MACHINING MONDAY

From Cigarette Paper to a $15 Brass Tool: The Edge Finder You Trust Every Setup

You dropped an edge finder into the collet this morning. Spun it up at 1,000 RPM. Watched the tip wobble, snap, then kick. Half the diameter, zero X, move on.

You did not think about it. You did not check the math. You did not look up the offset. You stopped checking ten years ago, the first week someone showed you how the kick worked.

That little brass-and-steel tool in your drawer is older than the CNC mill you set up on this morning. It outlived every electronic replacement that was supposed to retire it. The reason it is still in every shop in 2026 is not nostalgia. It is something the documentation never explains.

The Moment of the Kick

A wobble-type edge finder is two pieces: a cylindrical shank that goes into the collet, and a free-floating tip held to the shank by an internal compression spring. The tip is mounted off-center by design. When you spin the shank, the tip wobbles eccentrically because the spring is the only thing keeping it loosely concentric with the shank's centerline.

You bring the spinning tip toward the workpiece. The tip is still wobbling. The instant the tip touches the part, friction grabs the tip and snaps it into apparent concentricity. A fraction of a thousandth later, the spring force is overcome and the tip kicks violently to one side. That kick is the signal.

When the kick happens, the geometry is exact. The spindle centerline is offset from the part edge by exactly one half-tip-diameter. A standard Starrett 827B has a tip diameter of 0.200 inches. Half of that is 0.100 inches. You back off, zero the axis, and move 0.100 inches toward the part. You are on the edge. You learned the math the first week. You never looked it up again.

Most edge finders have a 0.200 inch tip. Some have 0.500 inch tips for heavy roughing setups where chip pile-up might trick a smaller tip. The half-diameter offset is the only thing that changes. Recommended spindle speed is 800 to 1,000 RPM. Go faster and the centrifugal force overcomes the spring and the tip flies eccentric without ever snapping concentric. Go slower and the kick is mushy.

Before the Edge Finder, Machinists Used Cigarette Paper

They actually did this. The problem of locating a part edge against a spindle is as old as the milling machine. Before the edge finder existed, the standard methods were slow and depended on feel even more than the edge finder does.

The cigarette paper method was the most common. A machinist would chuck a cutter, spin it slowly, and bring it toward the workpiece with a piece of cigarette paper held between the cutter and the part. When the paper started to drag, the cutter was within the paper's thickness of the edge, typically 0.0015 to 0.002 inches. The machinist would zero the dial, then add the paper thickness plus half the cutter diameter to get to the actual edge. It worked. It was slow. It depended on the paper being consistent thickness and the cutter being on-center.

Feeler gauges replaced the cigarette paper for tighter work. A 0.001 inch feeler between the cutter and the part gave a more repeatable feel for the contact. Same math, smaller error budget, but still a setup that took a minute or two per edge.

For rough setups, machinists indicated the side of the part with a test indicator mounted in the spindle. They scribed layout lines on the workpiece, then moved the table until the spindle centerline aligned with the scribe. Accuracy depended on the scribe lines, the operator's eye, and the patience of the day.

None of these methods were fast. A skilled machinist could pick up an X edge with cigarette paper in about a minute. The edge finder cut that to under fifteen seconds.

The Patent That Made It Standard

The earliest patent specifically titled "Edge Finder" was filed by James G. Radtke on January 26, 1945, and granted on December 5, 1950, as US Patent 2,533,198. The Radtke patent described the spring-loaded eccentric tip mechanism that snaps concentric on contact, the same mechanism every wobble edge finder still uses today. Before that filing, the citation chain traces back to a 1909 "Centering Device" patent by Charles Thoma Jr. (US Patent 1,040,715), the conceptual ancestor that later inventors built on.

Radtke did not file alone. Through the late 1940s and 1950s, a wave of related patents was granted to other inventors refining the same idea: Raymond N. Baugh's center finder in 1953, Leopold Freimark's edge-locating device in 1954, Kenneth J. Johnson's combined edge-and-center finder in 1958. The mechanism was in the air. Multiple shops were working on the same problem at the same time. The wobble-type design that won was the one with the simplest geometry and the lowest cost to manufacture.

The $15 Number Is Why It Won

By the 1950s, the wobble-type edge finder was standard equipment in American machine shops. L.S. Starrett began producing them in volume. Brown & Sharpe, SPI, and a dozen smaller tool companies followed. Cigarette paper was free, but a $15 edge finder cut setup time by 75%. The math justified itself in a single shift.

The design has barely changed in seventy years. A modern Starrett 827B looks almost identical to the ones from 1955. The internal spring, the tip geometry, the recommended spindle speed, all of it is the same. Some shops still own edge finders their grandfathers used. The tool does not wear out unless you drop it on concrete or run it at the wrong speed.

Where Electronic Probing Took Over

The wobble-type edge finder did not survive every battle. On modern CNC equipment with automated part location, it has been almost entirely replaced by touch-trigger probing. A Renishaw OMP60 spindle probe locates an edge in about three seconds without any operator interaction. It writes the result directly to a work coordinate offset. Repeatability is around 1 micron, or 0.00004 inches. The edge finder is good to about half a thousandth in the best hands. That is roughly twelve times the error.

For tombstone work, multi-part fixtures, and any setup where the machine locates the part automatically, the probe won. The math is not close. A shop running a probed setup can change parts in under a minute. The same shop running an edge finder setup takes ten.

Most production CNCs ordered today come probe-ready. The probe is in the tool changer. The macros are written. The machinist hits a button and the machine sets its own work offsets.

Why It Is Still in Every Drawer

Probing replaced the edge finder on the production CNC. It did not replace it everywhere.

On a manual Bridgeport, on a knee mill, on a toolroom lathe set up with a cross-slide vise, the edge finder still wins. It costs $15. It needs no batteries, no calibration, no software, no DRO integration. It works on the first day a new machinist owns it. The feedback is physical. You feel the kick before you see it. The tool teaches the operator something a probe never can: how a spindle relates to a part.

The edge finder is still in every CNC machinist's drawer too, for the same reason a journeyman keeps a pair of calipers in his apron even when the shop has a CMM. When the probe stylus breaks on a Saturday morning and the next freight from Renishaw is Monday, the edge finder picks up an X-zero in under a minute. When a programmer wants a sanity check on a work offset, the edge finder confirms the probe got it right. The two tools are not enemies. The cheap one teaches and the expensive one repeats.

What This Means for the Shop

A shop that has never let a new hire pick up an edge finder is a shop producing button-pushers. The operator can run a probed setup. They cannot run a manual mill. They cannot pick up an edge on a fixture the probe macro does not cover. They cannot sanity-check the probe when it gives a result that does not match the print.

The $15 brass tool is not just a backup. It is the cheapest training tool a shop owns. A machinist who has felt the kick understands what the probe is doing at a level a button-pusher never will. The shops that still teach on an edge finder are the shops where every hand on the floor can take over when the automation hiccups. The shops that skipped it have a much narrower bench.

That is the real cost of letting the edge finder out of the toolbox. Not fifteen dollars. The depth of the team.

The Tool That Teaches

The edge finder is a $15 tool that has taught more machinists how a spindle relates to a part than any probe, any sensor, or any software ever written. It is mechanical. It is honest. It kicks when it should and stops when it should. It does not need a firmware update.

The first time a new machinist picks one up, they learn a piece of the trade they will use for the rest of their career. The first time they feel the kick, they understand something about geometry that no probe documentation ever teaches.

Every machinist on your floor learned the trade on a $15 brass tool. That is not an accident. That is the tool doing what nothing else in the shop can.

Frequently Asked Questions

How does a wobble-type edge finder work?

A wobble edge finder has a free-floating tip held to the shank by an internal compression spring. The tip is mounted off-center, so when the shank spins, the tip wobbles. When the wobbling tip touches the part, friction snaps it concentric for an instant, then the spring force is overcome and the tip kicks sideways. At the moment of the kick, the spindle centerline is offset from the part edge by exactly one half-tip-diameter.

What RPM should I run a wobble edge finder at?

800 to 1,000 RPM is the standard range, per Starrett's own documentation and Machinery's Handbook. Below 800 the kick feels mushy and the snap is harder to read. Above 1,000 the centrifugal force overcomes the internal spring and the tip flies eccentric without ever snapping concentric, so you never get a clean kick to zero from.

What is the half-diameter offset on a Starrett 827B?

The Starrett 827B has a 0.200 inch tip diameter. Half of that is 0.100 inches. After the tip kicks at the edge, you back off, zero the axis, and move 0.100 inches toward the part to put the spindle centerline on the edge. Some edge finders use a 0.500 inch tip for heavier roughing work, which gives a 0.250 inch offset. The half-diameter math is the only thing that changes.

Is a wobble edge finder accurate enough for production work?

In the best hands, a wobble edge finder is good to about half a thousandth (0.0005 inches). For most manual mill and toolroom work, that is well within tolerance. For CNC production work with tight tolerances, a Renishaw OMP60 or similar spindle probe is more precise (around 1 micron, or 0.00004 inches) and writes the result directly to a work coordinate offset. The edge finder is still useful on CNC as a backup or sanity check when the probe is down or unavailable.

Who invented the wobble-type edge finder?

The earliest patent specifically titled "Edge Finder" was filed by James G. Radtke on January 26, 1945, and granted on December 5, 1950, as US Patent 2,533,198. The Radtke patent described the spring-loaded eccentric tip mechanism every wobble edge finder still uses today. The conceptual ancestor traces back to a 1909 "Centering Device" patent by Charles Thoma Jr. (US Patent 1,040,715). Through the late 1940s and 1950s, several other inventors refined the same idea, including Raymond N. Baugh, Leopold Freimark, and Kenneth J. Johnson.

Why do shops still use edge finders when CNC probes exist?

Cost, simplicity, training, and resilience. An edge finder costs $15 and works on the first day a new machinist owns it. A probe stylus costs hundreds of dollars and requires macros, calibration, and integration. When the probe breaks on a Saturday morning, the edge finder picks up an X-zero in under a minute. More importantly, the edge finder is the cheapest training tool a shop owns: a machinist who has felt the kick understands what the probe is actually doing at a level a button-pusher never will.