A different method of carrying out the principle of grinding the rollers together with water is sometimes resorted to, in order to allow of the rollers being subjected to the same degree of pressure during grinding, that they are intended to sustain when at work, as it is occasionally found that notwithstanding the strength of the rollers, they yield slightly beneath great pressure, so as to interfere with the accuracy of contact.

To avoid this interference the bearings of the upper roller instead of being suspended over the lower, so that the high points of the two rollers alone touch, are loaded so as to press the rollers in contact with the same degree of force that is required for pressing the paper. But under this pressure the surfaces in contact do not admit of being driven in opposite directions, because it is found that the two surfaces meeting each other, cause so much friction that the rollers are almost certain to be torn even when a very slow motion is employed. The rollers are therefore driven in opposite directions at different velocities, generally in the proportion of nearly 5 to 6, so that the surfaces in contact travel in the same direction, but the velocities being different they move over each other with a sliding action.

The adjustment of the velocity depends principally upon the degree of pressure employed, and the condition of the surfaces; if driven too rapidly the surfaces are liable to heat and tear each other, which in this as in the arrangement previously described is the principal difficulty to be contended with.

The water is supplied through a perforated tube extending the length of the rollers, and should any portion of the rollers appear to be grinding too rapidly, the action may be checked by stopping up some of the holes to reduce the supply of water at that part, but which is not generally resorted to, owing to the risk of the rollers being allowed to become too dry from neglect on the part of the attendant.

The grinding action appears to be principally due to the small particles of cast iron rubbed off by the friction, and which serve as the abrading powder. The progress of the grinding may be expedited at the commencement, by using the same water repeatedly over again, in order to bring a larger quantity of the grinding powder into action, but towards the conclusion of the process when the highest finish is required, clear water is alone used.

The process appears to be partly grinding and partly burnishing, and does not admit of being indefinitely pursued, as if continued too long the surfaces crumble away, which is also liable to occur if the castings be unsound, and therefore all such places should be plugged up with cast iron of the same quality and hardness as the rollers. Wrought iron should never be used for the plugging as it is but little acted upon by the water grinding, and the wrought iron plugs would stand out beyond the general surface of the rollers.

Accuracy of surface contact is the object desired in these rollers, and their absolute straightness is a matter of secondary importance, the rollers are therefore made simply to revolve upon their axes, and are not at the same time traversed through their bearings, as this would be liable to introduce a new source of error by wearing the journals into a screw-like form. A small amount of end adjustment is however sometimes adopted, should the rollers be found to wear into rings; with this view the bearings are so far separated as to allow of a little end motion in the journals, and the insertion of washers between the collars and bearings, allows of the rollers being shifted endways a small distance when required. This adjustment is however scarcely called for, as without it a pair of rollers may be ground so nearly straight, that the ordinary test of a straight edge would fail to detect any irregularity, and when three or more rollers are ground with their surfaces in contact, they mutually correct each other for straightness as well as circularity.

The cylindrical rims of pulleys employed for driving machinery by leather straps are usually turned to form in the lathe, and afterwards smoothed with emery applied on a stick as explained at page 1072. But in some cases these pulleys are wrought into the cylindrical form by the ordinary grindstone, assisted by a little mechanism, after the same general method as that employed for grinding superior cylindrical works. In the case of pulleys the grinding is resorted to not from any superiority in the method, but solely from motives of economy, the grindstone being more rapid in its action than the turning tool, when the object is merely to produce a level surface, without removing a greater bulk of the material than is necessary for that purpose, as the turning tool requires to penetrate sufficiently deep into the metal to remove the outer hard crust left in casting, as explained on page 375, Vol. I. The action of the grindstone is however little influenced by the hard crust, and consequently a much smaller quantity of material has to to be removed by grinding to produce the cylindrical form.

In Mr. James Whitelaw's machine for grinding pulleys,* the grindstone of about 4 feet diameter is mounted in fixed bearings as usual, and revolves about 180 times per minute, the pulley to be ground is fixed upon a mandrel parallel to the axis of the grindstone, and makes about 130 revolutions in the minute in the same direction as the grindstone, so that when the opposite edges of the pulley and grindstone are brought into contact, the two surfaces rub upon each other at their combined velocities, and at the same time the pulley is reciprocated a few inches backwards and forwards across the face of the grindstone, to equalise the wear of the latter, and ensure the cylindrical form of the pulley.

* Described in a communication to the Royal Scottish Society of Arts, 1838. See Trans. Roy. Scot. Soc. of Arts., Vol. I., page 235.