Slide Valves by Professor Green
by B. M. Green
July-August and September-August 1954
When I started some years ago, to build a miniature steam locomotive I decided to design it from the ground up using good engineering, or model engineering, principles. I selected slide vales, chiefly because of ease of manufacture. The next step was an investigation of suitable port widths and lengths and desirable areas of steam passages. Perhaps others would be interested in what I found.
Of two chief sources of information, one was "Slide Valve Gears" by Frederick A. Halsey, associate editor of American Machinist for many years. This little book was first published in 1889 when the interest in steam engines and in the design of slide valves was at its height. My copy is the thirteenth edition, dated 1920, and the book went our of print a few years later. It contains much of interest to the steam engine enthusiast. The other chief source was the writings and designs of our good friend "L.B.S.C.". he has never said much about how he designs his engines and i trust he will forgive me for analyzing a few of his designs to determine average proportions. There are also many articles by experienced designers of miniature locomotives in the old and lamented Modelmaker.
The slide valve is, of course, the means by which live steam is admitted to the engine cylinder and released from it at the proper times. If one plots pressure in the cylinder against piston position a diagram is obtained called a "card." Cards can be drawn for actual engines by an instrument called an indicator. Idealized cards for one end of an engine cylinder are shown in Figure 1.
Figure 1a shows what happens when an engine is operating at about 75% cut-off and Figure 1b illustrates the situation for about 25% cut-off. Since work equals force times distance the area within the complete curve times the piston area measures the work done by the steam during one stroke. It is evident that the work done at the earlier cut-off is less than that done at the later valve but, since a locomotive is only "hooked up" at the higher speeds, the work done per minute, or horsepower, may be greater for the earlier cut-off.
The "events" of a card are: *(1) Admission, when the valve opens to steam; (2) cutoff, when the valve closes to steam; (3) Release, when the valve opens to exhaust; (4) Compression, when the valve closes the exhaust. Looking again at Figure 1, the space from zero to the beginning of the stroke is proportional to the clearance volume in the cylinder, that is, the volume between piston and cylinder head plus the volume of the steam occupying it does no work on the piston and so is wasted. It is evident then that clearance volume should be as small as possible without getting the passage so small that there is a large pressure drop. The line marked "Atm" represents atmospheric pressure and the actual exhaust pressure will be a little steam passage up to the valve face. Clearance volume is expressed as a percentage of the stroke volume and in full size engines may be as much as 20%. The higher because of resistance at the blast nozzle. Halsey shows cards taken on locomotives which indicate a back pressure of about five pounds.
Comparing the card in Figure 1b with that in Figure 1a it will be noticed that cut-off and release are earlier in the forward stroke and compression and compression is earlier in the return stroke. These limitations of the plain slide valve must be accepted. The early release represents a waste of steam. The early compression means that more steam is trapped and compressed in the cylinder at the end of the return stroke. The resulting higher pressure may be a good thing because it aids in absorbing the energy of the faster moving piston and bringing it to rest at the end of the stroke.
Now, what about dimensions of ports and passages? Halsey has quite a bit to say about assumed average velocities of the steam through the ports and passages and this may be a good way when designing single speed stationary engines but he admits, himself, that the method can't be applied to locomotives with their widely varying piston speeds and cut-offs. A better way, for us at least, is probably to analyze a few locomotives and determine some proportions. Start by scaling the bore and stroke of the engine we wish to follow. Our bore is likely to be smaller than scale because our cylinder head screws are usually over scale and so require thicker cylinder walls. When the cylinder diameter is decided upon we can express port widths and lengths in terms of this diameter.
Figure 2 represents the port face for a cylinder using a slide valve. L is length of port (across the cylinder); S is steam port width; E is exhaust port width; and B is thickness of bridge wall. Let D be cylinder diameter. Halsey gives the dimensions for nineteen locomotives and also proportions of average stationary engine practice. Take first, the port length, L. In Halsey's locos L. varies from 0.85D to 1.0D with an average of 0.9D. In stationary engine practice L in often 0.75D. Halsey states that the ports of British locomotives are smaller for a given size of cylinder than are those of American engines and thinks that may be one reason why British engines are claimed to be more economical. (When we write). I have no data on British engines but I did investigate four of L.B.S.C.'s engines and in these L averages 0.57D. This small ratio may be partly due to over-size valve chest screws which reduce the available width inside the chest. However, a large L seems desirable. The steam pressure in the cylinder should rise rapidly as possible when the valve is just opening to steam and the longer the port the greater the available port area. I am not arguing for lengths greater than L.B.S.C's but the reasoning indicates why some amateur designs using a single drilled hole for a steam port will not pull anything.
Next, consider the width, S, of the steam port. Halsey's locomotives vary from 0.06D to 0.087D with an average 0.11D. Stop a minute to consider port area. Halsey's average length times width gives an area of 0.066D^2 while L.B.S.C.'s values give 0.063D^2, surprisingly close together. Stationary steam engines average about 0.15D for width. halsey's tabulation of locomotives gives no information on the width, E, of the xhaust port. His team velocity values would seem to indicate that it would be about 0.10D. L.B.S.C.'s locomotives average about 0.26D and E. T. Westbury uses 0.33D on his small stationary engines. When comparing all these figures it must be remembered that it was customary to make the ports on stationary engines considerably larger than those on locomotives, probably because the former usually exhausted at lower pressures and, therefore, the volume of steam was greater. A pretty good rule for small lcos is to make E equal 2S, or slightly greater. Regarding thickness, B, of bridge wall, Halsey suggests that it equal cylinder wall thickness on full size engines but this gives more than is needed for miniature engines and in such designs B is often made equal to S.
One more dimension is of interest and that is the area of the passage from the steam port to the end of the cylinder. I use area instead of linear dimensions in this case because this passage is often drilled in small engines and area will cover both size and number of holes. The average of Halsey's locomotives, which of course had rectangular passages, works out to 0.07 times piston area. The average factor for L.B.S.C.'s locos is 0.045 and for one of Mr. Westbury's engines it is 0.11.
So much for the ports and passages in the cylinder. You can see that the figures indicate what has been found to work; there is not theory involved. If, on the one hand, ports and passages are too large steam is wasted, and if they are too small there is to much pressure drop and the engine has too little power. Apparently very few experiments have been made on full size locos to determine optimum proportions.
Now that we have some basis for determining proper sizes of the ports and passages, let's consider the valve itself; how it acts and what its dimensions should be. Suppose we start by assuming the simplest possible mechanism to operate the valve. Both piston and valve will be driven by cranks of suitable length moving in slotted yokes attached to the piston rod and valve rod respectively, as illustrated in Figure 3 (a). Such a mechanism eliminates the distortion of the crank motion which is introduced by a connecting rod or eccentric rod and the "slotted crosshead", or "Scotch yoke" as it has been called, was actually used in early engines and survived until quite recently in steam fire engines because it is economical in space.
Looking at Figure 3a the piston is shown at the left end of the cylinder, with rotation indicated clockwise, and the valve is designed to just cover the steam ports. Note that the valve crank (or eccentric) leads the main crank in the direction of rotation by just 90 degrees. If the crankshaft is rotated slightly to the right, as in Figure 3b, the valve will move to the right and open the left hand steam port, admitting steam to the left hand side of the piston so that the steam pressure will act to continue the rotation. If, on the other hand, the crankshaft were rotated a little in the opposite direction from the dead center position, the piston moves to the right but the valve moves to the left, opening the right hand steam port and admitting steam to the right side of the piston. In this case the steam pressure pushes the piston to the left and the crankshaft moves clockwise. Thus, the valve crank is set correctly for clockwise rotation, as marked. With the primitive type of valve shown steam is admitted for the entire stroke and its expansive energy is not used so such a valve is wasteful.
Considering the exhaust side for a moment, Figure 3a shows that the cavity in the valve just seals both steam ports. In Figure 3b, with the valve to the right the right hand steam port is open to exhaust, as it should be for right hand rotations and in Figure 3c the left hand when the crank has moved 90 degrees more, bringing the piston to the right hand end of its stroke. The valve has now reversed direction and has moved half its travel to the left so that it just covers the steam ports once more and is ready to open the right hand port to steam.
Figure 4a shows the piston crank rotated 90 degrees, bringing the piston to mid-stroke. The valve has moved its entire travel to the right, opening the left hand steam port completely. Figure 4b shows the steam port is open to exhaust.
If a steam engine is to run economically the expansive energy in the steam must be utilized and this is done by arranging the valve to close the steam port before the end of the piston stroke. Two alterations in the primitive valve are ne3cessary: (1) the valve is given steam lap (or lap), and (2) the eccentric is advanced around the shaft in the direction of rotation beyond the 90 degree position already shown.
Figure 5a shows the change in the valve. The steam lap is the amount by which the valve extends past the edge of the steam port when in mid position. Figure 5b shows the eccentric moved through an additional angle sufficient to open the steam port at the beginning of the stroke with the new valve. note that the distance a-c is equal to the steam lap. When the valve closes the steam port it is in the same position as shown in Figure 5b but is then moving to the left instead of the right. Then, for cut-off the eccentric center must be at e directly below c and, during the time steam was admitted to the cylinder the eccentric (and with it the shaft and crank) must have turned through the angle cbe which equals a semi-circle less twice the lap angle. So, during the admission of steam the crank turns through 180 degrees less twice the lap angle. This means that cut-off occurs before the end of the stroke and the steam is used expansively for the remainder (or most of the remainder). Cut-off can be made earlier by increasing the steam lap.
Looking back at Figure 4a we see that the eccentric throw, (crank radius), is just enough to cause the primitive valve to open the steam port completely. In Figure 5b the valve does not begin to open the steam port until the eccentric center has reached point c so only the remaining motion, c-f, is available for port opening. Therefore, to get the same port opening, the valve with lap must be driven by an eccentric with a greater throw.
So far, we have been redesigning the primitive valve by adding steam lap and moving the eccentric forward to compensate, being guided by the desired steam evens. What has this action done to the exhaust events? With the primitive valve the steam port was opened to exhaust (called release) at the end of the piston stroke. With the improved valve the eccentric has been advanced on lap angle so release must now occur one lap angle before the end of the stroke. For the same reason compression now occurs one lap angle before the end of the return stroke. The exhaust events may be altered by modifying the inside, or exhaust, edge of the valve exactly as the outside, or steam, edge has been modified to alter the steam events. If material is added to the inside edge (called exhaust lap), making the cavity shorter, release is delayed, that is, it occurs later in the stroke. Compression, on the other hand, will occur earlier in the return stroke. From the standpoint of economy, the effect of exhaust lap is much less important than is the effect of steam lap. In full size engines exhaust lap may be positive, zero, or negative, depending on the results desired. In miniature engines it is most often zero, largely because of the difficulty of measuring accurately the very small valves needed; that is, the width of the valve cavity is exactly equal, or should be, to the distance between the near edges of the steam ports.
In the foregoing discussion, for the sake of simplicity, the valve and eccentric have been designed so that admission occurs as the crank passes dead center. Many designers of full size engines think that admission should occur slightly before dead center so that full pressure will become effective behind the piston as early as possible. If admission occurs before dead center then the valve is open an appreciable amount when the crank is on center. This amount of opening is called LEAD and it is measured in fractions of an inch. Lead is obtained simply by moving the eccentric forward around the shaft until the valve has moved from mid-position the desired amount. The total angle the eccentric has been advanced beyond ninety degrees ahead of the crank now includes lap plus lead. This angle is called the advance angle. The proper amount of lead is always a debatable point, especially with the Stephenson type of gear because this mechanism increases the lead as the link is brought toward the center to reduce cut-off. Halsey gives several "Link Charts" for particular engines. Here is part of such a chart, dated 1896, for engine No. 442, N.Y., N.H. & H RR:
- Cylinders: 20" x 24"
- Drivers: 73"
- Valve travel: 6"
|Lead||Valve opens||Cut-off inches||Cut-off percent|
Halsey also gives link charts for other locomotives showing negative lead at latest cut-off and says that this is done so that lead will not become too large at early cut-offs. In the case of miniature locomotives proper lead becomes so small that it is often not considered. That lead of 1/16" full size becomes only 0.004" for a scale of 3/4" to one foot. L.B.S.C. and other writers advise that the port edge should be just visible with the crank on center. This probably gives two or three thousands thead as the port edge could not be seen if exactly line-and-line with the valve edge.