The error in putting the crank on may occur from the key springing the crank out of true, and if the crank is shrunk on from too great an allowance for shrinkage or improper heating for the shrinkage or contraction, as it is sometimes termed. Referring to the error in keying, it is more liable to occur when the crank bore and its seat upon the shaft are made taper, than when made parallel, because it is a difficult matter to insure accuracy in the fit of the taper, and the key pressure will spring the crank over on the side at which it is the easiest fit. In [Fig. 2543] let a represent the end of the crank shaft; b the key, and c the crank shown partly in section: suppose the crank bore (whether made taper or parallel) has a slightly easier fit on the side d than on the side e, and the pressure of the key (supposing it to fit properly top and bottom) would spring the crank over in the direction shown in the figure, the axial line of the crank pin standing at the angle denoted by the line f, instead of parallel to the axial line of the shaft. Suppose the crank to be put on by hydraulic pressure, and the key to fit on the sides and not on the top and bottom, then its fit to its seat on the shaft would depend on the truth and smoothness of its bore and seat on the shaft, the amount allowed for the forcing fit and the amount of the error. If the latter amount was so small that the crank would fit at both ends, but simply fit tighter at e e than at d, the crank would remain true, but might possibly get loose in time. This would be especially liable to occur if the tool marks on the bore and seat were so deep that the contact was mainly at the tops of those marks or ridges which would be apt to compress. But if the surfaces were cylindrically true and smooth, and the amount allowed for forcing was sufficient as stated to give the bore and seat contact at d, with a key fitting sideways, the crank would probably remain tight and true.
Were the bore and its seat parallel the crank would remain true, no matter whether the key fitted on the sides or at the top and bottom, providing the key fitting top and bottom were bedded fairly from end to end.
When the surfaces are not smooth, but contain tool marks or ridges, an unequal pressure of the key at one end, as compared to the other, sets the crank over, as shown in the figure, because the key pressure compresses the ridges and lets the crank move over.
Fig. 2544.
Supposing the strain of the key, or keys, to be depended upon to hold the crank, they must fit top and bottom, and their accurate fit becomes of the first importance; because not only is it necessary that they fit equally at each end, but they must also fit equally across the width of the key at each end. For example, in [Fig. 2544] is a key binding most at the opposite corners, as denoted by the dotted surfaces a b, and the result will be that the key pressure would tend to twist the crank in the direction of d e, having c as a centre of motion, providing that the error was equal at a and b; but in proportion as the error was greatest and the fit tightest at a, or at b, would the centre of motion be moved nearer to either point.
Supposing now that the crank is to be shrunk, or contracted on, then the points of consideration are (supposing the crank to fit properly to its seat, whether the same be either parallel or taper) that the hub of the crank opposite to the throw is the weakest and is likely to give most in the process of contraction, so that if one part (as f) of the crank be made hotter than another (as g) it will give way more, and this will twist the crank. This is specially liable to occur if an excessive amount of difference in the bore and seat diameters has been allowed for contraction.
