I — What it is
The caliper reads finer than it measures
A graduated beam with a fixed jaw and a sliding jaw, read by a vernier, by a rack-and-pinion dial, or by an encoder and an LCD. It is the default modern hand tool and the only one of the three kinds that can emit data at all.
Vernier arithmetic is worth carrying: with the main scale graduated at s and n vernier divisions spanning (n−1)·s, resolution is s/n. A metric scale at 1 mm with 20 divisions gives 0.05 mm; with 50 divisions, 0.02 mm.
II — Resolution against specification
This is the fact to take away from the page. Under the caliper standard, a 0.01 mm-resolution instrument carries a default maximum permissible error around ±20 µm at 150 mm for external-face measurement.
There is a second error the standard treats separately and most people never hear about: a shift error applying on the move from the external faces to the internal, depth or step faces. It is larger than the external-face figure. That is the formal reason internal caliper measurements are worse than external ones, and depth and step faces each get their own test.
III — Why it is beaten by a micrometer
Not build quality. Geometry. A caliper is structurally Abbe-violating: the scale lies in the beam and the measurement happens out at the jaw tips, so any tilt in the slider is multiplied by the distance between them.
The canonical worked example: if slider tilt produces a jaw-slope error of 0.01 mm over a 50 mm jaw-guidance length, then at 40 mm jaw depth the tip error is (40/50) × 0.01 = 0.008 mm, comparable to the entire maximum permissible error of the instrument, produced by nothing but where on the jaw the part was held.
IV — Variants worth knowing
| Type | What it is for |
|---|---|
| Vernier | No battery, robust, slow to read. Still the right answer in some environments. |
| Dial | Rack-and-pinion driving a dial. Fast to read, and sensitive to grit in the rack. |
| Digital | Encoder and LCD. The default, and the only kind that can output data. |
| CM-type | Combined jaws with a nib and a second scale offset by the jaw thickness, so internal readings are direct — it removes the “add the jaw thickness” arithmetic error rather than asking the operator to remember it. |
| Point-jaw | Conical tips for uneven surfaces, webs and root diameters. |
| Blade-jaw | Thin flat blades for narrow grooves and recesses. |
| Groove / neck | Radiused or hooked noses, for internal and external grooves at depth. |
| Depth | Beam with a depth blade and a bridge. Inherits the caliper’s Abbe offset. |
| Gear-tooth | Depth slide sets chordal height, width slide reads chordal thickness. |
| Jenny / odd-leg | Not a measuring tool. A layout tool that scribes parallel to an edge or finds a shaft center. |
V — Technique that moves the numbers
- Inspect the faces for burrs and damage. A burr biases every subsequent reading rather than scattering them, which means it does not look like noise.
- Clean the beam, slider and faces, and the part. Dirt between the faces is the commonest zero shift there is.
- Let the part and the instrument equalize thermally. The error follows the difference between them rather than the room temperature.
- Adjust slider friction until motion is smooth with no play. play is Abbe tilt.
- Zero with the jaws closed and check it, against the light or by withdrawing a sheet of paper. That confirms face contact, which pressing the zero button does not.
- Hold the part with one jaw stationary, close to the beam. The Abbe instruction, again, because it is worth more than the rest combined.
- Apply light, consistent force at the thumb roller, never by pushing the jaw. No constant-force device exists on a caliper; consistency is the only available control.
- Rock gently for the external minimum or internal maximum, which finds the true diameter and removes chord error.
VI — Getting data out of it
A digital caliper with SPC output emits on request over a Digimatic frame: a value, a sign, a decimal position and a unit. The frame carries no serial number, timestamp or characteristic. Anything a record needs beyond the number is supplied by the layer above the tool.
Three practical consequences. The cable is model-specific: there is no universal Digimatic cable, and cable sprawl is one of the largest hidden costs in a first deployment. A wireless transmitter clipped to the port converts the same frame to radio and inherits the same silence about identity. And a keyboard wedge will type the value into whatever field has focus, which removes transcription error entirely and does nothing about the value landing in the wrong row.
VII — Sources and boundaries
- Maximum permissible errors for calipers are defined in ISO 13385-1 and the corresponding ASME standard. The single figure quoted here (the ±20 µm class at 150 mm for a 0.01 mm instrument) is cross-checked against a manufacturer’s published calibration note. The full tables belong to the standards and are not reproduced.
- The Abbe worked example is a manufacturer’s own, and the technique sequence follows published national-laboratory guidance.
- The Digimatic frame content is verified against an independent open-source decoder; see the protocol page for the byte-level detail and its sources.
- Instrument specifications vary by manufacturer, range and class. Treat every figure here as the shape of the answer and the instrument’s own calibration certificate as the answer.