Cylindrical Grinding Machine

Norton claims rigidity as the route to accuracy — the opposite of the intuition that delicate work needs a delicate machine. Write down why you would expect the opposite.

Hold the steel ruler flat, press its end sideways. It bends easily. Now turn it on edge and press again. Same steel, far stiffer. Shape beats material.

Clamp the ruler flat over the board's edge and press with a known steady push; measure the deflection with the dial indicator.

Repeat with the ruler on edge. Record the much smaller deflection. This ratio is why Norton's machine was heavy.

Sand the brass rod by hand for one minute, turning as you go.

Measure the rod at four points around and along with the micrometer. Record the spread.

Glue two blocks to the board so the rod is supported close to where it is being worked. Support near the cut is the whole trick.

Sand again, this time with the rod held in the rest and rotated steadily.

Measure the four points again. Compare the spread with step 6. Same abrasive, same hand, better geometry.

Roll the rod under the dial indicator on the surface plate. The needle sweep is out-of-roundness.

Remove one support and work the rod again. Measure. Chatter and taper appear immediately.

Set a feeler stack to your target and test whether the rod passes. Grinding without a gauge is just polishing.

Set a target such as ±0.05 mm and work until every measured point falls inside it. Record how many attempts it took.

US 762,838, “Grinding-Machine”, Charles H. Norton, Worcester, Massachusetts, filed 1 February 1901, granted 14 June 1904, assigned to the Norton Grinding Company.

A date correction. Popular accounts date this machine to 1900 and often imply the patent is contemporaneous. 1900 is when Norton designed it; the application was filed in February 1901 and took more than three years to issue.

What it did to industry. Norton's grinder could finish a cylindrical part far faster than anything before it while holding about a quarter of a thousandth of an inch. Henry Ford bought them in quantity for Model T production. Grinding a crankshaft or a bearing race to a repeatable diameter is precisely what makes engines mass-producible — the Timken tapered roller bearing linked here is one of the parts that depends on it.

Why rigidity beats delicacy. A cutting force pushes the tool away from the work. If the machine flexes, the cut is shallower than commanded, and worse, the flex varies as the force varies, so the error is not even constant. A heavy machine deflects less, so what you command is what you get. This is one of the genuinely counter-intuitive results in engineering, and it is why precision machine tools are massive lumps of cast iron rather than elegant frames.

Where it sits. Johansson's blocks, linked here, define the size. Norton's machine produces it. Without both, a tolerance on a drawing is a wish.