The angle of the valve seatings should be 45° to the vertical axis of the valve stem. If made more acute, there is some risk of the valve sticking in the seat; if much flatter, particles of carbonaceous matter may adhere, and so prevent the valve closing properly. The width of the actual seating may be about equal to half the thickness of the head of the valve, or 0.05 times the diameter of the valve opening. The wider the seating, in reason, the longer the valve will work without it being necessary to re-grind the seating. Also the pitting and erosion due to the rapid passage of the hot products of combustion is much more pronounced with narrow than with wide seats. The valves themselves are usually made of one piece of mild steel, but in the case of large valves the head is sometimes made of cast iron, or even nickel alloy, which is screwed and riveted to a mild-steel stem. This is claimed to make a more durable valve than one constructed entirely of steel, but the writer's observations go to show that the greater amount of pitting and erosion take place on the valve seat, and that a mild-steel valve head is practically as good as one of cast iron. It might be expected that nickel steel would give the best all-round results as a material for exhaust valves, but the writer has no data on this point. When the head of the valve is made separate from the stem, there will be a possibility of the head becoming loose on the stem, which is entirely avoided by the one-piece valve.

It is of the utmost importance that the guide, in which the stem of the valve works, should be perfectly concentric

with the valve seat. The practice of making the valvestem guide separate and screwing it into the valve box is to be deprecated, as it usually results in the guide being eccentric with the seat. Even if made true to commence with, the expansion and contraction due to the changes in temperature will, in the majority of cases, cause the guide to become eccentric with the seat sooner or later. If cast integral with the valve box, the guide can be made true with the seat once for all, and will remain so. If, however, the valve box is insufficiently water jacketed, or is cooled on one side more than another, there will be considerable risk of the valve-stem guide being warped when the valve box is heated.

Next to having the valves the correct size, the matter of timing their operation takes an important position. The exact moment at which the exhaust valve should open depends, for the most part, on the piston speed. A motor running with a high piston speed will require to have the exhaust valve opened considerably earlier in the cycle than when the piston speed is low. There is not much. data available on this point, but as a guide it may be stated that with a motor having a piston speed of 700 feet per minute, the exhaust valve should commence to open when the piston has completed eight-tenths of its stroke, and close when the piston has started on its suction stroke, and not exactly at the dead centre. The reason why the valve should be late in closing, although this is contrary to usual practice, is that with a high piston speed the products of combustion will not have all escaped from the cylinder at the termination of the exhaust stroke. The exact amount by which the closing of the exhaust valve should be delayed will best be determined by experiment. At slow piston speeds of, say, 500 feet per minute, the exhaust valve may be closed exactly at the dead centre. The writer has observed a decided improvement in power,

in the case of high-speed motors, when the closing of the exhaust valve has been thus delayed, hence it is reasonable to assume that the products of combustion are not entirely expelled by the time the piston has finished its in-stroke. From consideration of this point the writer has for some time been of the opinion that the mechanically operated inlet valve, as applied to high-speed engines, is a mistake. The usual practice is to open the inlet valve immediately the exhaust valve closes, without reference to the pressure existing in the cylinder. Thus if the pressure within the cylinder is above atmospheric at the moment the inlet valve is opened, in place of a fresh mixture flowing into the combustion chamber, there will be a rush of the residual, burnt gases into the carburettor. Before any

fresh mixture of air and gas can be taken into the cylinder, these burnt gases must be drawn back through the inlet valve, and the greater the piston speed, the more will this effect obtain. With a motor required to run at a low piston speed, the mechanically operated inlet valve undoubtedly gives better efficiency than the automatic valve, and in point of fact the chief claim made by the advocates of the mechanical valve is that it enables the motor to be run at a much slower speed than when it is equipped with automatic inlet valves.

The most rational method of overcoming the defects of the automatic and mechanically operated inlet valves would seem to be that in which the moment at which the valve opens should be determined by the pressure existing in the combustion chamber. In the valve gear invented by Mr. R. E. Phillips, the pressure of the spring which holds the inlet valve on its seat is relieved a short time before the completion of the exhaust stroke, and the inlet valve is kept closed by the pressure of the gases in the cylinder. When the pressure in the combustion chamber falls to that of the atmosphere the inlet valve is free to

open, either by its own weight, if inverted, or by the suction effect of the piston, without the restraining influence of any spring. A light spring may even be used to assist the valve in opening. Prompt closing of the valve is ensured by the valve spring being allowed to again resume its function. With automatic inlet valves the spring tension is a matter requiring careful adjustment. If too strong, the valve will only open, and remain open, while the pressure in the cylinder is below that of the atmosphere to an extent depending upon the strength of the spring, resulting in small charges and a consequent lowering of the compression pressure. With a weak spring the valve will open with a very slight vacuum in the combustion chamber, and thus full charges will be assured, but the closing of the valve may be so delayed that the greater portion of the charge will be returned into the carburettor. At the best only a compromise is possible, and the general tendency is towards using a fairly strong spring, and rightly so, as the lesser of two evils. The time taken by the inlet valve in closing, especially with high-speed motors, is important. This time may be calculated from the following formula :

in which S = the time in seconds, L = the lift of the valve in inches, W = the weight of the valve, and M = the average pressure exerted by the spring. W and M must both be taken in the same units, either ounces or pounds. Taking, for example, an engine running at 800 revolutions per minute, with a valve weighing 6 ounces, and having a lift of, say, = 0.34375 inches. Assuming the spring selected to have an average tension of 12 ounces, and substituting known values in the formula 21, we have-

= about 0.03 second

S = 0·0721 / 0·344 × 6

60

12

At 800 revolutions per minute the engine would make one revolution in 6% 0:075 second, or one stroke in 0.038 second; that is, the engine would make nearly one complete stroke while the valve is closing. Evidently a much stronger spring is required. To calculate the size of spring to be used the formulæ given by Professor Unwin

n

FIG. 7.

will be found of great utility. For the force required to compress, or extend, the spring, we have

in which F the force necessary to compress (or extend) the spring one inch, in ounces; d= the diameter of the wire in inches; r = the mean radius of the coil in inches; and n the number of coils. These proportions are graphically illustrated in Fig. 7.

The formula for the safe working load on the spring, in pounds, is