When a lead screw is long it requires to be supported, otherwise, either its weight will be supported or lifted by the feed nut in gear, or if that nut does not lift the screw, the thread cut will be finer than that due to the pitch of the lead screw, by reason of its deflection or sag.

A lead screw should preferably be as near as possible to the middle of the lathe shears, and as close to the surface as possible, so as to bring it as nearly in line with the strain on the tool as possible, but on account of the cuttings, which falling upon the screw would cause it to wear rapidly, it is usual to locate it on one side, so as to protect it from the cuttings. It is better to locate it on the front side of the lathe rather than on the back, because the strain of the cut falls mainly on the front side (especially in work of large diameter when this strain is usually greatest) and it is desirable to pull the carriage as near in a line with the resistance of the cut as possible, because the farther off the feed nut from the cutting tool point, the greater the tendency to twist the carriage on the shears.

To preserve the nut from wear, it should be made as long as convenient, as, say, five or six times the diameter of the lead screw; it is usually made, however, three or four diameters.

It is obvious that the pitch of the thread should be as accurate as possible, but it has not as yet been found practicable to produce a screw so accurate that it would not show an error, if sufficient of its length be tested, as, say, several feet.

Fig. 620.

If the error in a screw be equal, and in the same direction at all parts of its length, various devices may be employed to correct it. Thus [Fig. 620] represents a device employed by the Pratt and Whitney Co.

It was first ascertained by testing the lathe that its lead screw was too short by 7100ths of a revolution in a length of 2 feet, the pitch of its thread being 6 to an inch. Now in 2 feet of the screw there would be 144 threads, and since 7100ths (the part of a revolution the thread was too short) × 16 (the pitch of the thread) = 7600ths (which was called 185th), the error amounted to 185th inch in 144 turns of the screw. The construction of the device employed to correct this error is as follows: In [Fig. 620], a represents the bearing of the feed screw of the lathe, and b b a sleeve, a sliding fit upon a, prevented from revolving by the pin h, while still having liberty to move endways. c represents a casing affording journal bearing to b b, having a fixed gear-wheel at its end c′, and an external thread upon a hub at that end. d is the flange of c to fasten the device to the shears of the latter, being held by screws. e represents an arm fast upon the collar of the feed screw, and carrying the pinion f, the latter being in gear with the pinion c′, and also with g, which is a pinion containing two internal threads, one fitting to b at b, and the other fitting to c at c, the former having a pitch of 27 threads to an inch, the latter a pitch of 25 to an inch.

The operation is as follows:—The ordinary change wheels are connected to the feed screw, or lead screw, as it is sometimes termed, at j in the usual manner. The arm e being fast to the feed screw will revolve with it, and cause the pinion f to revolve around the stationary gear-wheel c′. f also gears with g. Now, f is of 12 diametrical pitch and contains 26 teeth, c′ is of 12 diametrical pitch and contains 37 teeth, and g is of 12 diametrical pitch and contains 36 teeth. It follows that the pinion f, while moving around the fixed gear c′, will revolve the pinion g (which acts as a nut), to an amount depending upon the difference in the number of its teeth and those of fixed gear c′ (in this case as 36 is to 37), and upon the difference in the pitches of the two threads, so that at each revolution g will move the feed screw ahead of the speed imparted by the change gears, the end of the sleeve b abutting against the collar of the feed screw to move it forward.