moving toward the right to shut the port. At ORk the displacement of .the valve is equal to the inside lap and com- pression occurs; the position of the valve is shown by Fig. 5, PI. II, and it is moving toward the right to shut the exhaust- port. At ORr the displacement of the valve is again equal to the inside lap and release occurs ; the valve now has the left- hand inside edge on the edge of the port and it is moving to open the port. At ORJ the displacement of the valve is equal to the outside lap and admission occurs at the crank end ; the position of the valve to give admission at the head end (cor- responding to the crank-position OR^ is shown by Fig. 4, PI. II. It is customary to associate the admission at OR^, in anticipation of the forward stroke, with that stroke, and to associate the admission at ORJ with the return stroke. lO VALVE-GEARS FOR STEAM-ENGINES. ' jr Though it is not customary to do so, the two arcs tt t'\ ss s may be added to find the crank-positions at which the edge of the valve is on the further edge of the port, and the port is wide open. The radius Ot' is made equal to the width of the port plus the outside lap ; when the crank-position passes through / or t" the displacement of the valve is equal to the lap plus the width of the port and the port is then wide open. In like manner the radius Os' is made equal to the width of the port plus the inside lap, and when the crank-posi- tion passes through s or s^' the port is wide open for exhaust. The diagram shows that the outside edge of the valve over- travels the edge of the port by the amount t'P. Some over- travel is desirable to give a free flow of steam to the cylinder and a rapid admission and a sharp cut-off. The over-travel of the valve for exhaust is s'P, which is greater than t'P by the amount of the difference of the outside and inside laps. The amount by which the port is open at any position of the valve is called the port-opening. The maximum port-opening for supply, which occurs when the crank coincides with OP, is equal to o'P, the difference between the eccentricity and the lap. The maximum port-opening for exhaust is equal to i'P. A slide-valve moved by an eccentric always has the maximum port-opening for exhaust at least as great as the width of the port, and it is commonly greater; the maximum port-opening for supply is also commonly greater than the width of the port, but it is sometimes less. A slide-valve moved by a gear that gives a variable cut-off, as will be seen later, may have both maximum port-openings less than the width of the port for some settings of the gear. An inspection of the diagram Fig. i, PI. Ill, will show that a change of the outside lap will affect both admission and cut- off ; thus, the cut-off is hastened and the admission is delayed by an increase of the outside lap, and conversely the admission comes earlier and the cut-off comes later if the outside lap is decreased. In a similar way, increasing the inside lap delays PLAIN SLIDE-VALVE. II the release and hastens the compression, while decreasing that lap produces a contrary effect. Since it is the relative propor- tions of lap and eccentricity which determine the cut-off, it is apparent that decreasing the eccentricity with a constant lap produces the same effect as increasing the lap with a constant eccentricity ; i,e. it hastens the cut-off and delays the admis- sion. Finally, it will be seen that increasing the angular advance hastens all the events of the stroke ; and that decreas- ing the angular advance delays all the events of the stroke. Should the inside lap be made nothing, so that in mid-gear the inside edge of the valve shall coincide with the edge of the port, then both compression and release will occur when the crank is at right angles to POP' . Sometimes, in designing or remodelling a valve, it will be found advisable to give a clear- ance to the valve instead of an inside lap, in which case the engine will for a short time, near the end of the stroke, ex- haust from both ends at once. Suppose that the circle ii'i'\ Fig. I, PI. Ill, represents a clearance, then release will occur at ORk and compression at OR,.. Expansion and Compression. — From cut-off at OR, (Fig. I, PI. Ill) to release at OR^-, the head end of the cylinder is shut off from both the supply and the exhaust. While the crank moves forward from R, to R^ the piston moves a corre- sponding amount, and the steam in the cylinder expands and experiences a loss of pressure in consequence ; this action is called the expansion. When the valve closes the exhaust at ORk the steam then caught in the cylinder is compressed ahead of the piston till a new supply of steam is admitted at ORJ ; this action is called the compression. On the return stroke, in a like manner, steam is expanded from OR, to ORJ, and is compressed from ORk to OR^- Rocker and Bell-crank Lever. — In the work thus far it has been assumed that the centre-lines of the crank and con- necting-rod with the cross-head and piston, and of the eccentric and eccentric-rod with valve-spindle and valve, coincide in the 12 VALVE-GEARS FOR STEAM-ENGINES. elevation, as shown by Fig. i, PI, I. This assumption is con- venient in giving the first description of the valve-gear and in discussing the action and the theory of the valve-motion ; and the design of the valve is commonly carried on as though such a coincidence existed, the deviation from such a coincidence being considered only in the mechanical problem of laying out the mechanism of the engine. If the centre-line of the valve-spindle passes through the centre of the shaft, and the eccentric-rod is connected directly to the valve-spindle, then the motion of both crank and con- necting-rod, and eccentric and eccentric-rod, referred to their own proper axes, will be the same as already found, even though their centre-lines do not coincide. Such a lack of co- incidence will make the angle between the eccentric and the crank more (or less) than 90° plus the angular advance, by the amount of the angle between the two centre-lines. This dif- ference needs consideration only in the process of setting the valve ; and if the angle between the centre-lines is small, it will require little or no attention at that time. Since such an arrangement involves a lack of parallelism between the paths of the valve and of the piston, the work of boring thecylin- der and facing off the valve-seat is more troublesome, and other machine-work is more difficult, unless special processes are pro- vided ; consequently this arrangement is seldom adopted. Very commonly the paths of the piston and the valve are parallel but do not coincide in the elevation ; thus, the axis of the crank and connecting-rod with the cross-head and cylinder may be XX' in the Figs. 2, 3, and 4, PI. Ill, while the centre- line of the valve-spindle may be xx' in the same figures. In such cases a rocker or a bell-crank lever should be used to transmit motion from the eccentric to the valve. The following method may be used in laying out a bell- crank lever. Let A be the position of the end of the valve- spindle when the valve is in mid-position ; lay off Aa = Aa' equal to the lap of the valve, and with a radius equal to the PLAIN SLIDE-VALVE. 1 3 length of the arm of the bell-crank lever draw arcs intersecting at C\ this gives the axis about which the bell-crank lever vibrates. This construction prevents the bell-crank lever from introducing any irregularity into the action of the valve at admission and cut-off; irregularities at other times are of less consequence. In laying out such a gear for a locomotive with a rigid valve-spindle that extends directly from the end of the bell-crank lever or rocker to the valve-yoke, it is important to have the bending or lateral motion of the valve-spindle as small as possible ; in such case the point C may be so chosen that the lateral motion of the end of the valve-spindle shall be half above and half below the line xx' . With a radius equal to BC^ the other arm of the bell-crank lever, draw an arc as shown, and draw a tangent to this arc from O. Draw perpendiculars OE and CB from O and C to this tangent ; then EB is the length of the eccentric-rod. If desired, the relation of the crank and eccentric may be found by laying off the angle XOR = ^, the angular advance, since the crank is at that angle before the dead-point when the valve is in mid-position ; the angle EOR will not be equal to 90° -f- S, but this is a matter that affects the valve-setting only, and even in that process the exact knowledge of the angle between the crank and eccentric is not of importance. If it is considered of importance that the eccentric-rod shall be some definite length, then the centre C, on which the bell- crank lever vibrates, may be shifted so as to give that length. If C is to be shifted a short distance, then a line parallel to XX^ may be drawn through B, and with a radius equal to the desired length of the eccentric-rod an arc may be drawn from E inter- secting that parallel line at a point B^ ; the whole bell-crank lever is to be shifted bodily to the extent BB\ and the length of the valve-spindle must be changed the same amount. In the figure the arm CB is made J of CA ; consequently the motion of the valve will be that which would be given by an eccentric equal to f of OE if the connection were direct. 14 VALVE-GEARS FOR STEAM-ENGINES. In designing and laying out the valve it is treated as though it were moved by such an eccentric. The ratio of the arms may be made anything desired ; they have commonly the same length. In laying out a rocker, the process is the same as that just described for the bell-crank lever, except that the arm CB, Fig. 3, is laid off on the side opposite A, and the eccentric follows the crank. Fig. 4 shows the equal-armed straight rocker with the centre of vibration C midway between the lines XX' and xx' ; it may be made a little longer so as to give a construction equivalent to that shown by aAa' , Figs. 2 and 3. Area of Steam-pipe and Steam-ports. — In order that the loss of pressure in the steam-pipe due to friction may not be excessive, it is customary to limit the velocity to 100 feet a second or 6000 feet per minute. The volume of steam supplied to an engine is calculated, for this purpose, on the assumption that the cylinder is filled at each stroke. For example, the diameter of the steam-pipe of an engine having a diameter of 18 inches and a stroke of 3 feet, and making 75 revolutions per minute, should be 5 inches. It may be found as follows : The piston displacement is 7r/i8\2 \^) X 3 = 17671 X 3 = 5-30 cubic feet; the volume of steam per second is 7c 2 X 5.30 X ^ = 13.25 cubic feet; the area of the steam-pipe should be 13.25 100 X 144 = 19.08 square inches, PLAIN SLIDE-VALVE. 1 5 and the diameter should be X 19.08) = 4.93 inches, or nearly 5 inches. Even though the cut-off occurs before the end of the stroke and the actual volume of steam used is less than the volume calculated by the above method, the area of the steam-pipe should not be reduced, for the rate of the flow of steam will be as great when the piston is near mid-stroke. The steam- pipe supplying an engine which has an early cut-off (one third stroke or less) may be made less than given by the above method, provided there is a large steam-chest, or a steam-drum near the engine. Seaton * states that a velocity of 8000 feet may be allowed in the steam-pipe of a marine engine; and that 10,000 feet may be allowed for very large engines. The area of ports and passages leading to the cylinder should be equal to that of the steam-pipe ; and the area of ports and passages leading from the cylinder should be double that area. If the steam-pipe is calculated on the assumption of a velocity of 6000 feet a minute, this rule will often make the size of ports and passages such that it is dif^cult, if not impossible, to provide for them in the design of a high-speed engine. In such case the area of ports and passages must be reduced, but they should never be less than half the area given by that rule. This is equivalent to calculating the area of a steam-port by the method given for a" steam-pipe, except that a velocity of 12,000 feet a minute is allowed ; and then making the exhaust-passage twice that area. Since both supply and exhaust pass through the ports of a plain slide-valve engine, the area must be made sufficient for the exhaust. For example : the engine mentioned above should have the area of the port 2 X 19.08 = 38.16 ; * Manual of Marine Engineering. l6 VALVE-GEARS FOR STEAM-ENGINES. and if the length of the port is made eight tenths of the diam- eter, or 14 inches, the port will be 38.16 -^- 14 =r 2.65 inches. If the area of the port is reduced to half as much the width will become if inches. When it becomes difficult or undesirable to give the slide- valve sufficient motion to open the steam-port wide for the supply, the maximum port-opening may be made from -f^ to y^^ the width of the port. When special valve-gears are used that open the valve rapidly and close it promptly, the area of the ports and pass- ages may be made smaller than the above methods provide^ but such reduction should be made only with complete knowl- edge of the action of the valve and of the effect of the reduc- tion of the flow of steam. Lead and Lead-angle. — The lead, or the amount that the valve is open when the engine is on a dead-point, varies with the type and size of the engine, from a very small amount (or even nothing) up to f of an inch or more. Stationary engines running at slow speed may have from -^-^ to -f-^ of an inch lead. The effect of compression is to fill the waste space at the end of the cylinder with steam ; consequently engines having much compression need less lead. Locomotive-engines having the valves controlled by the ordinary form of Stephen- son link-motion may have a small lead when running slowly and with a long cut-off, but when running at speed with a short cut-off the lead is at least J of an inch ; and locomotives that have a valve-gear which gives constant lead commonly have ^ of an inch lead. Zeuner's diagram does not admit of the use of the lead in solving problems that arise in designing valves, but we may use instead the lead-angle, or the angle that the crank makes with the line of dead-points at admission ; in Fig. i, PI. Ill, PLAIN SLIDE-VALVE. 1 7 XORa is the lead-angle. The lead-angle may vary from zero to 8° ; a convenient lead-angle for solving problems is 2^°. After the required problem is solved by the aid of the lead- angle the conditions may be varied so as to give a desired lead. Problems on the Slide-valve. — By assuming various ele- ments of the valve to be known, a series of problems relating to the plain slide-valve may be stated and solved. Of such problems oite has a real importance to the designer of valve- gears ; others are either so simple as to require no formal solu- tion, or they are problems that are not likely to arise in prac- tice. This important problem is given below as Problem II ; the other, Problem I, is given as a convenient introduction to it, for students who here approach the subject for the first time. Problem I. Give^t the eccentricity^ the lead-angle, and the crank-angle at cut-off, to find the angular advance, the lap, and the lead. In Fig. I, PL IV, draw, to any convenient scale, the arc XR,X' to represent the path of the crank, referred to the axes XOX' and OY. Lay off the angle XOR^ equal to the lead- angle, and lay off XOR^ , the crank-angle at the point of cut-off. Should the piston-position at cut-off be given instead of the crank-angle, lay off Or, the given piston-position at cut-off, and draw the vertical line rR,, to find the crank-position OR^ corre- sponding. With equal leads and laps the crank-angle at cut-off will be the same for each stroke, and the mean piston-position will be nearly equal to the piston-position at cut-off with har- monic motion ; hence the above construction. Bisect the angle R^OR, and draw the line OP', on it draw the valve-circle OoPo'" , and through the intersections of this valve-circle by the lines ORa and OR^ draw the lap-circle 00' 0" . The lead is Oa. In Fig. I the eccentricity is \\ of an inch, the cut-off is at f of the stroke, and the lead-angle is 2|^°. The lap is |f of an 1 8 VALVE-GEARS FOR STEAM-ENGINES. inch, and the lead is y^ of an inch ; the angular advance is Problem II. Given the crank-angle at cut-off, the lead-angle, and the maximum port-opening, to find the eccentricity, the lap, and the angular advance. As has already been stated, this problem is the one met by the designer in laying out a valve-gear. The angular advance is obtained by the same process as in Problem I ; namely, by bisecting the angle R^ORc- Now assume a trial eccentricity, preferably a little larger than the required eccentricity, and draw an assumed valve- circle OqP^ , and the corresponding lap-circle qq'q" . The maximum port-opening with the assumed valve-circle is q"P^ ; and from it the diameter of the required valve-circle, equal to the required eccentricity, may be obtained by the proportion • Assumed port-opening : actual port opening:: Assumed eccentricity : required eccentricity. A graphical solution may be made by aid of similar triangles, as follows : Lay off the line Oc^ = in a convenient posi- tion, and join C^c^ ; make Oc equal to the given maximum port- opening, and draw cC parallel to c^C^ ; then C is the centre of the required valve-circle and OP is the required eccentricity. The lap-circle oo'o'" is drawn through the intersections of the valve-circle by Ra and R, . In Fig. 2, PL IV, the angular advance is 3iJ°, as in Fig. i ; the assumed eccentricity is if of an inch, which gives a port- opening of f|- ; the required maximum port-opening is taken to be f of an inch ; the required eccentricity is ly'^^ of an inch ; and the lead is yig- of an inch, or a trifle less. Modifications. — In general it is not of great importance that the cut-off shall occur exactly at the chosen crank-angle or piston-position, and it is seldom necessary to know the angular advance except in drawing the valve-diagram. It is, however, very convenient if not necessary that the lead and lap shall be PLAIN SLIDE-VALVE, I9 some determined quantity stated in fractions of an inch that are commonly used in the machine-shop. By judicious modi* fications the designer may secure this for the valve without seriously affecting the point of cut-off. In Fig. 3, PI. IV, the lap, Oo^ is made f of an inch, and the lead is -^^ of an inch. At a the vertical aP is drawn, and from O, with a radius equal to i J of an inch, the vertical is intersected in P\ thus giving the diameter OP on which the valve-circle OaPo" is drawn. The cut-off comes at the crank-position OR^ , corresponding to the piston-position JCr instead of JC^ = f stroke, as required. The process of laying out a slide-valve will be considered in connection with the valve for equal cut-off. Equalization of Cut-off at the Expense of the Lead. — Let it be assumed that the cut-off shall occur at a given piston position for each stroke, taking account of the irregularity due to the connecting-rod. Draw the line j::X\ Fig. i, PI. V, on which choose for the centre of the crank and for the origin of coordinates, and draw the vertical axis VOY\ Draw the circle XVX' Y\ to any convenient scale, to represent the crank- circle. With a radius equal to the length of the connecting-rod, on the same scale, and with JC and X^ as centres, cut the centre- line X^x at X and x' ; this will give the stroke of the cross-head, equal to the stroke of the piston. Lay off the point on the stroke at which cut-off is to occur on both strokes, and with these points as centres and the length of the connecting-rod as a radius intersect the crank-circle at R^ and R/. In the figure the connecting-rod is taken to be five times the crank, and cut-off is assumed to occur at f of the stroke. It is at once apparent that the crank-angles XOR^ and X' OR^ are not equal, and ORc and OR^ are not one straight line. Choose a small head- end lead-angle XOR^\ in the figure it is 1°. Bisect the angle R^OR,, and draw the line POP' , on which draw the two valve- circles as shown. The eccentricity may be determined by a preliminary solution, assuming harmonic motion, as in Problem II, or the same solution may be made directly on the figure ; 20 VALVE-GEARS FOR STEAM-ENGINES. if the latter method is used, confusion is liable to occur from the number of circles drawn, especially if the eccentricity is modified to get some convenient dimension ; consequently it is better to make such a construction separately and transfer the results to the main diagram. Draw the lap-circle oo for the upper valve-circle, through the intersections of that circle by ORa and ORc ; and draw the lap-circle o'o' for the lower valve- circle, through the intersection ORJ with that circle. The admission at ORJ occurs at the intersection of the lower valve- circle and the lap-circle o'o' ; the crank-end lead is large, if not excessive. It is customary, in designing a valve for equal cut-off, to equalize the compression also. In Fig. I the compression is assumed to occur at f of the stroke, or at the crank-positions ORk and ORk The inside laps are it and i'i', so that the release occurs at R^ and i?/. The point of intersection i of the line ORk with the valve-circle may be determined by drop- ping a perpendicular Pi from P on ORj,. The valve-diagram in Fig. I gives the following dimensions : Eccentricity ij inch. Outside lap, forward stroke -|^ *' return stroke f T " Inside lap, forward stroke tV " return stroke -g^^ " Head-end lead A " Crank-end lead -^1- " To lay out a Slide-valve. — The valve for which dimen- sions were found in Fig. i PL V, is shown in section by Fig. 2. To lay out the valve, begin at the crank end and make ab =^ ^ of an inch, equal to the return-stroke outside lap ; make be = r^-^ of an inch, equal to the width of the port. The greatest displacement of the valve, equal to the eccen- tricity i|- of an inch, will carry the point a to a\ and when the PLAIN SLIDE-VALVE, 21 valve is in that position it must not overrun the edge of the bridge, but rather there must be width enough remaining to prevent leakage. The least width of bridge in the figure is f of an inch, and the width of \ inch is chosen to insure a joint. The forward-stroke inside lap, -f-^ of an inch, is laid off at cd. The greatest displacement of the valve will carry the point d to d\ and at that position of the valve the remnant of the exhaust-space should be at least as wide as the port, i.e. -^-^ of an inch as shown. The exhaust-space is commonly made wider than this construction gives ; it should not be unduly increased, since it will then make the valve large and the friction excessive. The valve is completed by making the width of the bridge \ of an inch and of the port, -f^ of an inch, as shown. If the eccentricity, \\ of an inch, be laid off toward the left, from the right-hand edges of the valve, it will appear that the right-hand bridge is wider than necessary, and that the remnant of the exhaust-space, when the valve has its maximum displacement to the left, is greater than the width of the steam-port. No inconvenience will occur from such an excess of bridge or exhaust-space ; but had the construction been begun at the right hand, then both the bridge and the exhaust-space would be too narrow. For constructive reasons, the bridge for any slide-valve maybe made wider than required to. prevent leakage. Fig. 3, PL V, gives the section (half-size) of a valve with equal laps which will give the same average cut-off as the valve shown by Fig. 2 ; the cut-off at the head end will be longer, and that at the crank end will be shorter. The outside lap for Fig. 3 is very nearly the mean of the unequal outside laps of Fig. 2 ; and the inside lap is very nearly the mean of the inside laps of the same figure. Such a valve may be laid off, begin- ning at either end. The height of the exhaust-cavity of the valve should never be less than the width of the steam-port ; it is commonly once and a half as high, if not more. The method just given for laying out the slide-valve has 22 VALVE-GEARS FOR STEAM-ENGINES. the apparent inconvenience that the centre of the exhaust- space cannot be directly located on the assembly drawing of the engine. This difficulty is, however, only apparent, for the section of the valve is commonly drawn separately and at full size, and then can be transferred to the assembly drawing, which may be to any convenient scale. The results obtained by laying out the work on the drawing-board should always be checked by a numerical calculation ; and if desired such a numerical calculation may be made first, but it should be checked by the subsequent laying out of the valve. Thus the width of the bridge should be greater than ^\ —\\ — tV = If' <^^ nearly | of an inch ; and the exhaust-space should have a width of tV + li + tV - ¥ = If of an inch. Equalization of Cut-ofT with Rocker. — A method of equalizing the cut-off without destroying the equality of the lead was devised by Professor Sweet, and is employed on the Straight Line engine. The same method may be employed with a bell-crank lever; the construction must, however, be carried out for each case separately. On Plate VI the centre-line of the crank and connecting- rod is xX' , while the centre-line of the valve-spindle is nn' , Assuming that the cut-off is to occur at f of the stroke, divide the diameter XX' of the crank-pin circle into fourths and draw the vertical line NR^ from N\ in the figure the eccentric-circle, drawn to a larger scale than the crank-circle, happens to pass through N. Make the angle XOR^ equal to the lead-angle, and bisect the angle RaOR^ by the line OP. Draw the valve- circle on (9Pand the lap-circle ^^', thus finding the outside lap. With C and C ,^\. f of the forward and return strokes as centres, and with the length of the connecting-rod as radius, intersect the crank-circle at R^ and i?/, to find the crank-posi- PLAIN SLIDE-VALVE. 2$ tions at cut-off. Produce the line R^O to RJ, the crank- position at admission for the return stroke. Since the gear is drawn with a rocker, the eccentric will follow the crank at an angle equal to XOP. The positions, r^ and r/, of the eccentric at cut-off will be found by laying off the angles Rfir, and RJOrJ, equal to XOP', and the positions of the eccentric, r^ and rj, at admission are found by making the angles RaOr^, and RJOrJ equal to the same angle. With r^ and r^ as centres, and with a radius equal to the length of the eccentric-rod, draw two arcs intersecting at e^ ; also with the same radius draw arcs from rj and rj intersecting at e] then if the end of the eccentric-rod be guided on a path passing through e and e^ the lead and cut-off will both be equalized. With a radius equal to the length of the arm of the rocker, 'draw arcs intersecting at T^ , and from T^ draw an arc through ee^ , and also the chord ee^ ; the length of the chord is greater than twice the lap ; consequently the other arm of the rocker should be made smaller than T^ey in proportion as the lap is less than half ee^ . The length of the other arm may be found by a numerical calculation, or by the following construc- tion: make ^^ equal to the lap, and draw st parallel to eT^\ then st is the required length of the rocker-arm. The construction may now be completed by shifting the centre of the rocker-shaft T^ , so that the other arm shall have the proper motion with regard to nn\ the centre-line of the valve-spindle. To make this construction, draw a line //' par- allel to nn', and at a distance equal to /^, taken from the triangle sqt or found by calculation from a proportion involving the side T^g of the triangle eT^g. With 6^ as a centre and with a radius equal to OT^ intersect the line //' at T; this gives the position of the centre of the rocker-shaft. With 7" as a centre the arcs da^d' and dc^d^, on which the ends of the rocker- arms travel, can be drawn. At the same time that T^ is swung to T, the point e„ is swung to c„ , found by intersecting the arc dd^ from O with a radius equal to Oe^ . From T draw Ta^ per- 24 VALVE-GEARS FOR STEAM-ENGINES. pendicularto ;2;2'; then c^Ta^ is the position of the rocker when the valve is in mid-position, and aa' is equal to twice the lap. The position of the eccentric will be found by moving it through an angle equal to TOT^ \ and in the same direction ; thus when the crank is at ORd (the crank-end admission) the eccentric is at r/^ found by making rJOrJ'^= TOT^. The heavy black lines are drawn to show the eccentric, eccentric- rod, and rocker at the crank-end admission. The true position of the eccentric is of importance only in setting the valve and need not be known exactly even then. The extreme positions of the eccentric-rod and of the rocker are to be found by trial, and from them the extreme positions b and b' of the head of the valve-rod or valve-spindle. Since this method of equalizing the cut-off introduces some irregu- larity, a complete study of the valve-motion, by aid of the valve- ellipse or otherwise, is desirable. Piston-valve. — If the section of a plain slide-valve and its seat, as shown by Fig. 3, PI. I, be supposed to revolve about an axis xx'y there will be generated a piston-valve with its cy- lindrical seat. Such a valve is represented by Fig. i, PL VII, which gives a section of the high-pressure cylinder and valve of the U. S. battle-ship Massachusetts. It will be seen that the outside shell of the cylinder, the lower cylinder-head, and the valve-chest form one casting, with feet attached for bolting to the engine-frame. A cylinder-liner is forced into and secured to the outer shell, with a space between to serve as a steam- jacket. The piston is of conical form, and the heads are shaped to correspond. Leakage is prevented by two packing- rings held in place by a junk-ring. The piston-valve is in the shape of two pistons connected by a pipe or sleeve through which the valve-spindle passes. The valve-spindle is prolonged beyond the valve and attached to a small balancing piston which relieves the valve-gear of the weight of the valve and attached parts ; the upper end of the balancing cylinder is connected with the exhaust. The valve- PLAIN SLIDE-VALVE. 2 5 seat is formed by two short hollow cylinders, forced into the shell of the valve-chest. The space surrounding each half of the valve-seat is connected with and forms part of the passage leading to the cylinder. The ports are cut through the cylin- drical valve-seat as shown. Steam is supplied to the middle of the steam-chest, and is exhausted from the ends through pipes shown by dotted circles. This arrangement secures the advantages that the supply and exhaust steam are kept well separated so that heat cannot easily pass from one to the other, and the valve-rod stuffing- box is exposed to the exhaust steam only ; such an arrange- ment is not advisable for a cylinder in which there may be a vacuum, since the leakage of air inward, past the stuffing-box, is more troublesome than the escape of steam. The laps con- trolling the admission and cut-off are by this arrangement placed inside ; while the laps controlling the exhaust and com- pression are outside. To avoid confusion, it is advisable to distinguish them as steam-lap and exhaust-lap ; the design of the valve-gear by the aid of valve-circles or otherwise may with this understanding be carried out as usual. It will be noticed that the top-end steam-lap is the larger; while the top end exhaust-lap is the smaller, and is here a negative lap or clearance. This arrangement is adopted to hasten the cut-off and compression on the down-stroke and delay them on the up-stroke, but is not carried far enough to produce complete equalization. The greater lead at the lower end helps to com- pensate for the shorter cut-off and the weight of the recipro- cating parts. Leakage past the valve is prevented by packing-rings, like those of the piston, which form the acting-edges of the valve, To prevent the valve-rings from springing into the ports, bridges are left as shown at A, Double-ported Valve. — It is frequently difficult or impos- sible to get sufficient width of port for engines having a large diameter and short stroke, if the common plain slide-valve is 26 VALVE-GEARS FOR STEAM-ENGINES. to be used. Fig. 2, PI. VII, shows a device, known as a double- ported valve, used in marine engineering to overcome this dif- ficulty. Each passage leading to the cylinder has two ports, and two slide-valves, joined together and forming one casting, to control the flow of steam through those ports. The inner valve resembles the common slide-valve, except that there is a communication through the top between its exhaust-space e^ and the exhaust-space ^of the outer valve. The outer valve is elongated enough to leave a steam-space {a and a') to supply the inner valve ; a bridge between e and a separates the exhaust of the outer valve from the steam-space of the inner valve, and is continued to the opening through the top of the inner valve. Fig. 3 gives, at the left hand, a transverse section through the exhaust space e, and at the right, through the steam-space a. The space a is drawn down toward the middle of the valve as shown, so that the valve may be made compact while providing sufficient area for the flow of steam. Allen or Trick Valve. — Fig. 4, PI. VII, shows a valve which is so made that a double admission of steam takes place at and near cut-off and admission. It is used with the link- motion and other gears which give a variable cut-off with the slide-valve, and is intended to remedy the defects due to the slow motion imparted to the valve at those points when the cut-off occurs early in the stroke. Through the body of the valve there is a passage ss' , and the valve-seat is cut away so that the distance from the outer edge of the passage to the edge of the valve-seat is equal to the outside lap of the valve. If the valve is displaced toward the right by the amount of the outside lap, the edge c of the valve is brought to the edge of the port a, and at the same time the edge of the passage / is brought to the edge d of the valve-seat. Consequently there is a double admission of steam to the port a, one in the usual way past the edge c, and the other under the right-hand end of the valve, past the edge doi the valve-seat, and through the passage s's. As the valve opens PLAIN SLIDE-VALVE. 2^ wider, the passage ss' is liable to be shut off by traversing past the farther edge of the port a, but when that happens the supply past the edge c is abundant. Near cut-off the passage ss' is again opened at s and gives a double supply of steam till cut-off occurs by the simultaneous coincidence of c with the edge of the port a, and of d with the edge of the passage /. A modification of this form of valve, used on the Arming- ton and Sims engine, is shown by Fig. 2, PI. IX. The valve is a piston-valve, taking steam at the middle and exhausting at the ends, and differs from the valve shown by Fig. i, PI. VII, in that the stem connecting the two ends of the valve is hollow, and in that there is a passage from this interior channel through the valve-face. In the figure the valve is giving admission to the head end, and steam enters the cylinder directly from 5 at a, and also from 5 at b, through the hollow stem and thence through the passage in the valve-face. Balanced Valves. — When the difference of pressure be- tween the steam and exhaust pipes is large, the force exerted to hold a plain slide-valve against its seat is very large, and the friction of the valve on its seat is excessive. This consumes a needless part of the work developed by the engine, throws a severe duty on the valve-gear, and makes it difficult to main- tain the acting-surfaces of the valve and its seat in good con- dition. Various methods of relieving valves from part or all of the steam-pressure on them have been devised, resulting in what are called balanced valves. The piston- valve (shown by Fig. i, PL VII) has no press- ure on it to hold it against its seat, and is consequently per- fectly balanced. It is very commonly used for the high-pressure and intermediate cylinders of triple-expansion marine engines, and on high-speed engines under the control of a shaft-gov- ernor. When well made, and provided with packing-rings, there is no more reason for leakage than exists with the piston of the engine. Small piston-valves are commonly made with- out packing-rings, and then depend on the fit in the valve-seat 28 VALVE-GEARS FOR STEAM-ENGINES, to prevent leakage. It is claimed that they do not leak when new, that the wear is insignificant, and that both valve and seat may readily be renewed when necessary. It is, however, probable that such piston-valves do frequently leak in common service. The double-ported valve (Figs. 2 and 3, PL VII) and the Allen valve (Fig. 4, PI. VII) have part of the pressure on the back relieved, and are known as balanced valves. The double- ported valve has a shallow cylindrical recess turned in its back. In this is a short cylinder or ring that is pressed by springs against a finished surface on the valve-chest cover. A bronze ring fastened to the valve and bearing against the vertical ring or cylinder is intended to prevent leakage. Communica- tion is opened between the enclosed space and the exhaust, so that the leakage may not accumulate in this space and destroy the balancing of the valve. The unbalanced pressure of the steam on the unenclosed part of the valve gives enough press- ure against the seat to prevent leakage. The Allen valve is commonly much longer than wide, and consequently a rect- angular balancing-frame is used to exclude pressure from part of the top of the valve. Leakage into the enclosed space is allowed to flow directly into the exhaust-cavity through a small round passage, shown by dotted lines. All such devices are somewhat costly to make and troublesome to maintain in good condition, and if allowed to get out of condition are liable to a large loss from leakage directly into the exhaust. Valve-setting. — A slide-valve is commonly set to give equal lead, or else equal cut-off. Sometimes the leads are made un- equal, so as to partially equalize the cut-off ; in this case the jnethod of setting is like that used for equal lead, except that the lead at each end is made the amount determined on. If the cut-off is equalized by aid of a rocker or bell-crank lever, as shown on Plate VI, the valve is set to give equal lead. As a preliminary to the setting of the valve, a method will be given for putting the engine-crank on the dead-centre. PLAIN SLIDE-VALVE. 29 To put an engine on the dead-centre. — In Fig. i, PI. VIII, let the circle C^CC^C be the path of the crank, and let A^A^ be the stroke of the cross-head ; while abed represents the edge of the fly-wheel or the face of the crank-disk, if the crank is so made. Set the engine with the cross-head near the middle of the stroke, and make reference-marks or take measurements so that it may be set again in the same position. Make a ref- erence-mark on the circle abed, and on some fixed object, at a and 0. Turn the engine round till the cross-head again comes to A ; the crank will then be at C\ and the mark made at a will be found at c. Make another mark at a, and bisect the arcs abc and adc at b and d. It is apparent that the angular distance of b from a is equal to the angular distance of C from C^ ; consequently the crank will be at the dead-point C^, if the mark at b is brought opposite 0. Also the crank will be at the dead-point C^ when d is brought opposite 0. In this process, and during all the operations of valve-set- ting, the engine and the valve-gear should always be moved in the direction in which the engine is intended to run, so that the lost motion or back-lash may be tal<:en up in the right way. Should the engine or the gear be moved too far at any time, then it should be turned back beyond the desired point, and brought up to that point with a motion in the right direction. Should the elasticity of the engine-belt interfere with the con- venient and accurate setting of the engine, it may often be possible to place a stick of timber under a fly-wheel arm, block up one end and place a jack-screw under the other, and so force the engine to the desired setting and hold it at will ; or some equivalent device may be used. To set a valve with equal lead. — First method. — Set the engine on a dead-point and give the eccentric the proper angular ad- vance, as near as may be ; making it too much rather than too little. Adjust the length of the eccentric-rod or of the valve- spindle to give the valve the proper lead. Move the engine forward to the other dead-point, and measure the lead ; if it is 30 VALVE-GEARS FOR STEAM-ENGINES. not right, then correct half the error by changing the length of the valve-spindle, and the other half by moving the eccen- tric. Repeat till the result is satisfactory. If a valve-gear has a rocker, then the length of the valve- spindle should be such that the rocker may swing as designed ; usually to an equal angle on each side of a perpendicular through its axis, to the centre-line of the eccentric-rod motion. In such case the eccentric-rod only should be changed in setting the valve ; a small change of the valve-spindle may be allowed. Second method. — A valve that has harmonic motion will give the same maximum port-opening when set with equal lead. Such a valve may be set for equal lead by the following method. Valves which do not have harmonic motion cannot be so set ; as examples may be quoted a slide-valve having equal lead and with the cut-off equalized by aid of a rocker or bell-crank lever, and a valve controlled by a link-motion or radial valve-gear ; the two last forms will be described in future chapters : Loosen the eccentric on the shaft, and turn it around till it gives the maximum port-opening first at one end and then on the other. If the maximum port-openings are not equal, make them so by changing the length of the valve-spindle by half the difference; this operation adjusts the length of the valve- spindle. When that adjustment is complete, set the engine on a dead-point and give the valve the proper lead by turning the eccentric on the shaft ; this adjusts the angular advance. This method is convenient when it is difficult to turn the engine. To set a valve for equal cut-off. — With the crank on the head-end dead-point, give the eccentric the proper angular ad- vance, and give the valve the proper lead. Move the engine forward till cut-off occurs, and measure the displacement of the cross-head from the beginning of the stroke. Move the engine forward, again, till cut-off occurs on the return stroke, and measure the displacement of the cross-head from the crank end PLAIN SLIDE-VALVE. 3 1 of the stroke. Should the cut-off be earHer at the head end than at the crank end, the valve-spindle is too long ; and con- versely it is too short if the crank-end or return-stroke cut- off is the earHer, In either case, change the length of the valve-spindle by an amount which it is estimated will correct the inequality ; it may be convenient to draw a valve-diagram to aid in making an estimate for a large engine. Set the engine again on the head-end dead-point, and adjust the lead by moving the eccentric. Try the cut-off again, and repeat till the result is satisfactory. It is apparent that a valve that is designed for equal cut-off Avill be properly set if the leads are made what the design gives for them. When such information is at hand, the process of setting the valve will be the same as the first method except that the lead at each end is to be made the proper amount, with the addition that the displacement of the cross-head is to be determined for each stroke, and the adjustment is to be com- pleted by the method just given. To set a valve with the steam-chest cover on. — Figs. 2 and 3, PI. VIII, show an arrangement by aid of which the valve may be set with the steam-chest cover on, and, if convenient, with steam applied to the engine. In preparation the valve is set to give the proper lead at each end, and a centre-punch mark IS made on the valve-spindle outside of the stuffing-box. A pair of trams are made of heavy wire, and adjusted so that they shall reach from the mark on the valve-spindle to a mark on the steam-chest, one at one dead-point and one at the other. In setting the engine, the first method for equal lead is to be used, with the difference that the valve is set to give the lead, when the engine is on a dead-point, by aid of the trams. A varia- tion of this method maybe used, for which one tram is required and two centre-punch marks are made on the valve-spindle. The method is convenient for locomotives, and it is customary to provide trams for this purpose in the tool-chest of the loco- motive, so that the valves may be set, if necessary, on the road. 32 VALVE-GEARS FOR STEAM-ENGINES. It should be noted that the method of setting the valve with equal lead or equal cut-off insures that the action of the valve shall be what is desired when opening or closing. Any error of design due to the neglect of the angularity of the eccentric-rod is therefore transferred to some other part of the motion of the valve, namely, to a place where the valve is open or closed, and any irregularity of motion is then of little con- sequence. CHAPTER 11. ADJUSTABLE ECCENTRICS. Reversing with Loose Eccentric. — The device shown by Fig. 4, PI. VIII, was used for reversing some of the earliest locomotive and marine engines ; it is to-day used for reversing small and unimportant engines, and with some modifications to secure positive action, is used on engines of considerable power. As shown, the crank is at C, and the eccentric has its centre at E, so that the engine will run in right-handed rotation as shown by the arrow. The eccentric is loose on the shaft and has a pin at B, which engages with the end of a circular slot in a disk back of the eccentric, so that the eccentric is driven by the disk. To reverse the engine, it is stopped, and the eccentric or the engine is turned till the pin B engages with the other end of the slot at B\ The valve- circles for forward and for backward motion are drawn at OP and OP', and the lap-circle is nn^n^ ; the cut-off occurs at OR, on the forward stroke when running right-handed, and at 0R\ on the forward stroke when running reversed. The valve-circles for the return stroke are omitted to avoid confusion. Shifting Eccentric with Variable Lead. — A shifting eccentric, like that shown by Fig. i, PI. X, may be used for reversing an engine, and it possesses also the property of giving a variable cut-off. The eccentric is swung on a pivot S, a point on the centre-line of the eccentric and eccentric-rod motion, and is slotted to clear the shaft ; the angle OSO' is made equal to ESE', so that the centre of the eccentric may be brought to the point E' when the engine is reversed. Let 33 34 VALVE-GEARS FOR STEAM-ENGINES, the lap of the valve be equal to Ob ; then the displacement of the valve when the engine is on a dead-point is Oa^ found by drawing the vertical EaE' , and the lead is ba. In Fig. 2 the valve-circle OP represents the valve-motion when the eccentric-centre is at E. The cut-off occurs at the crank-position OR^ or at the piston-displacement xa, assuming harmonic motion ; with crank and connecting-rod the piston- displacement will be longer at the forward-stroke cut-off, and shorter at the return-stroke cut-off, but in such case xa is nearly the mean for the two strokes. When reversed the motion of the valve will be represented by the dotted valve- circle OP' . Suppose now that the eccentric-centre is shifted to E^ , Fig. I ; the angular advance is YOE^ , and the eccentricity is OE^, The valve-circle OP^ will represent the motion of the valve ; it has the angle YOP^ equal to the angular advance, and the diameter OP^ equal to the eccentricity when the eccentric- centre is at E^ , Fig. i. The point P^ is evidently on an arc of a circle having its centre on XOX^ produced, and drawn with a radius equal to ES, Fig. I. In like manner the valve-circles OP^ and OPq represent the motion given to the valve when the centre of the eccentric is at E^ and at E^, respectively. The piston-position at cut-off for the valve-circle OP^ is a^ ; the cut-off for the valve-circle OP^ is at a^ , and for the valve- circle OP^ is at a^. Thus it appears that the cut-off may be made to vary from the piston-displacement xa to the piston- displacement xa^ ; that is, from ^ to ^ oi the stroke. All the other events of the stroke, namely, compression, release and admission, vary at the same time as the cut-off, and in a simi- lar manner, though to a less degree. Of these other events, the admission varies the least ; an examination of the figure will show that the lead-angle increases from about 2° to about 40°. The lead increases from ba to bE^ , Fig. i, and the increase is as clearly shown by Fig. 2. A circle representing the inside lap would, if drawn, show the change of release and compression ; ADJUSTABLE ECCENTRICS. 35 it is omitted to avoid further complexity. Such a circle would show that the compression varies less than the cut-off, and that the release varies more than the admission. It is appar- ent that the increase of the lead depends on the radius SE (Fig. i) of the arc on which the eccentric-centre moves, and may be diminished by moving the pivot 5 away from the axis of the shaft. The gear is said to be in full-gear forward when the eccen- tric is at OE^ Fig. i, and at full-gear backing when the eccen- tric is at OE' , When the eccentric is at OE^ the gear is said to be at mid-gear; intermediate positions may be called grades. An examination of Fig. 2 will show that, since the centre of the valve-circle OP^, is on the axis XOX' , the crank-angle at admis- sion is equal to the crank-angle at cut-off, and this, with other considerations, will indicate that the mid-gear position of the eccentric does not give a proper motion of the valve for either forward or backing motion of the engine. Shaft-governor. — At the present time the shifting eccen- tric is widely used on high-speed engines, which have the valve under the control of a shaft-governor. Fig. i, PI. XI, shows the valve-gear of the Straight Line engine. The line XX' is the centre-line of the piston, crank and connecting-rod ; xx' is the line of the valve-spindle. The cut-off is equalized for one grade by aid of a rocker, and is not far from equal at other grades. The eccentric is pivoted at S, so that OS is the centre-line of the eccentric and eccentric-rod motion. The rocker oscillates on a centre at 2", and the arms Ta and 7"C are connected, respect- ively, with the valve-spindle xx' and the eccentric-rod CE, The governor consists of the weight Wand the spring LQ. The governor-lever WNM is pivoted at iV, and is connected by the link MVL to the eccentric casting at F, and to the five-leaved spring QL at L. When the engine is at rest the weight W lies against the fly-wheel boss as shown, and the eccentric is at full-gear. When the engine comes up to speed, the centrifugal force acting on W enables it to compress the spring QL, and 36 VALVE-GEARS FOR STEAM-ENGINES. at the same time to move the eccentric toward mid-gear, and thus the governor adjusts the cut-off, and consequently the steam-supply, to the load. Since the engine is never reversed, the eccentric is slotted to clear the shaft only far enough to bring its centre to mid-gear. Fig. 2 shows the valve-diagrams for this gear drawn to an enlarged scale. In order that a shaft-governor may be able to control the valve of an engine, and maintain the speed nearly uniform, without being of excessive size, the valve must be nicely balanced and must move very freely. The Straight Line engine has a flat valve that moves between its seat and a cover-plate with enough clearance to avoid friction, but not enough to allow of leakage, just as a piston-valve moves in its cylindrical seat. A device similar in effect to the passage through the body of the Allen valve gives a double admission at and near cut-off and admission. The piston-valve, usually without packing-rings, is commonly used in connection with the fly-wheel governor. The action of the reciprocating parts of a high-speed engine is of great importance. A considerable part of the work of the steam is expended in imparting motion to the reciprocating parts during the first half of the stroke, and this stored energy is restored during the second half of the stroke as the recipro- cating parts come to rest. In order that they may come to rest quietly at the end of the stroke the piston should be cushioned by compression. Now a valve that gives a variable cut-off and a variable compression is likely to have too little compression at full-gear and too much at a short cut-off. An engine with a large clearance will suffer less from this difficulty than one with a small clearance ; consequently the clearance of high-speed engines with shaft-governors is often made large, but a large clearance is not conducive to economy in the use of steam. Now, lead acts, as does compression, to stop the reciprocating parts and to fill the cyHnder with steam, so that in general the more compression an engine has the less lead ADJUSTABLE ECCENTRICS. 37 it will need. But it has just been seen that the shifting eccen- tric shown on Plate X has an increasing lead toward mid-gear, that is, at the time when it is least needed on a stationary engine. Had the pivot 5 been placed on the other side of the shaft, then the lead would have decreased toward mid- gear. Fig. I, PI. XII, gives the valve-diagrams for such an eccentric. Some forms of the Straight Line engine have their valve-gears so arranged in order that the decreasing lead toward mid-gear may partially compensate for the increasing compression. Shifting Eccentric with Constant Lead. — Fig. 2, PI. XII, shows an eccentric that has a motion square across the shaft, thus carrying the centre of the eccentric on the straight line EE^E' from full-gear forward to full-gear backing. It is at once apparent that the lead is constant. Fig. 3 gives the valve- diagrams for the full-gear, mid-gear, and two intermediate grades. Though the lead is constant, the lead-angle is not so ; a comparison with Fig. 2, PL X, will show that the variation is not so much as that of the lead-angle for an eccentric with increasing lead, but it is more than for an eccentric with de- creasing lead, as may be seen by a comparison with Fig. i, PL XII. A shifting eccentric with constant lead must be slotted to clear the shaft ; the line 00' is made equal to EE\ in order that E may pass to EE' when the engine is reversed. If the engine runs always in one direction, the slot, from centre to centre, may be only as long as EE^. The shaft-governor of the Armington and Sims engine, shown by Fig. i, PL IX, is equivalent to a shifting eccentric with constant lead. The eccentric ^ is on a casting, ab, loose on the shaft ; the eccentric E^ is loose on the eccentric E. The weights W and V compress the springs N and My and move out toward the rim of the wheel as the engine comes up to speed. The links ac and bd take hold of the eccentric-casting ab^ and the link st takes hold of the eccentric ^,. As the 38 VALVE-GEARS FOR STEAM-ENGINES. weights move out the linkage cabd takes a new position, such as ^V^'^', turning the eccentric E toward the left; at the same time the link st takes the position s'f^ turning the eccen- tric E^ toward the right. The proportions of the mechanism are so chosen that the centre, e, of the outer eccentric, E^y moves straight across the shaft to a new position, e', and the cut-off is shortened. It will be seen that the effect is the same as that of the shaft-governor of the Straight Line engine, ex- cept that the lead is constant. CHAPTER III. LINK-MOTION. The valves of locomotives, marine engines, and other re- versing engines are commonly controlled by a mechanism called a link-motion ; this mechanism has also the property of giving a variable cut-off. The mechanism consists essen- tially of two eccentrics, one for full-gear forward and one for full-gear backing, together with the eccentric-rods and the link. The eccentric-rods are attached to the link, at or near the ends, and the link is slotted or otherwise arranged to receive a block on the end of the valve-spindle, on a radius- rod, or the end of a rocker, as the case may be. The link- motion takes two forms ; in one, known as the Stephenson or shifting link, the link is moved on the block to reverse the engine or vary the cut-off ; in the other, known as the Gooch or stationary link, the block is moved in the link to accompHsh the same object. Stephenson Link. — The usual form of Hnk-motion for American locomotives is shown by Figs, i and 2, PI. XIII. The valve is moved through a rocker so that the eccentrics follow the crank ; thus, the centre of the crank is at C, and the go-ahead eccentric has its centre at E, while the backing ec- centric centre is at E' , The link-pins P and P' , to which the eccentric-rods are attached, are set back from the link-arc, and the link may move over the link-block B so far as to bring the centre of the block opposite the centre of the link-pin, as shown by Fig. i ; in which position the motion of the valve is 39 40 VALVE-GEARS FOR STEAM-ENGINES. controlled almost entirely by the eccentric E, and has essen- tially the motion of a plain slide-valve. The link is suspended by a link or hanger, nm^^ from a reverse-shaft centred at S; the hanger takes hold of the saddle-pin m^ on a plate that is com- monly at the middle of the link. A locomotive has two cylinders, with pistons acting on two cranks set at a right angle, and thus has two engines each of which must be provided with its own link-motion. Both links are suspended from one reverse-shaft, which has an arm SR from which a rod runs to a reverse-lever conveniently located in the engineer's cab. The reverse-lever moves over a notched arc, and by aid of a latch engaging with the notches the link may be set and secured in any desired position. Fig. 2, PI. XIII, gives an end elevation of the link, the hanger, and one arm of the rocker carrying the link-block B. English locomotives commonly have the link act directly on the valve-spindle, without the intervention of a rocker. In such case the link-pins should be placed on the link-arc, as shown by Fig. 3, PI. XIII, and consequently the link-block cannot be opposite one of the link-pins and cannot receive the full motion of the eccentric. Consequently the eccentricity, and with it all the dimensions of the link-motion, must be larger to give proper motion to the valve. The link-motion for the high-pressure cylinder of one of the engines of the U. S. S. Maine is shown by the figures on Plate XIV. The parts are lettered as on Plate XIII ; thus E and E' are the centres of the eccentrics, and P and P' are the link-pins, which in this case are on the link-arc ; vS is the reverse- shaft, and iVPis the drag-link or bridle which takes hold of the go-ahead link-pin. The link, which is known as the Scotch or side-bar link, is shown in plan by Fig. 2. The link-block is between the side-bars, and is pivoted directly on the end of the valve-spindle ; thus the link can be set so that the axis of the link-pin coincides with that of the link-block pivot, and the full motion of the eccentric can be given to the valve. The LINK- MO TION. 4 1 head of the valve-spindle is guided by cast-iron jaws, as shown in Fig. I. The end of the reverse-arm is slotted and provided with a sHding-block, screw, and hand-wheel as shown, so that the cut-off may be adjusted in a manner to be described later. Open Rods and Crossed Rods. — If the eccentric-rods of a link-motion are connected as shown in Fig. i, PL XV, the rods are said to be open ; on the other hand, the rods are said to be crossed when connected as shown by Fig. 2. In both figures the crank is on the crank-end dead-point, and the valve- gear has no rocker. A link-motion with a rocker is said to have open rods when the eccentric-rods are connected as shown by Fig. 3; aTb is the rocker, and be the valve-spindle-, the crank is on the head-end dead-point. In all these figures the link-pins are on the link-arc. Since half a revolution will apparently cross the rods for Figs, i and 3, while it will appar- ently open the rods for Fig. 2, the nomenclature seems to be unfortunate. There is, however, a real difference in the methods of the connection of the rods, and that difference has an im- portant influence on the action of the valve, for open rods give an increasing lead from full toward mid gear, while crossed rods give a decreasing lead from full toward mid gear. In Fig. I, the full lines Ep^, E'pJ, and the arc/^^/, show the eccentric- rods and the arc of the link at mid-gear, while the thin lines Ec, E'p' , and the arc cp' , show them at full-gear forward. Since the valve and valve-rod have the same motion as the link-block, it will be sufficient to trace the motion of the latter. At full- gear the link-block will be at c^ found by intersecting the line of centres with ^ as a centre, and with a radius equal to the length of the eccentric-rod. The eccentric-pin p' is located by drawing arcs from E' and c, with the lengths of the eccentric- rod and the length of the link as radii. At mid-gear the link- block is ac^\ the points p^ and // are at a distance from the centre-line OX^ equal to half the chord of the link-arc, and the link is erect. The increase of lead from full-gear to mid-gear is apparent from the diagram. A similar construction in Fig. 42 VALVE-GEARS FOR STEAM-ENGINES. 2 shows the decrease of lead from full-gear toward mid-gear for crossed rods. In the figure the decrease is greater than the full-gear lead, so that the valve is shut at the dead-point when the link is at mid-gear. Long and Short Rods. — The variation of lead from full- gear toward mid-gear is due to the curvature of the link-arc, and is more pronounced for a link with short radius than for one with long radius ; now the radius of the link-arc is usually equal to the length of the eccentric-rod, hence the variation is more for short than for long rods. In Fig. i, PI. XV, it is apparent that c"c^' is greater than cc^ ; a similar construction will show that the decrease of the lead from full-gear to mid-gear for crossed rods is more marked for short than for long rods. Radius of the Link-arc. — An analytical discHssion of the link-motion shows that the radius of the link-arc should be equal to the length of the eccentric-rod ; if the link-pins are on the link-arc, then the radius should be the distance from the centre of the eccentric to the link-pin ; but if the pins are back of the arc, the radius is the distance from the centre of the eccentric to the link-arc, i.e. the length of the rod plus the distance the pins are back of the arc. The same discussion establishes also the fact that open rods give increasing lead, and that crossed rods give decreasing lead, from full-gear toward mid-gear ; but the demonstrations given are believed to be useful, and a similar demonstration will be given of the proper radius for the arc. In Fig. I, PI. XV, the link-block is at c at full-gear when the crank is on the crank-end dead-point ; when the crank is on the head-end dead-point a similar construction will give for the position of the link-block the point c' , The point Oy half way between c and c' ^ corresponds to the mid-position of the valve, and from o the lap on^ on' may be laid off on each side, giving nc = n'c' for the lead. At mid-gear the head-end lead is nc^ , and a similar construction for the head-end dead- point will give n'c^ = nc^ for the head-end mid-gear lead. If LINK-MOTION. 43 now a diagram is drawn for some intermediate gear of the link, it will be found that the lead is the same at the two ends, and that it is intermediate between nc and nc^. Fig. i is drawn with the radius of the link-arc equal to the length of the eccentric- rod, and any diagram drawn with dimensions chosen at random will in Hke manner show equal leads under like conditions. Constructions for a link with crossed rods will also show equal leads if the radius of the Hnk-arc is equal to the length of the eccentric-rod. Consequently it may be inferred that the one requirement for equal leads is that given, i.e., that the radius of the link-arc shall be equal to the length of the eccentric- rods. A natural inference is that any other radius for the link- arc will give unequal leads for some grades of the link, and such will be found to be the case if constructions are made. Analytical Discussion. — The best idea of the nature of the motion given by a link to the valve is obtained from an analytical discussion due to Zeuner, taken, with slight variations, from his Treatise on Valve-gears. On Plate XVI the Figures i and 2 are drawn to represent link-motions with open and with crossed rods. The diagrams in thin lines give the positions of the parts when the crank is on the crank-end dead-point ; and the diagrams in heavy lines show the positions when the crank has moved through the angle d. The eccentricity for each eccentric is r^ and the angular advance is S. The link-pins are on the link-arc, the length of the eccentric-rods is /, and the radius of the link-arc is p which may or may not be equal to /. The length of half the link-arc is Cy and the displacement of the link-block from the middle of the link-arc is d. In the discussion it is assumed that the link is supported and guided by the link-block so that the point n remains on the line XX' . It is also assumed that the chord joining the link-pins is equal to the length of the link-arc between those pins, and that in like manner the displacement of the link-block from the middle of the link may be measured 44 VALVE-GEARS FOR STEAM-ENGINES. indifferently either on the link-arc or on the chord. The error of this assumption may be estimated as follows : A common proportion is r = J^ = y^^/, or ^ = |^/, so that the arc subtends an angle of about lo°, and for that angle the arc is 0.1745 of the radius and the chord is 0.1736 of the radius, and the error is a little more than half of one per cent. The distance from the centre of the driver-axle to the middle of the valve, in either Fig. i or Fig. 2, PI. XVI, is Ob = Om -{- mn ■\- nb ^^ Op — mp -f- mn -\- nb, , (7) in which the length of the valve-spindle iib may be replaced by Sy and the value of the other terms may be conveniently deter- mined as follows : First, the term mp is determined by the equation mp = mP sin mPp = (c — d) sin a, , » , (8) Now pp' Op - Op' sm« = ^, = ' ^9> Prom Fig. i, PI. XVI, Op=Oe^epz:^Oe-V'(EP' '-(Pp'-'Eef\''\ . (10) and from Fig. 2, 0p=.0e^ep=0e^ \EP - (/> + Eey\\ . (il) LINK-MOTION. 45 In either figure Oe = r sin {S -\-d\ Ee — r cos (B -\- d), EP = /, and Pp =z mP cos mPp ^^{c — d) cos a ; these values substituted in equations (10) and (11) give Op = rs\n{e-\-d)-[- {P — \{c — a)cosa ^ r cos(e-{-d)J\i', {12) the upper sign being taken for open and the lower sign for crossed rods. Expanding the term with a fractional exponent by the binomial theorem, and rearranging terms with the higher powers of / in the denominator, gives (c — dY cos' a Op = rsm{e + S)-\-l-^ ^^ (^ _ d)r co s {6 + S) co s a r" cos ' {B + d) ± 1 H Yl Now the terms containing cos a in the numerator have / in the denominator, and are small compared with r sin {6 + 6\ while a is not more than 30°, for which the cosine is 0.866 ; consequently we may replace unity for cos a without much error. With that change and some expansion, c' cd d^ op = r sm(e^ 6) + I --^^-j--^ (c -d)r cos {O + S) r" cos' {S + 0) 46 VALVE-GEARS FOR STEAM-ENGINES. Similarly Fig. i and Fig. 2 give respectively equations (14) and (15): Op' = e'p' - Oe' = \E'P' - {Ff - E'eJY - Oe' ; (14) Op' = // - Oe' = \E^" - {P'p' + E'e'y\^ - Oe\ (15) In either figure 0/ = r sin {6 - 6), E'e' = r cos (B - S\ E'F = /, and P'p' = mP' cos mP'p' z=z{c-\- d^ cos a ; these terms substituted in equations (14) and (15) give (9/= -rsin(^ - ^)+ {/'-[(^+^) cos a':^r cos(^-d)]'}* ; (16) the upper sign being taken for open and the lower for crossed rods. As in the previous work, the cos a may be replaced by unity ; expanding by the binomial theorem, and rejecting terms? with the higher powers of / in the denominator, gives Op' = ^r sin (B.-d) J^l- ^'^J^' {c + d)r cos {B — 6) r" cos' {6 — S) ^ 7 Yl ^/="--(^-^)+^-S-?-S (c + d)r cos (6' - S) _ r" cos" (B ^ S) LINK-MOTION. 47 Substituting in equation (9) the values for Op and Op\ r sin id -\-&)-^r sin {S — d) 2cd sma = + ^ (c -d)r cos {d-\-6)-{c-{- d)r cos (^ — 6) ^ 2cl r' cos" {d-\-d)- r' cos' ( 6- d) T T dt .'. sin a = — cos (^ sin ^ =F -7 sin (5^ sin <9 ip -- cos S cos ^ + 7 - ^/cos" (« + d) - cos' (^ - d)]. (18) To find the value of the term mn^ which in Fig. 3, PL XVI, is seen to be nearly equal to mj,^ we have mn = m^i = m^n^ — tn^ ; Pm, m c ^ / 1 X f . /. mn = — Tf. — T^ = (nearly). . (10) Substituting in equation (7) the values of the several terms, c' cd d" Ob = r sin S cos 6 -\- r cos ^ sin 6 -\- 1 J ~\~~T j K^-^). a ^ ' n ■ ^x ^'C0s'(d+^) ± -^ — -^cos ^ COS ^ — sm sm d) /^ if T dv — COS S sin ^ =F y sin d sin ^ =F — 7 cos d cos 6^ + / - ^/ 1"°'' (^ + <^) - <^°^" (^ - <^)3 } + — - — + ^; 48 , VALVE-GEARS FOR STEAM-ENGINES. Ob = r[s'm d ± — i — cos ^ ± (^ — d)—^ cos d] cos fi c — d , - c — d ^ c — d . «-,.„ -Y ^Lcos o =F — ^^ — sin o cos o ± — - — sin o] sin u - -^ I 2^ cos^ (p^-Q)-{c- d) [cos'((9 + <^) - cos' (O-d)-] I — ^'p -|- 2cdf> — d*f> — 2(<; — d)dp + ^'^ — <^V + Wp + ^+^; / c" —d" \ d ,\ Ob = Asm d ± — cos dlcos -\- r— cos S sin 6 - ^jiif + d) COS' {S + e) + {c-d) cos' (^ - LINK-MOTION. 49 If the engine is on the crank-end dead-point, then B is zero ; and it is 180° at the head-end dead-point. The special values of the crank-angle give Ob' = r[sm S ± --^j- cos d) + {/ - d'')-^ + /+ ^ ; (22) Ob" = - r[sm d ± ^--^ cos c^) + {c^ - ^)^ + l+s. (23^ The mid-position of the valve should be midway between b' and b", Figs, i and 2, PL XVI ; half the sum of Ob' and Ob'' is *=^^^ = (--^/-ir"+'+>. . ,,) in which the only variable is the term containing p, the radius of curvature of the link-arc. If p be made equal to /, then this term disappears, leaving Oo = /+s (25) Wttk equal laps^ the necessary and sufficient condition for equal leads, at all grades, is that the radius of the link-arc shall be equal to the length of the eccentric-rod. Applying this condition to equation (21) gives / c^ — d^ \ d Ob = r( sin d ± ^ — cos djcos B -\-r— cos d' sin B -\- 1-\- s, (26) 50 VALVE-GEARS FOR STEAM-ENGINES. The displacement of the valve from mid-position is e=Ob— Oo\ ,\ e = z*! sin S ± — cos 8\ cos 6 -{- r— cos 8 sin Q. (27) General Equation for Valve-motion. — The equation (6) gives for the displacement of a plain slide-valve moved by an eccentric, e •= r sin {p -\- ^'i expanding the parenthesis, e =L r cos d sin 6 -\' r sin 6 cos 6^ ; , , , (28) which may be written e = A cos 6 -\- B sin 6, (29) since r and S are constant for a given slide-valve gear. It has been shown by the aid of Fig. 7, PI. II, that the motion of a plain slide-valve may be represented by a valve- circle, and a comparison of that figure with the equation (29) will show that the constants in the equation are the coordi- nates of the end P of the diameter of the valve-circle. Thus Og = rsind = A; (31) Pq z=r cos d = B , (32) LINK-MO TION. 5 1 It may be concluded that any valve which has its displace- ment represented by an equation of the same form as equation (29) has a harmonic motion and may have its motion repre- sented by a valve-circle. Zeuner's Diagram. — A comparison of equation (27) with equation (29) shows that a valve controlled by a Stephenson hnk-motion has a harmonic motion, and that its displacements from mid-position, at any grade of the link, may be represented by a valve-circle, having for the coordinates of the end of the diameter A = r\sm ± ——^ — cos ^j ; .... (33) ^ = r-cos6' (34) At full-gear d =^ c, which applied to equations (33) and (34) will reduce them to equations (31) and (32). Thus the valve- diagram for a link-motion at full gear is identical with the dia- gram for a plain slide-valve under the control of an eccentric having the same eccentricity and angular advance as one of the eccentrics of the link-motion ; which coincides with our conceptions of the link-motion, derived from the drawings on Plates XIII and XIV. In Fig. 4, PI. XVI, and Fig. i, PI. XVII, Oq and ^Pare made equal to the coordinates of the end of the valve-circle diameter at full-gear, when d=^ c\ i.e., Og = r sinS = Af qP = r cos d =1 B, At mid-gear d becomes zero, and the coordinates the end of the diameter of the valve-circle become A^=irsmd ±jC0s6\ (35) ^0 = (36) 52 VALVE-GEARS FOR STEAM-ENGINES. The upper sign is taken for open rods, and the lower sign for crossed rods. Fig. 4, Pi. XVI, corresponds with the first case, and Fig. i, PI. XVII, with the second case. The ends of the diameter of valve-circles for intermediate grades of the link may be found by assuming values for d and calculating the co- ordinates by aid of equations (33) and (34), if desired ; but an inspection of those equations shows that they give the co- ordinates of a parabola having its vertex on the axis XX\ Two points, P and P^^ are already located and an arc of the parabola may be passed through them by the ordinary geo- metrical construction ; or, since the arc is quite fiat, there may be substituted for it the arc of a circle having its centre on the axis XX' . The centres of the valve-circles have for their co- ordinates, \A and \B, and consequently lie on another para- bola with its vertex on XX' ; an arc of a circle centred on XX' may be substituted for the arc of the parabola, and will be more convenient to draw since its radius is half the radius of the circular arc substituted for the parabola through P and P.- Fig. 4, PL XVI, and Fig. i, PL XVII, exhibit the variation of lead which was pointed out in Figs, i and 2, PL XV. The same thing is evident from an inspection of equation (33), taking the upper sign for open and the lower sign for crossed rods. For convenience the fact is stated as follows : A Stephenson link-motion with open rods gives increasing lead from full-gear toward mid-gear ; with crossed rods it gives de- creasing lead from full-gear toward mid- gear. Valve-circles showing the motion of the valve at inter- mediate grades of the link are drawn at OP^ and OP^ , on Fig. 4, PL XVI, and on Fig. i, PL XVII. On both figures the lap- circles are nn'n" y showing cut-off at the crank-positions ORy ORy^ , ORf y and OR^ ; neglecting the influence of the connect- ing-rod, the corresponding piston-displacements are xa, xa^ , xa^ y and xa^ . An inspection of the figure will show that the LINK-MOTION, 53 lead-angle increases as the cut-off is shortened, accompanied by an earlier admission. If an inside lap-circle were drawn, it would show that an early cut-off is accompanied by an early release and a large compression. A comparison of these dia- grams with the valve-diagrams shown by Fig. 2, PI. X, and Fig. I, PI. XII, will show that a Stephenson link-motion is equivalent to a shifting eccentric with variable lead. Gooch Link. — Fig. i, PL XVIII, shows the Gooch or sta- tionary link, as applied to locomotive engines. E and E' are the forward and backing eccentrics, from which the eccentric rods lead to the link-pins P and P' , The link is suspended by the link mUy from a fixed pivot n^ and has its convex side turned toward the axle O. The link-block B is carried by a radius-rod BD, which is connected directly to the head of the valve-spindle at D, By means of a reverse-arm ST and hanger TU, the engineer may place the link-block B opposite the link- pin P for full-gear forward, opposite the link-pin P' for full- gear backing, or at any intermediate position. The action of this link-motion is therefore equivalent to that of the Stephen- son link-motion. The details of the mechanism are varied somewhat by different makers. In the figure the link is sus- pended from a saddle-pin on or near the chord joining the ends of the link-arc ; and for this purpose a plate or bridge similar to that shown by Fig. 2, PL XIII, is employed which permits the passage of the link-block. The saddle-pin is some- times placed behind the link-arc toward O, so as to avoid the use of a plate or bridge. The link-pins are placed behind the link-arc to allow the link-block to be brought opposite the link-pins. They may be placed on the link-arc, using a link like that shown by Fig. 3, PL XIII, but turned so that the convex side is toward the axle O ; in which case the full action of the eccentric cannot be given to the valve. Sometimes a box-link, shown by Fig. 3, PL XVIII, is used, and then the saddle-pin and link-pins may be placed in any desired positions 54 VALVE-GEARS FOR STEAM-ENGINES. without interfering with the Hnk-block ; this device is equiva- lent to the side-bar link shown on Plate XIV. Open and Crossed Rods. — As was found to be the case with the Stephenson link, the rods of the Gooch link may be open or crossed. Fig. i, PL XVIII, has open rods, and Fig. 2, PL XVII, has crossed rods. Radius of the Link-arc. — The common and proper prac- tice is to make the radius of the link-arc equal to the length of the radius-rod ; and when so made the lead is constant for all grades of the link. This property is at once evident from in- spection of Fig. I, PL XVIII, and Fig. 2, PL XVII, one having open and the other crossed rods ; for it will be seen that when the engine is at a dead-point and the link is erect, the link- block may be moved from one end of the link-arc to the other without moving the valve. Analytical Discussion. — Making use of the same notation as in the analytical discussion of the Stephenson link, let e be the eccentricity, and S the angular advance for each eccentric ; let c be the half-length of the link, and d the displacement of the link-block from the middle of the link ; let / be the length of the eccentric-rod, the link-pins being assumed to be on the link-arc ; let /, be the length of the radius-rod, and s the length of the valve-spindle. Assume that the link is so suspended that m, the middle point of the chord, shall remain on the central line XX' y and that the length of the link is sensibly the same whether meas- ured on the chord or on the arc. Assume also that the rods are open. In Fig. 2, PL XVIII, the diagram in fine lines represents the link-motion when the crank is on a dead-point ; and the diagram in heavy line, represents it when the crank has moved iYOvaC^ to (7 through the angle 6, The distance from the origin to the middle of the valve b is Ob = Op-pk^kq-\-qS-^Sb (37) LINK-MOTION. 55 The term//^ is determined by the equation pk = KP sin pPK ={c — a) sin or. (38) Now pp' Op - O p' . . sin or = -^ = (39) But Op =^ Oe -{-ep = Oe -^\EP ~(Pp-Eey\\', . (40) Op' =-0e'-\- e'p' = - 0/ + {£'r - {Py - Eey\K (41) in which EP=E'F=l, Pp = Pp' = ccosa; Oe = r sin {d + d), Oe' = r sin {0 — S)\ Ee = r cos (^ + S\ Ee' = r cos {0 - ^). Substituting these values in equations (40) and (41) gives Op = r sin (d -\^ d)-\-\P -\c zo^a - r cos {6 + S)-\'\\ ; (42) Op' z=-r sin (6> - d) +|/» _ [c cos or - ^-cos (^ - d)]''}i (43) A comparison of equations (42) and (43) with equations (12) and (16) shows that they differ in that the coefficient of cos a does not contain d, and that only the upper sign of the double $6 VALVE-GEARS FOR STEAM-ENGINES. sign appears before the last term in the bracket ; this last be- cause the discussion applies only to open rods. Consequently the value of sin a may be obtained from equation (i8) by omit- ting terms containing dy and using only the upper sign of the double signs ; hence r r sin or = — cos S sm 6 — -j sin S sin 6 c I -^^[cos'(^+')- LINK-MOTION. 57 To obtain the last equation, it may be admitted that qk (Fig. 2, PL XVIII) is nearly equal to QK^ which is nearly equal to tm. Now Pm is half of a chord bisected by a diam- eter, of which one segment is nm and the other is 2/^ — mn ; consequently Pn^ = nm{2l^ — mn) ; /. mn = -J- (nearly) ; and in like manner nt = —7- (nearly). /, tm= kq = -j J (nearly). Substituting the values of the several terms in equation (37), if f — sin 6^ cos d — y sin 6^ sin d + ^^[cos'(^_^)^cos^(^+ ~ ~7~ + ^i + -^ - J^cos' (^ + ' (^^ zz= r( sin <^ + -T cos Sj cos ^ -| 1 cos ^ -- y- sin 6 VALVE-GEARS FOR STEAM-ENGINES. TABLE V. Link acts directly on valve-spindle without rocker. Link-pins on link-arc. Cut-off equalized at ^ stroke. Saddle-pin \ of an inch forward of link-arc. Distance between link-pins, 12 inches. Cut-ofiE. Lead. Slip. Travel. Port Open- ing. a H.E. b C.E. Diff. a&Lb. 6 6f 8 8 f 3 T6 2t\ 13 "ST 9 9iV 1 T6- f ^ ol3 if 12 I If i s 8 \ .7I9 2^2" 1^ 15 I4f f w 5 T6 2f H 18 i7i 1 1 1 32 3 8 3f H.E. I C.E.I- 21 20A if 7 "3^ i 4t H.E. If C.E. l\ 22i 21^ f -h f 6tV H.E. 2f C.E. 2,% This example was chosen to represent the type of link-motion which acts directly on the valve-spindle without the interven- tion of a rocker, and, as is customary, has the link-pins on the link-arc. It may be compared with Table IV, which has the link-pins the same distance apart. The equalization of cut-off from i to f stroke is good, though not perfect, and the inequal- ity at and near full-gear cannot have a very bad effect, though it is much larger than in the fourth example. The slip is less than for that example at all grades ; but probably the slip could be much reduced in such a link-motion with a rocker, if the link-pins were placed nearer the link-arc or on the link-arc. Such an arrangement would show greater inequality in cut-off than is found in Table IV. The sixth example was chosen to show the effect of placing the link-pins ahead of the link-arc, on a link which acted direct- LINK-MOTION. 87 ly on the valve-spindle without the intervention of a rocker. Such an arrangement would require the use of a side-bar link or a box link (shown by Fig. i, PI. XIV, and Fig. 3, PL XVIII), and might involve some mechanical difficulties. TABLE VI. Link acts directly on valve-spindle without rocker. Link-pins 3 inches ahead of link-arc. Cut-off equalized at \ stroke. Saddle-pin if of an inch ahead of link-arc. Distance between link-pins, 12 inches. Cut-off. Lead, Slip. Travel. Port Open- ing. a H.E. b C.E. Diff. a &. b 6 6f f 3 8 t23 ^3¥ 23V 8 8 9 Q31 ^3 2^ sV f iH 2f 1^ 12 I HI 1 1 6 1 1 t3 1 2f tV 15 I4f i ^ t21 ^32 2if fi 18 I7H 1^ i If 3i 1 21 20H -h 3 TB" If 4f lA 22-1- 22f i h If 6| 2H The equalization of the cut-off must be considered to be satisfactory, but it is attained at the expense of an excessive slip of the link-block. Probably a compromise between the link-motions shown by Tables V and VI, having the link-pins an inch or an inch and a half ahead of the link-arc, would be found to give a fair equalization of the cut-off without exces- sive slip. Also greater distance between the link-pins (14 in- stead of 12 inches) could be used with advantage. CHAPTER IV. RADIAL VALVE-GEARS. The name radial valve-gear has been applied to a number of reversing-gears that differ widely in detail and in general ap- pearance, but agree in that they derive the mid-gear motion of the valve from some source that is equivalent to an eccentric with 90° angular advance, and they combine with this motion another that is equivalent to that of an eccentric with no angu- lar advance. The general conception of this form of valve- gear is most easily obtained from an example. Walschaert Gear. — This gear is chosen as the first ex- ample of the type because the elements are easily distin- guished. In Fig. I, Plate XXII, H is the engine cross-head, and a is the head of the valve-spindle. The valve is moved through a radius-rod, one end of which carries a block that may be set at any position in a slotted link dF, and the other takes hold of a combining-lever af^ that receives motion from the engine cross-head. The slotted link swings on a fixed trunnion at G and is moved by an eccentric OE, which has no angular advance. In Fig. 2 the diagram in thin lines shows the gear at a dead-point, and the diagram in heavy lines shows the gear when the crank has moved through the angle Cfic= e. If the motion of the engine cross-head can be considered to be harmonic, then it is clear that the motion that it gives to the valve could be derived from an eccentric with 90° angular advance ; this motion is made equal to the lap plus the lead. RADIAL VALVE-GEARS. 89 If the block d is at the middle of the link, the valve will derive motion from the cross-head only and the mechanism will be at mid-gear. The radius of the link-arc is made equal to the length de of the radius-rod, consequently the lead is constant for all settings of the gear. If the point h of the guiding-link hf were a fixed point, then the valve would receive motion from the eccentric OE^ which has no angular advance. By placing the link-block nearer the trunnion G the motion is reduced ; for example, the motion communicated from the eccentric OE will be half as much if the block is half-way be- tween d and G, If the link-block is below G the motion is reversed. The displacement from mid-position of a valve moved by an eccentric is ^ = r sin (^ -[- <^). Now the motion derived from the cross-head is equivalent to that from an eccentric having 90° angular advance, provided the cross-head motion is assumed to be harmonic. Conse- quently the valve derives a displacement from this source of e^ = r, sin {6 -|- 90°) = r^ cos d. ... (53) From the proportions of the combining-lever and the length R of the crank, we have ae „ The displacement of the valve from the influence of the eccentric OE is e^ = r, sin {0 -\- 0°) == r, sin 6^, . . . . (54) in which dG af ' GF ef 90 VALVE-GEARS FOR STEAM-ENGINES. The entire displacement e of the valve at any crank-angle is the sum of the displacements from the two independent sources. .-. e = e^-\- e^ = r^ cos 6^ -|- r^ sin 6^ ; . . . (55) and since r^ and r, are constant for any grade of the link, equation (54) is a special case of equation (29), r^ and r^ being the coordinates of the diameter of a valve-circle for that grade of the gear. In Fig. I, Plate XXI, the valve-circle OP^ is drawn with a diameter r^ and represents the mid-gear action of the valve. The circles Op, Op^ , and Op^ represent the motions derived from the eccentric OE^ at full-gear and at two intermediate gears. The circles OP, OP^, and OP^ represent the actual dis- placements of the valve, derived from both sources. It is evi- dent that OP = \0P, + Op [*> and that OP^ and OP^ may be obtained in a similar manner* A comparison of Fig. i, PI. XXI, with Fig. 3, PL XVII, shows that the action of the Walschaert gear is equivalent to that of the Gooch link-motion. To aid in this comparison, the dimen- sions OP^ and P^P were transferred from Fig. 3, PI. XVII, ta Fig. I, PL XXI, and consequently the diagrams are identical. As actually constructed, this gear does not give harmonic motion to the valve, for the motion of the cross-head of the engine with the usual proportions of locomotives has consid- erable irregularity from the angularity of the connecting-rod ; also some irregularity is introduced by the combining-lever a/. Consequently such a diagram as Fig. i, PL XXI, can be of use only in roughly blocking out a gear. The real action of the gear can be determined either by constructing diagrams similar to Fig. 2 on as large a scale as convenient, or by aid of a modeL RADIAL VALVE-GEARS. 9 1 A combination of the two methods, similar to the skeleton model for link-motions, may be found convenient for this pur- pose. Since part of the motion of the valve is derived from the cross-head, the adjustment of the gear to give equal cut-off will generally be easier than for a link-motion. In laying out a Walschaert gear, the combination-lever af should be made vertical when the cross-head is at the middle of its stroke ; the guiding-link hf should be made to vibrate equal angles above and below a horizontal line ; a line from G to F Qw the link should be made vertical when the engine is on a dead-point; and the supporting-link with the reverse-arm ST should be so laid out that it may be guided nearly on a horizon- tal line, unless the adjustment may be found to require a differ- ent arrangement. The length of the combining-lever should be so chosen that its angular vibration shall not exceed 60°. The main dimensions of the gear for any engine will be imposed on the designer by the general proportions of the engine and its frame. There are, however, two elements over which the designer will have more or less control : they are the position of the axis of the trunnion G, which in the figures is on the link-arc, but which may be placed either forward of or back of the link-arc ; and the reverse-shaft T, which may often be located at will, within limits. The first will be found to have the most influence on the action of the gear. The eccentric OE is sometimes replaced by a return-crank from the engine-crank C. The link is sometimes turned the other way, in which case the radius-rod extends forward from the head of the valve-rod. Marshall Valve-gear. — Plate XXIII shows the Marshall valve-gear as applied to the U.S.S. Yorktown. In Fig i, XX' is the axis of the cylinder, O is the centre of the shaft, and C is the crank-pin. The eccentric centred at E gives motion to a short and massive eccentric-rod EG, which is guided at o by the link oDy and is connected to the valve-spindle K by a valve-rod GV, The guiding-link oD is supported by the bell- 92 VALVE-GEARS FOR STEAM-ENGINES. crank lever DoSj having its axis at o ; the rod ST gives con- nection with the reverse-shaft arm TU. In the figure the engine is at a dead-point, so that the guided point of the eccentric-rod coincides in projection with the axis of the bell- crank lever. Fig. 2 shows the centre-lines of the gear in two positions ; the heavy lines are for the full-gear of the valve-motion, and the fine lines are for a gear between that and the mid-gear. OC is the centre-line of the crank ; E is the centre of the eccentric; FD is the guiding link; and oD is the arm of the bell-crank lever, having its axis at o. When the bell-crank lever is set to give the mid-gear action of the valve, D is found at D^ , and the guided point F moves on an arc of a circle that nearly coincides with the line OV; the point E describes a circle, and all other points of the eccentric-rod describe ovals that are more or less elongated as they are near or removed from the guided point F. In this setting of the gear the horizontal motion of the point G is made equal to the lap plus the lead, so that the valve receives a motion like that given by an eccentric having 90° angular advance ; as 6^ is beyond F, the eccentric E properly coincides with the crank. At any other gear than mid-gear, for example with D set for full-gear, the vertical displacement of F will have two -components, one along the axis OV, and one perpendicular to it. The second component with some modification is trans- ferred to G and gives to the valve an additional displacement like that from an eccentric with no angular advance. This gear is consequently of the general type described at the be- ginning of the chapter, but the various irregularities of the gear are so marked that the valve-diagrams similar to Fig. i, PI. XXI, cannot be used at all in designing and laying out the gear. Since the guided point i^is always brought into coin- cidence with the axis of the bell-crank lever when the engine is on a dead-point, the lead is the same for all gears. RADIAL VALVE-GEARS. 93 The cut-off is shortened by making D approach D^ ; the thin lines show the gear set for a cut-off at about f stroke. The engine is reversed by carrying D beyond D^ toward D" . Fig- 3 gives the valve-ellipses for full-gear and for a short cut-off, corresponding with the diagrams in full lines and finer lines shown by Fig. 2. The valve-ellipses show the defect of the gear, which is a marked inequality in the maximum port- opening. In spite of this defect, the gear has been much used for horizontal marine engines, as it interferes less than would a link-motion with the access to the working parts of the engine when running. In some cases the point G is taken between the eccentric and the guided point, in which case the eccentric is set oppo- site the crank. The design of the Marshall valve-gear must be carried out by the aid of diagrams or a model, or by a combination of the two methods, which appears to be well adapted for this work. Hackworth Valve-gear. — This gear differs from the Mar- shall gear in having the guided point carried by a block that slides in straight guides and thus avoids the irregularity due to the guiding-link. The irregularity of the valve-motion is less than when the Marshall gear is used, and the maximum port-openings can be made nearly equal. The pressure on the sliding-block is large, especially at full-gear, and unless ample wearing surface is provided the friction and wear are liable to be excessive. In some cases the sliding-block has been pro- vided with rollers to reduce the friction. Joy Valve-gear. — An example of the Joy valve-gear used on the Pennsylvania Railroad Company's tugboat Delaware is shown by PI. XXIV. XX' is the centre-line of the crank and connecting-rod, and xx' is the centre-line of the valve-spindle. The lever abc is guided on a flat arc by the rod go, and is attached to the connecting-rod at the point a^ which describes an oval having the length a^a^ equal to the stroke of the engine. This oval, which is omitted to avoid confusion of the diagram. 94 VALVE-GEARS FOR STEAM-ENGINES. is symmetrical with regard to the axis XX' j and is slightly more pointed at the cross-head end than at the crank end. The point ^, which describes the irregular oval bb^b^ , takes the place of the centre of the single eccentric used with the Marshall valve-gear (PI. XXIII), and acts on the lever bze. The point t of the lever bie is guided on the circular 2.x o, ff^ by the sliding-block B^ and the point e^ which describes the oval ee^e^ , carries the valve-rod ed. The connecting-rod CD^ the valve-rod ed, and the rod eg are in one plane ; the levers ac and be and the curved guide-bar ^^ are double, one system being on each side of the connecting-rod ; in the figure the system of levers in front of the connecting-rod is omitted to show the construction more clearly. The guided point i could evidently be guided on the arc ff^ by a link centred at k \ such a construction is frequently used in marine engines. A comparison of this gear with the Marshall valve-gear will show much similarity. The essential points of difference are : (i) the radius of the guiding-arc ff^ is always equal to the length of . the valve-rod, and (2) the irregularity due to the angularity of the lever bie is compensated by the action of the lever ac, somewhat in the manner that the linkage known as Watt's parallel motion is made to give nearly a straight-line motion. These advantages are attained at the expense of greater com- plication and cumbersomeness ; in passing it may be remarked that the inequality of port-opening, which is the notable defect of the Marshall gear, may be nearly if not quite remedied by making the length of the guiding-link equal to the valve-rod, but such a construction is usually impracticable since it requires either an impossible length for the guiding-link, or else a short valve-rod and a long valve-spindle that must be guided at the outer end. The guiding-bars _//"i are hung on trunnions with the axis at the point i^ , and are connected at/" to the reversing- lever. The gear is shown at full-gear for left-handed rotation ; it may give a shorter cut-off if the guiding-bars ff^ are given less inclination from the horizontal or mid-gear position, and RADIAL VALVE-GEARS. 95 when in mid-gear it will give the valve a motion equal to the lap plus the lead ; if the guiding-bars //"^ are inclined the other way the engine will be reversed. This gear, when properly proportioned, gives a rapid motion to the valve when opening and closing, less compression at short cut-off than does a link-motion, and the cut-off can be made nearly equal for all grades of the gear. Like all other radial valve-gears, it gives a constant lead. Its defects are, the number of parts and of joints that are liable to wear loose, and the obstruction that it offers to inspection and care of the crank-pin and cross-head when the engine is running. To lay out a Joy valve-gear : choose a point a on the con- necting-rod, having a transverse throw equal to twice the maxi- mum displacement of the valve ; make the length of the lever ac such that the angle a^c^a^ shall not be more than 90° ; draw the centre-line xx'' of the valve-gear, and locate the point e^ op- posite the middle-point of the line a^a^\ lay off e^e^ = e^e^ equal to the lap plus the lead ; lay off aj)^ = aj)^ equal to about one third of a^c^ ; and draw the lines ej}^ and ejb^ intersecting at i^ : this last point locates the axis of the trunnions carrying the guiding-bars ff^. If the point i is guided by a link, then the arm carrying the link must be centred at i^ ; in the figure the position of such an arm is shown by the line ij^. Finally, the valve-ellipse should be drawn for several grades of the gear; in the figure the valve-ellipse 00^ has its length equal to the stroke of the engine, and the valve-displacements are magnified fourfold, i.e. st = /\d^d. Usually the axis xx' of the valve- motion is determined by the general design, and cannot be changed much, if at all. The length of the lever ac will com- monly be as great as desirable if the angle a^cjj^ is something less than 90° ; it cannot be made shorter without throwing ex- cessive stress on the links and levers. The transverse motion of the point a is properly twice the maximum displacement of the valve, in order that the inclination of the guide-bars //"^ may not be more than 25° or 30°; should such a location 9^ VALVE-GEARS FOR STEAM-ENGINES. bring the lever ac too near the shaft, as may be the case when the crank is counterweighted, then a may be placed nearer the cross-head, but at the expense of more inclination of the guiding-bars at long cut-off. The location of the point b is under the control of the designer, and may be used to equalize the cut-off either at that grade at which the engine is to run habitually, or el^e to give nearly equal cut-off at all grades. The equalization of the cut-off must be made by trial, and no general rule can be given, since the elements, such as the length of the lever ac and the distance between the axes xx' and XX' y over which the designer has little control, have a large influ- ence. A skeleton model may be used to advantage in this work. It should have rods to represent the connecting-rod, the levers ac and be^ and the links eg and ki. The point k will be located on an arc of a circle centred at i^ . The points i and b on the levers be and ac should be made adjustable ;: the first by mounting it on a sliding-block that can be clamped in any desired position, and the second by that method, or by a series of holes to receive the screw representing the pin at b. It may be found advantageous to provide pieces, properly guided, to represent the cross-head and head of the valve- spindle, but a simple model may be made by placing these points by hand on the lines XX' and xx' for the several po- sitions of the model, and the centre C may in like manner be placed on the circle CC'C* CHAPTER V. DOUBLE VALVE-GEARS. A PLAIN slide-valve, set to give an early cut-off, is liable to give either an excessive compression or an early release, or both. A single valve under the control of a gear that gives a variable cut-off, such as a shifting-eccentric or a link-motion, is open to the same difficulties ; and in addition the compres- sion varies with the cut-off, though to a less degree. For a stationary engine a large compression may be undesirable, and a varying compression is always so. To avoid these difficulties two valves are frequently used ; one, called the main valve, has an unvariable motion, and gives the admission, release, and compression ; the other, called the cut-off valve, gives the cut-off only, which may be varied without affecting the other events of the stroke. The cut-off valve may be placed in a separate valve-chest, as shown by Fig. 3, PI. XXV, or it may be placed on the back of the main valve, as shown by Fig. i, PI. XXVI ; thus giving rise to two separate types of double valve-gears. It is impor- tant to obtain a clear conception of the principles of double valves, and then all existing forms of double valve-gears may be readily understood, and a gear for a given purpose may be easily designed, or else it may be shown that a satisfactory design is impossible. Cut-off Valve in a Separate Valve-chest. — The usual arrangement of this valve-gear is shown by Fig. 3, PI. XXV. The main valve, which receives motion from an eccentric with 97 98 VALVE-GEARS FOR STEAM-ENGINES. constant angular advance and eccentricity, is designed to give the desired release and compression, and is set to give equal lead ; it will be observed that either the release or the com- pression may be equalized. In the figure there is no inside lap ; this arrangement may be frequently found desirable, but it is chosen here for the sake of simplicity, and will be adhered to throughout the chapter ; attention will be given exclusively to the cut-off, since the other features of the gear are the same as for a plain slide-valve and have received sufficient attention in the first chapter. In the figure the valves are shown disconnected from the eccentrics and both in mid-position ; they cannot both be in such position when the gear is connected up, but such a draw- ing is convenient in laying out the valves. The cut-off valve is a rectangular open frame having the acting edges inside. The distance / from one edge of the valve to the opposite edge of the port is the clearance of the valve ; when the valve is dis- placed from mid-position an amount equal to the clearance, the cut-off valve gives either cut-off or readmissio7i. The right- hand edge gives cut-off for the head end of the engine, and readmission for the crank end ; it is important that the read- mission by the cut-ofT valve should precede the admission by the main valve, in order that the second steam-chest shall then be properly filled with steam at full pressure. The steam-porj; a^ is for the passage of full-pressure steam only, and may consequently be made from |^ to f of the area of the port a, through which exhaust steam also must pass. Two or more ports are frequently provided in the cut-off valve- seat, and the valve is then known as a gridiron valve ; in such case the combined area of all the ports a^ should be a little in excess of what would be given to one port, to allow for the greater friction in numerous narrow passages. A gridiron valve acting on several narrow ports will require a proportionately less throw. In Fig I, PL XXV, OP is the diameter of the valve-circle DOUBLE VALVE-GEARS. 99 for the main valve which gives admission at OR^ and OR^, and cut-off at OR. The eccentric acting on the cut-off valve is given a negative angular advance, i.e. it is less than 90° in advance of the crank ; consequently the displacements of the cut-off valve from mid-position are given by a valve-circle, such as OP^, hav- ing its diameter laid off at an angle 8^ (the negative angular advance) away from the crank. If the clearance is equal to (94, then the cut-off by the cut-off valve occurs at OR^ and the readmission at OR. A variable cut-off by aid of this gear may be obtained by varying the clearance of the valve, or the throw of the eccentric, -or by using a shifting-eccentric. Now the same effect is pro- duced by increasing the lap (or decreasing the clearance) with a constant eccentricity as is produced by decreasing the eccentricity with a constant lap ; consequently an investigation of one of the two methods will serve for both. It will appear that neither will give a good variable cut-off. On the other hand, a satisfactory gear may be had by using a shifting- eccentric. In Fig. I, PL XXV, the clearance 01^ gives cut-off at OR^ and readmission at OR, coincident with the cut-off by the main valve : and that is the latest admissible cut-off ; for if the •clearance is made equal to 01^ in order to obtain a cut-off at OR^ , the readmission will occur at OR^ , before the main-valve ■cut-off, and a double admission of steam will occur. Such a double admission causes a large waste of steam and an irregular action of the engine and cannot be tolerated. The earliest admissible cut-off is obtained by a clearance equal to 01^, which gives cut-off at OR^ and readmission at OR^ . An in- spection of the figure shows that the angle RfiR^ must, from symmetry, be equal to ROR^\ the latter angle depends on the lap of the main valve, and it is at once evident that only a limited range of cut-off can be obtained with such an arrangement. In practice a gridiron cut-off valve was commonly used and the cut-off was obtained by varying the travel of the valve. 100 VALVE-GEARS FOR STEAM-ENGINES. For this purpose a fixed eccentric was connected to the end of a slotted lever or link, and motion was communicated to the valve from a link-block that could be set at any desired place in the link. To shorten the cut-off, the link-block was moved toward the fixed end or fulcrum of the link. An idea of this arrangement may be obtained by supposing the radius-rod of the Walschaert gear (Plate XXII) to act on the cut-off valve^ spindle ; the eccentric should, of course, have a negative angular advance, and only half of the link will be required. The effect of such an arrangement is the same as though the eccentricity were varied, and, just as has been shown to be the case for a gear with varying lap, the range of variation of cut-off depends on the lap of the main valve. Cut-off Valve with Shifting-eccentric. — If the cut-off valve in a separate valve-chest is moved by a properly de- signed shifting-eccentric, the readmission may be kept within the proper limits, i.e. before the admission and after the cut- off by the main valve, and at the same time any desired range of cut-off may be had. In Fig. 2, PI. XXV, let OP be the main valve-circle, giving cut-off at OR and admission at OR^. Suppose that the cut-off is to vary from OR^ , corresponding with ■§• stroke, to OR co- incident with the cut-off by the main valve. Bisect the angle ROR^ by the line OP^, and bisect the angle ROR^ by the line OP^ ; then a valve-circle centred at 67/ can be made to give a cut-off coincident with that by the main valve, and readmission coincident with the main-valve admission ; while a valve-circle centred at C^ can give cut-off at OR^ , and readmis- sion coincident with the main-valve cut-off at OR. It is convenient to have the shifting-eccentric on an arm centred on the centre-line of the crank, or rather on that line produced, for then the cut-off gear will work equally well in forward and in backing gear on a reversing-engine ; but for a stationary engine such an arrangement is not essential. We now have three elements of which one may be chosen at DOUBLE VALVE-GEARS. lOI pleasure and the other two will then be determinate. In the figure the diameter OP^ of the valve-circle to give early cut-off is made if inches, equal to the eccentricity for the main eccentric ; the clearance 01 of the cut-off valve is then deter- mined by the intersection of that circle by the lines OR and OR^ , and is if of an inch. The centre C^ of the valve-circle to give the cut-off coincident with that of the main valve must be so chosen that it shall pass through O and I" ; it is on a perpendicular to 01" at its middle point. Draw AS perpen- dicular to the middle point of an imaginary line joining P^ and P/; then 5, the intersection of this line with the axis XX' , is the centre of an arc on which the point P^ will travel as the cut- off is shortened. The shifting-eccentric must be swung from a point on the centre-line of the crank produced, and at a distance from the centre of the shaft equal to OS (ij of an inch) ; when the crank is at OX this point will be at T. With ordinary proportions for an engine, the point 2"is liable to fall inside the shaft, or el^ so near that the construction of an arm for the eccentric will be impossible. A reversing-engine may have the following arrangement : fast to the shaft may be an eccen-^ trie with an eccentricity equal to OS and with its centre opposite the crank-pin ; this eccentric may carry another having an eccentricity equal to ^5 (if |- of an inch); the second or outside eccentric will be turned toward the crank so that its angular advance will be negative. A stationary engine that always runs in one direction may have the swinging arm, for the eccen- tric controlling the cut-off valve, centred at any convenient point on the line AS produced. There now remains the determination of the width of the cut-off valve to prevent leakage past the outside edge. The greatest displacement of the valve from mid-position is equal to OP^ = i|- of an inch ; consequently the distance in Fig. 3, PL XXV, from the outside edge of the valve to the nearest edge of the port should be somewhat greater than if of an inch ; it is made ly^- of an inch. 102 VALVE-GEARS FOR STEAM-ENGINES. Cut-off Valve on back of Main Valve.— Fig. i, PL XXVI, shows a cut-off valve on the back of the main valve, both being disconnected from their eccentrics and placed in mid-position. When the gear is connected up, such a position of both valves at the same time does not occur, but it is convenient to make a drawing of the valves in that position to show the laps and other dimensions of the main valve, and the clearance and length of the cut-off valve. The main valve is designed to give the desired compression and release, and is set to give equal lead ; either the compression or the release may be equalized. The cut-off valve is connected to an eccentric having a large angular advance, so that it is nearly (sometimes exactly) op- posite the crank. In Fig. 3, OP is tlie diameter of the main valve-circle, and OPf^ is the diameter of a valve-circle showing the absolute dis- placements of the cut-off valve. At any crank-position, such as OR', the chord Oc intercepted by the main valve-circle shows the displacement of the main valve; represented in Fig. 2 by £. The absolute displacement of the cut-off valve is shown by the chord Ob, intercepted by the cut-off valve-circle ; repre- sented by e^ in Fig. 2. Both of these displacements are toward the left, but the former being the greater, the relative displace- ment e^ of the cut-off valve with regard to the main valve, and measured from the centre of the main valve, is towards the right, tending to shut the port in the main valve. Now the displacement of the main valve, with an eccen- tricity r and an angular advance d, is, for the crank-angle 6^ e = r sm{S -\- 6) = r sin 6 cos 6 -\- r cos d sin B. . (56) The cut-off valve-eccentric has the angular advance S^ and the eccentricity r^ , and for the crank-angle 6 the displacement of the cut-off valve is e, = r, sin {6^ -\- 6) = r, sin d^ cos ^ -f r„ cos 6^ sin 6. (57) DOUBLE VALVE-GEARS. IO3 The relative displacement of the cut-off valve measured from the middle of the main valve is e^-=e—e^={r sin d— r^ sin d^ cos d-\-{^ cos d— r„ cos d^ sin d, (58) which may be written e^=^ A cos 6^ -|- ^ sin 6^ ^^M Now it has been shown on page 50 that any valve-motion that can be represented by an equation having the form of equation (59) is harmonic ; and can be represented by a valve-circle^ having for the coordinates of the end of the diameter of the valve-circle A and B. Since r„ cos S^ is longer than r cos S, A will in this case be negative and must be laid off to the left of the origin. Conse- quently the circle representing the relative motion of the cut- off valve may be located in Fig. 3 by laying off Ov = A and vP^ = B^ and then by drawing the diameter OP^ on which the circle is to be drawn. This circle is called the auxiliary circle ; it is to be borne in mind that the angle YOP^ is not an angular advance, nor is OP^ an eccentricity. Auxiliary Valve- circle. — Although the auxiliary valve- circle can always be drawn by the process just stated, the usual and convenient method is as follows : In Fig. 4, PI. XXVI, let OP and OP^ be the diameters of the valve-circles for the main valve and the cut-off valve ; with /* as a centre and with a radius equal to OP^ , and with 6^ as a centre and with a radius equal to PP^ , draw arcs intersecting at P^ ; then OP^ is the diameter of the auxiliary circle. The figure PxPPfi is of course a parallelogram, and the process just described will be called completing the parallelogram. To prove the method just given : Draw P^v and Pu parallel to OY, and draw Ps and Pj: parallel to OX. It is evident that Ov = P^u — Pj; — Ps =r^ sin d^— r sin 6 =: —A ; P^v = Pu = Os — Ot = r cos ^ — r^ cos S^ = B. 104 VALVE-GEARS FOR STEAM-ENGINES. When the displacement of the cut-off valve from the middle of the main valve is equal to the clearance /, Fig. i, PI. XXVI, the edge of the valve coincides with the edge of the port and we have either cut-off or readmission. The crank-position at cut-off and readmission can be found by drawing the circle llj^ with a radius 01 equal to the clearance /. In Fig. 3 cut-off oc- curs at the crank-position OR^ , corresponding to a piston-dis- placement xa^ for harmonic motion ; the valves are then in the position shown by Fig. 6. Readmission occurs at OR'' ; the edge of the cut-off valve is on the edge of the port as in Fig. 6, but the valve is then moving towards the left to open the port through the main valve; the main valve is not in the position shown by that figure. It is important to know when cut-off by the main valve occurs ; in Fig. 3 it occurs at OR, correspond- ing, with harmonic motion, to a piston-displacement xa ; the valves are then in the position shown by Fig. 7. The readmission by the cut-off valve must not occur before cut-off by the main valve, otherwise a double admission of steam will take place. It should occur before the admission by the right-hand end of the main valve ; i.e., before the crank comes to the position ORa , Fig. 3. The readmission for the left- hand end of the main valve is of course given by the left-hand edge of the cut-off valve. Meyer Valve. — A double valve-gear, known as the Meyer valve, is shown by Fig. 2, PI. XXVIII. The cut-off valve is made in two parts on a valve-spindle with a right and left screw, so that the position of the plates may be adjusted by rotating the valve-spindle ; thus the clearance may be changed, and consequently the cut-off may be varied. In order that this may be done while the engine is running, there is a swivel-joint in the valve-spindle between the valve-rod head and the valve-chest, and the tail of the valve-spindle is carried through the head end of the valve-chest, where it reciprocates through a hand-wheel as shown by Fig. i ; the DOUBLE VALVE-GEARS, IO5 valve-spindle is squared so that it may be rotated by turning the hand-wheel. In Fig. I, PI. XXVII, the valve-circle OP, the auxiliary circle OP^, and the diameter OP^ for the valve-circle showing the absolute motion of the cut-off valve, are transferred from Fig. 3, PL XXVI. The cut-off by the main valve occurs, as in that figure, at OR. With the same clearance 01, the cut-off by the cut-off valve is at the crank-position OR^ , 90° from the dead-point. If the clearance is increased to the amount 01^ , the cut-off is delayed and occurs at the crank-position OR^ . With this clearance the readmission occurs at OR, coincident with the cut-off by the main valve, and a later cut-off would be accompanied by a readmission ; consequently the latest admissible cut-off is at the crank-position OR^ . In order that cut-off may occur at a given crank-position, for example at OR^ corresponding to a piston-displacement -equal to Xa^ , the clearance must be made equal to the chord (7// = 01^ , cut from the line OR^ by the auxiliary circle. If the clearance becomes zero, so that in mid-position the edge of the cut-off valve coincides with the outer edge of the port in the main valve, then the cut-off comes at a crank-position 0R\ perpendicular to the diameter OP^ of the auxiliary circle. For an earlier cut-off, the line of the crank, for example OR^ , will cut the auxiliary circle at a point beyond the origin O ; and the cut-off valve will then have a lap equal to O// = 01^ when in mid-position. For a cut-off at the dead-point a lap equal to 01^ will be required. The crank-position at readmis- sion is at the second intersection of the auxiliary circle by the clearance-circle ; for example, with a clearance equal to 01^ , the readmission occurs at OR. When the valve has a lap in mid-position, for example 01^, the readmission occurs at a crank-position found by drawing a line from the second inter- section of the clearance-circle and the auxiliary circle, towards the origin O and thence to the crank-pin circle. The readmis- sion usually comes in the second or third quadrant, and so long I06 VALVE-GEARS FOR STEAM-ENGINES. as it is later than the cut-off by the main valve, it is of little importance to know just where it occurs. Design of a Meyer Valve. — The main valve shown by Fig. 2, PI. XXVIII, has an outside lap of half an inch, and is moved by an eccentric with an eccentricity of i J of an inch ; the inside lap is zero. It is set with -Jg- of an inch lead. With these dimensions the valve-circle OP in Fig. 3 can be drawn,. and the cut-off will be found to occur at ORc\ corresponding,, with harmonic motion, to 0.89 of the stroke of the piston. The steam-port in the valve-seat is f of an inch wide, and the steam-port through the main valve may be taken to be f as much, or \ of an inch. The diameter of the auxiliary circle may be assumed to be one inch. With an eccentricity of if of an inch for the cut-off valve-eccentric, the parallelo- gram PPfiP^ may be drawn locating both the auxiliary circle OP^ and the diameter OP^ of the circle showing the absolute displacements of the cut-off valve. The cut-off valve-eccentric has the angular advance YOP^, equal to 55°; it is convenient to know this angle approximately in setting the valves. The dimensions for OP^ and OP^ are chosen by trial to give a con- venient location of the auxiliary circle, with its diameter placed beyond OR^ . Were the auxiliary circle placed with its diam- eter coincident with OR^, as shown by Fig. 5, PI. XXVI, then a readmission would be impossible ; such a disposition is recommended by Zeuner, but it has the disadvantage that the motion of the cut-off valve is very slow when giving a long cut-off. This defect is mitigated by placing the auxiliary circle beyond OR^, as in Fig. 3, PI. XXVIII; the earliest possible readmission is clearly at OR^, The largest clearance is 01^ 01' = W of an inch; the least clearance, or the greatest lap, will depend on the earliest required cut-off. Let it be assumed that the earliest cut-off shall be at OR^ , corresponding to a piston-displacement Xa = -J of the stroke, for harmonic motion ; then the cut-off valve must have a lap equal to 01^ = 01 ^ =yV ^^ ^^^ inch, nearly. DOUBLE VALVE-GEARS. 10/ The lower face of the main valve is laid out as for a plain slide-valve, but with the additions demanded by the passage through it. The least width of bridge is equal to the eccentricity less the sum of the lap and the width of port, or i i — (i + f ), = i of an inch ; the width used is \ of an inch. In like manner the width of the exhaust-space is ij- + | — J = if of an inch ; the width used is 2 X t| = ij of an inch. In order that the edge c of the passage through the valve may not reduce the passage through the port to less than \ of an inch, the distance ac is made 1 4 H" i" = 2 inches; this feature is fre- quently overlooked and dc is carelessly made equal to ef. In order that the edge g of the valve shall not come to the edge b of the port, the distance bg is made ii-'+ |- = if of an inch. The valve-face is cut away at a point h, if of an inch from g^ thereby giving an overtravel of \ of an inch. In order that the space efdc may be made small, the height of the exhaust- space is made only ly^ of an inch, a dimension that is probably too small to give a perfectly free exhaust. The width of the cut-off valve must be enough so that steam cannot leak past the inside edge when the valve is set to give the earliest cut-off and also has its maximum displace- ment. In Fig. 2, PI. XXVIII, the position of the valve to give cut-off at \ of the stroke is shown by dotted lines ; its lap is ^ of an inch, and its left-hand edge /^ is i -|- -J = i|^of an inch from e. The length of the cut-ofT valve is \\ -\- ^ -\- ^ ^=. 2-^-^ inches. The cut-off valve is shown in section with a clearance of W of an inch, which is proper for giving the longest cut-off coin- cident with that of the main valve ; its left-hand edge is 2^ -j- |i- =: 3gL inches from the edge /" of the port ef', the distance of the edge f of the port from the middle of the valve is made '^^-^ inches. The half-length of the main valve, over all, is made 4J inches, and provides an overtravel of J of an inch for the cut-off valve when set to give the earliest cut-off. The valve-spindle is provided with a right-and-left-hand screw, of which the right-hand part is shown. The thread I08 VALVE-GEARS FOR STEAM-ENGINES. should be cut only far enough to give the desired variation of cut-off, or some other stop should be provided in order that the engine attendants may not move the cut-off valve too far out, and so get a leakage or even admission of steam past the inner edge. The spindle is shown in two parts, joined by a right-and-left screw and circular nut or sleeve with pins to prevent the joint from jarring loose ; this arrangement is to facilitate the erection of the valve-gear. In this design the inside lap is made zero and the compres- sion and release are neglected ; in practice these features should receive the same attention as is accorded to them in designing a plain slide-valve. Again, the irregularity of the piston-motion due to the angularity of the connecting-rod has been ignored, and the clearance (or lap) of the cut-off valve has been made the same at both ends. This method is commonly followed in practice, but by using proper pitches for the threads on the valve-spindle the cut-off may be equal- ized at two points of the stroke, for example at ^ and at \ stroke, and will then be found to be more nearly equal for all parts of the stroke, except for long cut-off, when inequality is of less importance. Meyer Valve with Cut-off at Inside Edge. — Sometimes the Meyer valve is designed to cut off at the inside edge, as shown by Fig. 4, PI. XXVIII. It is then convenient to con- sider that the valve has a lap ab which diminishes as the cut- off is lengthened, and which may become zero and finally change to a clearance, shown by ac when the valve is in the position indicated by dotted lines. The eccentric is given a negative angular advance, i.e. it is set somewhat less than 90° ahead of the crank. In Fig. 2, PI. XXVII, the main valve-circle is (9P, giving a c'ut-off at ORy w^ith a lap On = On". Let it be assumed that the earliest required cut-off is at OR^ , and that the latest read- mission must be at OR^ ; then the auxiliary circle may have its diameter at OP^., on a line bisecting the angle R^OR^. Tiie DOUBLE VALVE-GEARS. IO9 diameter of the valve-circle for showing the absolute displace- ment of the cut-off valve will be found at OP^ by completing the parallelogram PP^OP^\ the eccentricity for the cut-off valve-eccentric is OP^ , and the negative angular advance is YOP^. The auxiliary circle may be placed lower down, thereby giving an earlier readmission and at the same time a longer eccentricity, but it cannot be placed higher up. There is no danger of a double admission of steam at a long cut-off, as was found to be the case with the ordinary form of Meyer valve. Meyer Valve for Reversing-engines. — If a Meyer valve is designed to give a satisfactory action in forward gear for a reversing-engine, the backing gear is liable to be very unfavor- able. In Fig. 5, PL XXVI, the main valve-circle for forward gear is OP, and with a lap-circle nn'n" the cut-off comes at OR,^ corresponding to f stroke with harmonic motion. The auxil- iary valve-circle has its diameter coincident with OR^ to avoid the possibility of a double admission of steam ; the eccentricity is found to be equal to OP^, and the angular advance is equal to YOP^. In backing gear the main valve-circle is at OF , and cut-off by the main valve is at OR^ , corresponding to f stroke, but the auxiliary valve-circle is now OPJ' , found by complet- ing the parallelogram /*'Po 6^/*/'. With a clearance equal to (9/ = 01' = 01" y the cut-off in forward gear occurs at half-stroke, but in backing gear the cut-off is at OR^ , corresponding to \ stroke. Such an arrangement cannot be used, for with it the engine cannot run at full power m backing gear. The action of the cut-off valve may be made the same in forward and backing gears by giving its eccentric 90° angular advance. In Fig. 5 the parallelogram PP\OPJ is drawn with PP^ parallel to ORc, and with PPJ parallel to XX', thus giving the diame- ter of the auxiliary circle coincident with OR^ and giving 90° angular advance, but both the cut-ofT eccentricity and the relative travel of the valve are thereby made excessive. In practice the cut-off eccentric is often given the same eccen- no VALVE-GEARS FOR STEAM-ENGINES. tricity as the main valve-eccentric, and then with 90° angular advance it is liable to give a double admission at long cut-off. Cut-off Valve with Loose Eccentric. — Let the cut-off valve receive motion from an eccentric which may turn freely on the engine-shaft and which is under the control of a shaft- governor ; let the clearance of the cut-off valve be unalterable : then the cut-off can be varied by changing the angular advance of the cut-off eccentric. In Fig. 3, PI. XXVII, the main valve-circle is OP, and with a lap on = on' the cut-off by the main valve occurs at OR^^ The cut-off eccentric may have its angular advance changed from YOP^ to YOP^, and the auxiliary circle may change from OP^ to OPJ. When the position of the diameter OP^ of the circle showing the absolute displacement of the cut- off valve is known, the auxiliary circle may be located by completing the parallelogram PP^OP^\ or the centre Cx of the auxiliary circle may be located by completing a parallelogram CC^OC X on the half-diameters OC and OC^ of the main valve- circle and the cut-off valve-circle. Since the side CC^ of this last parallelogram is equal to OC^^^\OP^ , it is at once apparent that the locus of the centre of the auxiliary circle is the dotted circle C^CJ drawn from the centre C of the main valve-circle, and with a radius equal to half the eccentricity of the cut-off eccentric. Again, since PP^ is equal to OP^ , the locus of the end of the diameter of the auxiliary circle is a circle drawn from P as a centre and with a radius equal to the eccentricity of the cut-off eccentric. The locus of the centre of the auxil- iary circle is the more convenient for use in solution of prob- lems. Let it be assumed that the cut-off shall vary from the crank- position ORc , coincident with the cut-off by the main valve, to OR^ , corresponding to \ stroke for harmonic motion. Assume the clearance of the cut-off valve and draw the circle ll'l". Erect a perpendicular 5(7^ at the middle of the line 01" \ it will intersect the locus C^CJ at C^ . the centre of the auxiliary DOUBLE VALVE-GEARS. Ill circle that will give a cut-off by the cut-off valve coincident with the cut-off by the main valve. In like manner, erect a perpendicular at the middle of the line 01' ; it will locate the centre C^ of the auxiliary circle that gives a cut-off at OR^ , corresponding to \ stroke. In designing a valve-gear of this type, the clearance of the cut-off valve may be chosen, usually somewhat larger than the width of the port in the main valve ; and then the auxiliary circle for maximum cut-off may be given such a diameter that a satisfactory action may be had in that gear. The eccentricity of the cut-off eccentric will be found by completing the parallelogram in the usual way ; should the result be an undesirable dimension it may be modi- fied, since the diameter of the auxiliary circle may be varied to a considerable extent. Finally, the auxiliary circle to give the earliest cut-off may be found by the process just stated ; it is liable to have a large diameter, and the travel and wear of the cut-off valve is likely to be excessive. In the figure the auxil- iary circle which gives a cut-off at J stroke is one third larger than the main valve-circle, and it would be still larger for a shorter cut-off. The maximum diameter of the auxiliary circle is equal to the sum of the eccentricities for the two eccentrics. If this gear is used with a shaft-governor, the cut-off by the main valve will commonly be earlier than that shown in Fig. 3 — a circumstance that will make the design of a satisfactory gear easier. Moreover it may be possible to limit the maxi- mum cut-off to half-stroke or less, even though the main valve gives a cut-off beyond half-stroke ; in that case the valve mechanism must be so arranged that the cut-off valve cannot act beyond the assumed range of cut-off, otherwise a double admission may occur. These observations are applicable also to the next type of valve-gear. Cut-off Valve with Constant Travel. — It is desirable that a valve shall overtravel its seat in order that the seat and the face of the valve may wear evenly and remain true. This is seldom possible for a Meyer valve with the common propor- 112 VALVE-GEARS FOR STEAM-ENGINES. tions, or for the cut-off valve under the contro] of a loose eccentric. It has been seen that the design of the cut-off valve is conveniently begun by choosing the diameter and position of the auxiliary circle ; it will be found that the design of a double valve-gear for a given purpose may be worked out by first finding how the auxiliary circle must be located or changed to give the desired action, and then finding how the cut-off eccen- tric must move to produce such an auxiliary circle. Suppose that the auxiliary valve-circle is to have a constant diameter, and that the variation in cut-off is to be produced by swinging the auxiliary circle around the origin O, Fig. i, PL XXIX, from the position OP^ to OP^. With a clearance equal to (9/= 01' = 01", the first-named auxiliary circle will give cut-off at OR^ , coincident with the cut-off by the main valve ; and the other auxiliary circle, OP J, will give cut-off at OR^ , corresponding to \ stroke. The diameters of the cut-off valve-circles, showing absolute displacements, are OP^ and OP^, found by completing the parallelograms PP^OP^ and PPJOP^'. It is evident that the locus of the point P^ is the circle P^PJ drawn from P as a centre and with a radius equal to the diameter, of the auxiliary circle. The arrangement of the eccentrics for this type of valve- gear is shown by Fig. 2, PI. XXIX. The centre of the engine- shaft is at 6^ ; on the shaft is the fixed eccentric centred at Ey for giving motion to the main valve ; the cut-off eccentric is carried by the main eccentric, and is shown by the full-line circle with its centre at ^/, corresponding to OPJ in Fig. i, while the dotted circle shows it with the centre at E^ , corre- sponding to OP^ in Fig. i. The cut-off eccentric may readily be placed under the control of a shaft-governor. The main valve represented by Fig. 3, PI. XXIX, has a lap of ^ of an inch, and is moved by an eccentric having i^ of an inch eccentricity, and is consequently a reduplication of the main valve for the Meyer valve-gear shown on PI. XXVIII, at its lower surface ; the top is of course laid out after the cut-off DOUBLE VALVE-GEARS. II3 valve has been designed. The main valve-circle has its di- ameter at OP, Fig. I, and the cut-off by that valve occurs at ORc. The cut-off valve is a double-ported or gridiron valve, each of the ports being \ of an inch wide. The clearance of the cut-ofT valve is assumed to be f of an inch, represented by the circle U'l" . As the auxiliary valve swings toward the right to give an earlier cut-off, the readmission moves through the same angle toward the line OR^, the crank-position at cut-off by the main valve ; and it is at once evident that the readmis- sion must not be earlier than ORcy otherwise a double admis- sion may occur. The smallest admissible auxiliary circle will have its centre on the line Os bisecting the angle R^OR^, and it will pass through the points I' and l'^ at the intersection of the clearance-circle by the lines OR^ and OR^. The diam- eter chosen for the auxiliary circle is f of an inch, or twice the clearance of the cut-off valve. The extreme positions, OPJ and OP^y of the auxiliary circle are so located that they shall pass, one through r and the other through l"\ they give cut-off at OR^ and at OR^. The corresponding diameters of the cut-off valve-circles are OP^' and OP^, found by complet- ing the parallelograms P^PP^O and PJPP^'O. The cut-off eccentric is mounted on the main eccentric and has an ec- centricity, referred to that eccentric, of J of an inch, equal to the diameter of the auxiliary circle. In laying out the cut-off valve and the upper surface of the main valve, it is convenient to begin in Fig. 3 by placing the port a in a convenient position near the exhaust-space e. From the right-hand edge of this port lay off f of an inch to c, the left-hand edge of the outer part of the cut-off valve ; this is equal to the greatest displacement of that valve, and insures that the edge c shall not overrun and contract the port a. From c lay off somewhat more than f of an inch, in this case if of an inch, to the left-hand edge of the port b ; this gives the necessary length of the bar cd in order that leakage may not occur past the edge c at the maximum displacement 114 VALVE-GEARS FOR STEAM-ENGINES. toward the right. The clearance, f of an inch, is laid off from the right-hand edge of the port b^ to determine the edge d of the bar cd. The inner right-hand bar is made as wide as cd. The left-hand half of the main valve and cut-off valve is a counterpart of the right-hand half. The cut-off valve-spindle takes hold of a lug on one of the bars of the gridiron cut-off valve ; and the main valve-spindle passes through a tube or passage cored out through the middle of that valve. To Set a Double Valve-gear. — Set the main valve to give equal lead ; the cut-off by that valve has little influence on the running of the engine, and requires little or no atten- tion. If the cut-off valve is designed to cut off at a definite point when the engine is running under normal conditions, equalize the cut-off by that valve at that point. If the load on the engine and the cut-off are variable within a limited range, the valve should be set to give the least irregularity within that range ; it will usually be sufficient to equalize the cut-off for the middle of the range. If the range of cut-off is wide it will often be impossible to get a good action for the entire range, and then it will be advisable to equalize the cut-off for some early point in the stroke of the piston. It has already been pointed out that the Meyer valve may have the cut-off equal- ized at two points of the stroke by using unequal pitches for the screws on the valve-spindle. CHAPTER VI. DROP CUT-OFF VALVE-GEARS. In this chapter there will be given descriptions of a few- special forms of valve-gears, selected, partly at random, from the large variety of such gears employed by the builders of automatic cut-off stationary engines. All are of the four-valve type of valve-gears, and all give a drop or disengagement cut- off. A description and analysis of these few forms will enable the student to analyze and understand other gears of similar types. Brown Engine Gear. — Fig. i, PI. XXX, gives a section through the head-end valves and valve-chests of the Brown engine ; the crank-end valves and gears are a duplication of those for the head end. The admission-valve F is a five-ported gridiron valve on a vertical valve-seat, and the exhaust-valve is a three-ported gridiron valve on a horizontal seat. Both are controlled by valve-gears on the shaft O, which is driven by the engine-shaft through a pair of equal bevel-gears and makes one revolution for each revolution of the engine. It is clear that four such valves might be driven directly by one eccentric on the engine-shaft, or by four eccentrics on the shaft Oj and that in such case the four valves would be equivalent to one plain slide-valve, and would be designed by the princi- ples laid down in the first chapter. The eccentric E, which moves the steam valve-gear, is set to one side of a vertical through ^, so that it gives a rapid 115 Il6 VALVE-GEARS FOR STEAM-ENGINES. Upward motion to the lever />. The toe of the lever /> catches under the edge of the latch Z, and lifts the valve V through the spindle SV. When the tail of the latch strikes the pin dy the valve is disengaged from the lever />, and it falls shut ; a dash-pot P checks the motion of the valve and prevents jar. The pin d on the arm bd is under the control of the governor through the herizontal shaft b. It is commonly said that the governor on an engine with a detachment cut-off gear has only the light duty of setting the stop (in this case the pin d) that unlatches the gear and releases the valve ; the friction of the governor and the attached parts is, or should be, small. Most such gears throw a shock on the governor, tending to disturb it and make it race when the cut-off valve is released ; and the governor should be sufficiently powerful to resist the shock. In this gear, when the tail of the latch L strikes the pin d^ the shock tends to open the latch and to throw the pin toward the left ; both will yield, but the motion of d, and consequently of the governor, is slight. The exhaust-valve is moved by the cam C, which consists of a groove, in the face of a disk, in which works a roller on the end of the lever trS' , The end S' of the lever is slotted and provided with a block to avoid bending the valve-spindle V S' . The action of this cam is equivalent to that of an eccen- tric, except that there are periods of rest when the valve is open or shut. Fig. 2 shows two ways of laying out such a cam ; it is intended to show general principles only, and would require some modification to fit it to the engine shown by Fig. I. Let it be supposed that the cam acts directly on the end of a horizontal valve-spindle, such as V S' ^ Fig. i, and that its centre is on a prolongation of the path of the valve-spindle. Suppose further that the cam turns toward the left, and that the valve shall begin to open when the line Od is horizontal,, and be wide open when Ob is horizontal. To give a uniform motion to the cam, make the curve i, 3, 7 an arc of an Archi- medean spiral ; this is done by dividing the angular space Mand DROP CUT-OFF VALVE-GEARS. 11/ the linear space 1^7 into the same number of equal parts, and by drawing intersecting arcs and radii, 6'6, 6^6, 5^5, 6^5, etc., as shown. The cam from ^ to <: is a circular groove, so that the cam remains at rest till the line Oc comes into coincidence with the path of the valve-spindle. The groove from the line Oc to the line Oa is so designed that it gives a harmonic motion. On the line \'j a semicircle is drawn, and its arc and the angular space cOa are divided into the same number of equal parts; arcs and radii are drawn intersecting at i, 2, 3, etc. Finally, the cam from Oa to Od is a circular arc, giving a period of rest. The second construction, giving harmonic motion, is to be preferred for heavy valves having a rapid motion, in order that they may start and stop easily and quietly ; for valves that move slowly and have a large frictional resistance, the first construction may be preferable, but the cam should be modified by rounding the corners at i and 7, to avoid a shock at starting and stopping. The positions of the lines Ob, Od^ Oa, and Oc may be chosen by the designer so that the time and rate at which the valve opens and shuts may conform to the requirements of his design and to the dictates of his judg- ment and experience. The cam in the figure has a symmetry with regard to the axes xx' and yy' that suggests the resemblance of its action to that of an eccentric. It is neither necessary nor customary to balance valves of the type used on the Brown engine, for they have little press- ure on them to produce friction when they are moving, and when they are shut they are at rest. It is usual and advisable to set the exhaust-valve to give compression nearly up to the steam-pressure in the steam-chest, so that the pressure under the steam-valve is nearly equal to the pressure on it at admis- sion. The valve drops shut at cut-off, and after it is at rest the steam-pressure in the cylinder is reduced by expansion. The expansion is carried down to within a few pounds of the back-pressure, so that at release the pressure on the exhaust- Il8 VALVE-GEARS FOR STEAM-ENGINES. valve is not excessive ; at compression the pressure in the cyl- inder rises after the valve is at rest. A feature common to many detachment cut-off gears can be well shown by reference to Fig. i, PL XXX. Let it be supposed that the eccentric E has no angular advance, and that the valve has no lead ; then the valve will open at the be- ginning of the stroke, and will have its greatest displacement, provided that it is not sooner released, when the piston is at or near half-stroke. If the latch has not then struck the pin d, it will not strike it at all, and the valve will remain connected to the gear, and will close at or near the end of the stroke. It is also evident that giving angular advance to the eccentric and lap to the valve will limit still further the range of cut-off. In this gear, however, the cut-off may be continued beyond half- stroke by giving a negative angular advance to the eccentric and a clearance to the valve. If an engine with a detachment cut-off that is limited to the first half of the stroke is over- loaded, there is a liability that a failure to cut-off will occur, in which case the sudden increase of work due to the steam following the piston to the end of the stroke will make the en- gine run very irregularly. Corliss Valve-gear. — Of all types of detachment valve- gears, that invented by Corliss has been most widely known and has received the most favor. A modification of this gear designed by Mr. Edwin Reynolds is shown by Plate XXXI, which represents the valve-gear on the intermediate cylinder of the triple-expansion engine in the Engineering Laboratories of the Massachusetts Institute of Technology. The Corliss type of engine has two steam-chests, 5 for the supply and ^for the exhaust; the latter is separated from the cylinder in order that it may not be chilled by the exhaust steam. This arrangement produces a somewhat rectangular casting containing the steam-chests and the cylinder, at the four corners of which are placed four valves, two of which, V and V, are admission- or steam-valves, and the other two, W DROP CUT-OFF VALVE-GEARS, II9 and Wj are exhaust-valves. The valve-seats are bored cylindri- cal and the faces of the valves are turned to fit ; the valves bear on half a circle or less, and are so connected to the valve- spindles that they may follow the valve-seats without cramping the valve-spindles. The valve-spindles, which are at right an- angles to the axis of the cylinder, project through stuffing- boxes and carry cranks on the ends, by means of which the valves are turned on their seats. The exhaust-valves IV and w have their valve-cranks WD and wd connected directly to pins A and a, in a wrist-plate Oy which receives a harmonic os- cillation from an eccentric on the engine-shaft. The admission- valves take steam on their inner edges as shown at v, and their cranks carry blocks as shown at the crank end. In the figure a section is taken just behind the crank Vh, which is represented by a dotted line only, in order to show the disen- gagement-latch zTh, which engages the block h and is carried by the bell-crank lever EVT\ the lever EVT is connected by the link EB to the wrist-plate. The latch is opened when the finger Tz strikes the stop x on the ring xr ; the ring is placed under the control of the governor through the cut-off rod NM and the double-armed bell-crank lever Mint, to which the gov- ernor-rod is attached at /. The linkage made up of the valve- crank, valve-rod, and wrist-plate, for example Oadw, is designed to give a slow motion when the valve is closed, and a rapid motion when opening or closing. The figure shows the wrist- plate and valves in mid-position, the eccentric being erect. The exhaust-valve w has its edge on the edge of the port ; its crank moves through the angle dwd^ , while the wrist-pin oscil- lates through the angle aOa^ , but that crank has only the an- gular motion dwd^ while the wrist-plate moves through the angle aOa^ = aOa^. The admission-valves have a similar ac- tion as shown at e^ve^. If it be supposed that the governor- balls are at their lowest position (at which the disengagement- gear does not act) it will be seen that this gear differs from the plain slide-valve gear in two points : first, it has four valves; I20 VALVE-GEARS FOR STEAM-ENGINES. and second, these valves have a more favorable action when opening and closing, on account of the linkages just described. From a pin at g in the crank Vh, a dash-pot rod represented by the line gi leads to a vacuum dash-pot shown by Fig. 2. This dash-pot has two pistons,/ and P\ the lower piston fits nicely in a closed cylinder from which air is excluded ; the upper piston works in a larger cylinder that is open to the atmos- phere through a series of orifices i, i^, and i^ and the pipe O. When the valve-crank Vh, Fig. i, is raised, it lifts the double piston Pp and a partial vacuum is formed under p, while air enters freely, through the orifices 2, i^ , i^ , to the annular space under the piston P. When the valve is disengaged, the weight of the dash-pot and the dash-pot rod, aided by the vacu- um under the piston /, closes the valve promptly ; while the air under the piston Tracts as a buffer and prevents a shock. The pipe O is provided with a hollow plug as shown, by aid of which the escape of air through the orifices i, i^ , and i^ may be regulated. A large number of detachment-gears have been devised and used by Corliss, and by others who have used this type of valve-gear. The one shown by Fig. i, PL XXXI, was invented by Mr. Reynolds and has the advantage that the latch mechan- ism is centred on the same axis as the cut-off stop; conse- quently the finger z always strikes the stop ;i;at the same angle, and the same force is required to disengage the cut-off valve. The block, when disengaged, slides along the plate y. On the return motion the plate y slides over the block till it can snap on to it, under the influence of the spring st. It is customary to give a small lap to the steam-valves ; consequently, as with a plain slide-valve, the eccentric has a small angular advance. With such an arrangement the eccen- tric-centre will be on the line of dead-points, and the valves will have their greatest displacement when the crank has moved through 90° less the angular advance, and before the piston is at half-stroke. If the detachment-gear has not been released r DROP CUT-OFF VALVE-GEARS. 121 before the valve has received Its greatest displacement, the valve will not be disengaged at all, but will remain under the •control of the linkage connecting it to the wrist-plate ; and cut-off will occur near the end of the stroke, and will be deter- mined by the lap and angular advance as with a plain slide- valve. It is therefore evident that the range of cut-off for the ordinary form of the Corliss gear is from the beginning of the stroke to half-stroke or less. When a longer cut-off is desired, for example, on the low-pressure cylinder of a compound engine, two wrist-plates may be used : one wrist-plate, moved by an eccentric with a small angular advance, has control of the exhaust-valves, and gives release and compression near the ends of the stroke ; the other wrist-plate is moved by an eccentric with a negative angular advance, and has control of the steam-valves which have a clearance instead of a lap : with this device the range of cut-off may be extended beyond half- stroke. Expertness in laying out Corliss valve-gears can be obtained only by experience, with good examples for models. The steam-port may be made from J^- to -^-^ of the area of the pis- ton, and the exhaust-port may be made y^ to Jy of that area. The exhaust-valve commonly has no lap ; the admission-valve has a small lap, \ of an inch, in Fig. i, PI. XXXI. In that figure ve is the mid-position of the steam valve-crank, and e' e^ is the maximum valve-displacement, equal to the lap plus the port-opening ; ve, is the extreme position of the valve- crank. The points b and b^ are found by Intersecting the arc bj}^ by arcs drawn from e and e^ with a radius equal to the length of the link BE. The arc bb^ is made equal to bb^ , and ve^ , the extreme position of the crank when the valve is shut, is found by intersecting the arc e^e^ by an arc drawn from b.^ with a radius equal to the length of the link. The linkage Oadw is laid out in a similar way, except that the angle aOa^ is from necessity equal to bOb. The lengths of the valve-cranks and the radii from the centre of the wrist-plate to the pins a 122 VALVE-GEARS FOR STEAM-ENGINES. and b depend partly on the proportions of the engine and partly on the habit and discretion of the designer; the longer they are the less will be the force exerted on the links, ad and be, and on the pins which they connect. The angle veb should be nearly a right angle, so that a rapid opening of the valve maybe obtained. The pin C in the wrist-plate receives motion from the eccentric either ' directly or through a carrier or single-armed rocker that magnifies the throw of the eccen- tric in about the proportion of i : ij. The chord of the arc C^C^j through which C swings, is not longer than the radius OC in order that the angle CfiC may not be more than 30°. The linkages Oadw and Obev are to be laid out by trial to give as nearly as may be the desired motion to the valves. It will be noticed that the radius Oa^ and the link a^d^ are in one straight line at the extreme position of the exhaust- valve ; and in like manner Ob^ and b^e^ are in one line ; should the linkages be carried beyond these positions, a double oscil- lation would be given to the valve-cranks, which is considered to have a bad appearance. The system of rods and levers connecting the rings xqN and vn with the governor is laid out so that the cut-off may come at the beginning of the stroke when the governor stands at the top of its range of motion, and when the governor is at the bottom of that range the cut- off may come at or after half-stroke, i.e. the valve will not be released. Though it is not always done, it will be advisable for an inexperienced designer to draw the valve-ellipse for the steam- and exhaust-valves. The ellipse, or more properly the oval, will have a form like that shown by Fig. 3. The valve will be found to open rapidly and to nearly its full width early in the stroke of the piston. A line nn' drawn at the distance xn, equal to the lap from the axis xx' , will show that the cut- off occurs near the end of the stroke. The valve will be found to be nearly at rest during the greater part of the time when it is closed. The steam-valve is disengaged when the latch holding it is DROP CUT-OFF VALVE-GEARS. I23 released, but cut-off does not occur till the edge of the valve comes to the edge of the port, which is an appreciable time later. In Fig. 3, PI. XXXI, let a represent the point of dis- engagement ; then, under the influence of the dash-pot, the valve falls with an accelerated velocity till it is checked by the air-cushion in the dash-pot. Representing the motion of the piston by abscissae, and the motion of the valve by ordinates (just as in drawing the ellipse), the action of the valve in closing may be represented by the dotted line abc\ the point of cut- off is represented by b at the intersection of this line and the lap-line 7in' . The piston-displacements may be readily found from the dimensions of the engine and Its speed of rotation, but the forces acting on the valve and its resistances cannot be estimated. The forces are the weight of the dash-pot and attached parts, together with the pressure of the atmosphere on the area of the piston/. The resistances are friction of the valve, of the dash-pot, and of other parts of the mechan- ism, and the varying pressures under the pistons P and/; the pressure under P is due to the escaping air, and under/ to the air beneath it when at its lowest position. Though the line abc cannot be determined by calculation or construction, it may be found experimentally by an apparatus described on page 7 for making an engine draw its own ellipse. The action of the valves of a Corliss engine is commonly investi- gated by aid of a steam-engine indicator; if the indicator-dia- gram shows a sharp cut-off, and if the other features are good, the action of the valves is considered to be satisfactory. Putnam Valve-gear. — The Putnam engine has four double poppet-valves, two for admission and two for exhaust. Plate XXXII shows a section through one of the admission-valves, and its valve-gear. XX' is a casting bolted onto the cylinder- casting. The space SS' is the steam-chest, and the space P leads to the cylinder. The two valves V and F", are made of composition and, when closed, rest on composition seats let into the casting. The seat of the valve V is large enough to 124 VALVE-GEARS FOR STEAM-ENGINES. pass the valve V^, so that the valves maybe readily withdrawn through the hand-hole H, The unbalanced pressure, which must be overcome when the valve is opened, is that on the excess of the area of the upper valve over that of the lower valve. The valve-spindle ab is made of iron to avoid unequal expan- sion and consequent leakage ; for, if the distance between the valves is not exactly the same as the distance between the valve-seats, one or other of the valves will not come properly to its seat. The valve-spindle ab is stepped into a frame mn^ shown in section. The arms ^ and gq form a bell-crank lever, one arm of which, under the influence of the spring Ik^ presses on the frame inn\ a pin/' is interposed to reduce friction. The other 3inn, g-f, of the bell-crank lever carries the cam-lever />, which acts on the frame mn through the interposed sliding- block d and the pin / ; this cam-lever is driven by the double cam C. This cam C, and three others, one for the other steam- valve and two for the exhaust-valves, are carried by a shaft which is parallel to the axis of the cylinder and which is driven from the engine-shaft through bevel-gears, so pro- portioned that the cam-shaft makes one turn for two revolu- tions of the engine. The figure shows the cam in contact with the cam-lever, and the valves on their seats ; the engine is consequently at admission. As the engine moves forward, the cam-shaft turns as shown by the arrow and raises the valves, giving admission of steam, till the cam slides past the corner j/ of the cam-lever ; the valve is then released and falls shut under the influence of the spring k/. The governor-rod >takes hold of the pin k at the end of the lever ^^. When the speed of the engine increases and the governor rises, the lever /ig is thrown down and the cam-lever e/ is pushed to the left, so that the cut-off comes earlier. No shock is thrown on the governor when the valve is released, but as the edge of the cam is rounded to avoid cutting the cam-lever, there is a tendency to disturb the governor which the governor must be DROP CUT-OFF VALVE-GEARS. 1 25 able to resist. Should the valve fail to close for any reason, the other end of the cam will strike on i and close the valve before the engine makes a return stroke. The exhaust-cam is shown at A. At each end the cam is cylindrical, so that it holds the exhaust-valve open till near the end of the stroke. The exhaust cam-lever is not placed under the control of the governor, but can be set to give a fixed compression. Gaskill Valve-gear. — One of the steam-valves for the high-pressure cylinder of the Gaskill horizontal pumping- engine, and part of the valve-gear, are shown on Plate XXXIII. The valve is shown in section by Fig. 2, and the seat is shown in section and half-plan by Figs. 2 and 3. The valve is of the Cornish type and differs from the double poppet-valve only in detail. Like that valve it consists of two valves joined together, the inner valve being small enough to pass through the valve-seat of the outer or upper valve. The unbalanced pressure to be overcome when the valve is opened is that on the difference of areas of the two valves. When open, both valves give admission of steam. The valve-seat vS is bolted to the cylinder-casting, and a passage leads directly to the end of the cylinder. The valve is covered by a small cylindrical valve-chest ; there are two such chests, one at each end of the cylinder, supplied by a branched steam-pipe. The valve-gear is shown by Fig. i, in which E is an eccen- tric on a shaft parallel to the axis of the cylinder, and driven from the engine-shaft through equal bevel-gears, so that it makes one turn for each revolution of the engine. The eccen- tric-strap has the cut-off toe a at one end and a lug b at the other. From b the rod bC leads to one end of an equal armed lever, and the valve-spindle d is hung from the other arm ; the distance between the rod C and the valve-spindle is several times as far as shown in the figure. The lever ^7, centred at h, is under the control of the governor through the rod ki. 126 VALVE-GEARS FOR STEAM-ENGINES. The eccentric and eccentric-strap with the lever il form a radial-detachment cut-off gear. Suppose first that the lever hi is thrown so far to the right that the toe a does not touch it ; then as the centre of the eccentric describes a circular path around the point Oy the point b^ of the line b^Ea will move on an arc that sensibly coincides with the axis XX' ^ and the point a^ will describe an oval a^a^a^ ; the valve meanwhile will remain shut. On the other hand, if the lever hi is supposed to be so far to the left that the toe a may remain always in contact with its curved end, and if by some means it is prevented from rising from that surface, then the point a will travel on a circular arc nearly coincident with the axis XX' , and the point b will describe the oval bjbj?^ ; such an action is of course im- possible when the gear is connected up, as the valve is on its seat when the point b is on the axis XX' and consequently b cannot rise above that axis. With the lever hi in the position shown in the figure, the toe a describes the oval a^a^a^ till it comes in contact with the curved surface at the end of the lever ; and then a slides along that surface, while the point b describes the arc hj?^^ and the valve is opened as shown by Fig. 2. When the toe comes to the edge of the surface /, it slips off and the valve is thrown shut by the action of a spring and dash-pot. The toe falls from a\.o a^, and the point b^ returns to b on the axis XX' , It is evident that this form of valve-gear can give a range of cut-off varying from the beginning to the end of the stroke, and that the release does not throw a shock on the governor. On the other hand, the sudden opening of the steam-valve when the toe a comes in contact with the lever hi throws a shock on the valve-gear that might be troublesome at any but the low speeds at which pumping-engines are commonly run. INDEX. Admission, 5 Allen link-motion, 61 Allen valve, 26 Angular advance, 5 Area of steam-pipe and steam-ports, 14 Auxiliary valve-circle, 103 Balanced valves, 27 Bell-crank lever, 11 Brown engine gear, 115 Clearance, 5 Compression, 6 Corliss valve-gear, 118 Crank and connecting-rod, 2 Cut-off, 5 " equalization of, 19 Cut-off valve on back of main valve, 102 " " in separate chest, 97 " '* with constant travel, in ** " with shifting eccentric, 100 Dead-centre, to set engine on, 29 Designing Meyer valve, 106 " link-motions, 67 Diagram, elliptical, 6 " sinusoidal, 7 " Zeuner's, 8 Double-ported valve, 25 Drop cut-off valve-gears, 115 Double valves, 97 '* '* auxiliary circle, 103 Eccentric and eccentric-rod, 3 Equalization of cut-off, 19, 22 Events of the stroke, 5 Exhaust-space, 4 Expansion and compression, 11 Gaskill valve-gear, 125 Gooch link-motion, 53 Hackworth valve-gear, 93 Joy valve-gear, 93 Lap, inside and outside, 4 Lap and lead, for link-motion, 74 Lead, 5, 16 " of link-motions, 52 Lead-angle, 16 Link-arc, radius of, 42, 54, 67 Link-motion, 39 Allen, 61 ** analytical discussion, 43, 54 ** designing, 67 ** for locomotives, 71 " for marine engines, 68 ** Gooch, 53 " lap and lead, 74 ** location of reverse-shaft, 76 " location of rocker, 75 " saddle-pin, 75 " modifications, 65 127 128 INDEX. Link-motion, open and crossed rods, 41, 54 port-opening, 78 skeleton model, 72 slip, 78 Stephenson, 39 to set, 80 Zeuner's diagram, 51, 58 Link-pins, 65 Loose eccentric, 33 Marshall valve-gear, 91 Meyer valve-gear, 104 " " cut-off at inner edge, 108 " " designing, 106 ** *' for reversing en- gines, 109 Model for link-motion, 72 " modifications, 79 ** applications, 81 <( i c (I i( Piston-valve, 24 Port-opening, 78 Ports, 4 " area of, 14 Putnam valve-gear, 123 Radial valve-gears, 88 Radius of link-arc, 42, 49, 54, 67 Reduction of slip of link-motion, 69 Release, 5 Reverse-shaft for link-motion, 66, 76 Rocker, 11 " equalization of cut-off by, 22 " location for link-motion, 75 Saddle-pin, 66 • Shaft-governor, 35, 37 Shifting-eccentric, constant lead, 37 " " variable lead, 33 Sinusoidal diagram, 7 Skeleton model for link-motion, 72 Slide-valve, i, 4 '* " problems, 17 ** " 10 lay out, 20 Slip of link-motion, 78 " , reduction of, 69 Steam-pipe, area of, 14 Stephenson link motion, 39 Stroke, events of, 5 To set a double valve, 114 " ** link-motion, 80 " " slide-valve, 28 " " an engine on a dead-point, 29 Trick valve, 26 Valve, Allen or Trick, 26 '* balanced, 27 " double-ported, 25 *' piston, 24 Valve-circle, auxiliary, 103 Valve-ellipse, 6 Valve-gear, Brown engine, 115 " " cam, 116 " Corliss, 118 ** " double, 97 " *' drop cut-off, 115 *' " Gaskill, 125 " ** Hackvvorth, 93 " Joy, 93 " Marshall, 88 " *• Meyer, 104 " " Putnam, 123 " radial, 88 " Walschaert, 88 Valve-setting, for double valves, 114 '* " for link-motions, 80 '* ** for slide-valve, 28 Walschaert gear, 88 Zeuner's diagram, 8, 51, 58 H A X' Plate V. CO Plate VI. T Plate VI. Plate VII. n Plate VIII. Fig.l -^"^ ^ Fig. 2 Plate YIII. aV^""^""" - A Aj Plate IX. Plate IX. Fig. 2 1 " ; — •^" "■ ■ ^^-^ Plate XII. f Plate XIII. 1 ^ f^ I i \ L ... 1 ^ r y^ \ 1 1 1 1 FiB.. \ g J zoj \ 1 / 1^ ) t JM u . Plate XIV. E- 3 -m m Plate XV. Po Plate XYI. ninmP ^' Q^ X' hob V' X' X' Plate XIX. n "0 ij/pl. -J ll IP' I \ in -X-S'o / \ / \ / N °^^^. Plate XX. Plate XXL \\\ s ' X^ '""X^^^^"^^. 1 V\ \ &^- / i-^^^ / U 1 ' "^-^'?//'/ '/ S 1 ' I ' 1 / 1 1 j 1 \ \ 1 I 1 ^1 1 \ 1 1 1 p '■ \ ' ' ^ - III \ :'^ 1 '' '^ / ' / . I /// ^"'^^^^^ bi) i\ W> /'/ / ' ^ /// / / o^y^ iiu 'l / ( / / 1 ' \ / "/ / / ___\ / { i ^' / ''// / -> 'iS-^ -^^\ s / ] / 1 A// / /' /^^^""^-^ --"Z^ \ / / / / ^^^"''*'--^ ^ Vj / \^ \ ; 1 / / i/ f / / / // / ^^'-'■Y^>cr^""**'^\v\ ' 1 / / l\ > ^ly^ fo'>=^ o / 1 t! \ 1 n \ f'k^^^^^'i \ - ;d^^^-^^j ^y"'^^ ///''' I 1 / 1 \' 1 \ \ \ \ \.^ ■^^^^^^^^^^"-""'^ \ 1 ^/^"^ /O. y // / /' 1 '\ 1 ^ \ '•■W\ ^v^^ """"^ \ y"^ \'/ ^"^)C^- i-^^ . 1 ^ \l ^ \ '\ X"-"""q=\;"^>^^ /'' Ji' / / / w --^ \ bb \ ^^^ / ^fejl" fu \^^ / 1 \ 1 1 ^ ¥' i 'i ^"^-^k_ X ' \ ^ '' '^ \ ^--^ ' ^4-^ / \ / ^~- — 1 -'' X Plate XXII. Fig. 1 Plate XXIIl. V Plate XXIV. Plate XXV. Plate XX VJ. Fig. 2 czzzza e ' ' K — I >f ^ ^^ ttr 07 Fig. 7 "ZZZZ Plate XX V_I. Fig-. 1 Fig. 2 yyyy/yyy/yA sX ^xx-;. \\\\\\^:i^^ Fig. 3 a tto rr' ^ Y ■— -T? ^ N 9^ Rp / y^ FV.^ 7i \ :::::- <—6 ^ ^ ^ \ \ \ \ \ •-- 1 \ V ^ 1 ( > / \ X Fig. 4 X\v \ Fig. 6 Fig. 7 ^^^M '^mmmm^ % mm ■^mmm W/MWd ^ Plate XXVII. \l Plate XXX. Plate XXXII. Plate XXXII. O H ^ #\^^ ^ v^' :/; "oQi ,.% ,0 o ^Q^r. 1 - ? ?5 %^. .'^ V*'- %/:,.., r -^A V .^' ;8^- ^ -.'b^ -v „# o 0^ -n^ V^ ^' V ^/. '^c ^/ u * s ^ .^^^ ^ ^ „ c^^. ^0 .•V %^ %, /■ ,0 <^^ "bo^ - ^ ' i, t, ^ \^ A'" ^ ^^ "" X •>, ''t'