The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

446 subject, and their dimensions were made such as to give ample rigidity. For instance, under full load_ the flexure of the cast-steel floating frame in the vertical plane is so slight that the end bearings will be lowered relatively to the center 'by not more than 1/2000 in. There is, moreover, a certain com- P, TRAE Marine REVIEW beams supporting the floating frame the pinion shaft tilts in a vertical plane parallel to that containing the axis of the large gear. If there were no floating frame, and all bearings of the pinion and gear were cast in one bed plate, the longitudinal dimen- sions remaining unchanged, an error of alignment of the axes in the ver- 'ave Fic. 5--Pinion SHAFT AND FLEXIBLE COUPLING. pensation for this deflection, owing 'to the action of the lubricant. Pro- fessor Osborne Reynold's researches on the action of the lubricating films in Beauchamp Tower's' well-known experiments on lubrication seem clear- ly to ifdicate that at 1,500 revo- ae Cc SIT]! Curve A 8, FLOATING FRAME AND INVOLUTE TEETH, + | CURVE AC, RIGID BEARINGS LE eas g |i a 5 o f i 1 Oot ! : z res 4 LIMIT OF ALLOWABLE OPENING OF CONTACT ZEA Ii a2 i ao ! oO 5 j w 5 Bf hives 2 ee ep NCH A ; ae ) 01.02" 037 04" 057 .067 .077 .08" 09" 10" SCALE OF ERROR OF END GAUGES Fic. 6. minute of bearings will be the present gear' the oil-film . in 1/1000 in, thick, When the load on the gear is increased, the end-bearings will be forced down, bending the pinion and making the teeth bear harder at the Thus more than the proper force will ibe exerted on the center bearing, and less on the outer 'bearings. This will at once lead to a thinning of the oil- film on the center bearing, and to its thickening on the outer bearings. As the total thickness of the oil-film is more than twice the yie'd of the outer bearings, this change in the thickness of the film will go far to compensate for the elasticity of the floating frame. S Given that the axes of the gears are parallel to start with, the action of the teeth will, it is claimed, force the axis of the pinion into exact alignment with that of the large gear, provided that in the flexure of the I- lutions _the pinion in well = over per mer ends 1 and "(". tical plane not exceeding 1/1000 in. in the length over 'the bearings would entirely upset the uniform distribu- tion of the pressure along the faces of the teeth. Hence, with the: most careful work, the above assumptions may not be fulfilled with sufficient accuracy. Even if they were exactly fulfilled when the gear was first com- pleted, inequality of the side wear of the bearings might cause the axes to stand at a senstble angle in . the horizontal plane. Provision is made for measuring such errors of alignment as follows: Outside, the end bearings and both pinion and gear, as shown at gl, g2, g3, g4 in Figs. 2 and 3, there are vertical and horizontal gages placed so that errors: of alignment can be readily measured when the gear is running. lithe horizontal gages at g3, should show the axes too close by, . say, 1/10 in, and those at 64, shovld show them too open by 1/10 STEAM PER HORSE POWER HOUR Yj "Wii YY WW) WU My, yy My REVOLUTIONS OF TURBINE PER MINUTE Fie: 7. in., giving a total error of alignment of 1/5 in. in length between: the gages, it may be shown that, with involute teeth, which are here used, they will bear hard at I and I, Fig. @ and lightly .at. J. and J If the teeth were perfectly rigid, there would of course, be point contacts November, 1909 at I and I', and the teeth would ac- tually stand open by a small amount at J and J'. This opening Messrs. Melville and Macalpine call the "opening of contact." If in changing from perfect align- ment the pinion turned about either a vertical or a horizontal transverse axis, it may readily be shown that 'the opening of contact would be a considerable fraction of the sum of the errors shown by the end gages. On the other hand, the inventors state that the action of the floating frame in combination with involute teeth is such as to constrain the pin- ion to turn about an axis parallel to what is known as the line of action of the teeth--in the present case, an axis inclined 14 deg. 30° min. to the vertical--and this is precisely the movement which reduces the opening of contact to a minimum. The result is shown in Fig. 6. THe error of each end gage is indicated by the 'horizontal. scale," the total length representing 1/10 in. corre- sponding to..a..total error. .of align- ment in the horizontal plane of 1/5 in. between the gages, as explained above. The vertical scale shows the "opening of contact" if the teeth are rigid, this scale being enormously magnified, as the quantities to be rep- resented are so emiall. The curve A B represents the resultant opening of contact -- forgiven gage errors for the present gear. The extreme opening of contact shown is 1/6500 in. and this for a gage error which should never be approached. -At half this gage error, still large and easily measureable, the opening of contact is reduced to one-quarter of the fore- going, or 1/25000 in., a quantity far below the limit of accuracy of the very finest machining. Consider next the effects of the elastic strains of the teeth and shaft- ing. Steel. gears of 114 in. pitch are ¢eommonly run under a load of 1,000 Iy per in, of. tooth face at - ordi- nary pitch-line speeds. As all gears are liable to errors of alignment, this average pressure actually means that at times there will 'be a maximum pressure very much greater. In the gear under discussion the _ highest mean value. used. is -less than 500 Ib. per in... of -tooth. contact. _Sup- pose. then that by.errors of align- ment the actual value varied from zero: to 1,000- Ib. . per which is far within the disturbance in ordinary gears. It may be shown that the yield of these teeth under 1,000 Ib. per in. is such that the dis- tribution of pressure would close up in., occurring