Documentation / Instruments / Micrometers

I — What it is

A screw of known pitch, and three structural advantages

Rotation becomes axial translation through a precision screw. Metric standard pitch is 0.5 mm with a 50-part thimble, giving 0.01 mm per division; the sleeve carries a second row of graduations offset by half a millimeter to disambiguate the half-millimeter steps.

Reference · Shared error physics at Instruments

That offset second row is the classic gross misread. Miss it and the reading is wrong by exactly 0.5 mm: a clean half-millimeter step rather than scatter. It is large, it is systematic, and it looks like a real measurement of a different part.

II — Why it beats a caliper

By roughly an order of magnitude, and for three structural reasons rather than because it is better made.

Reason	What it does
Abbe compliance	For a straightforward external measurement the screw axis, the spindle axis and the line of measurement are one line. There is no offset for a tilt to act on, which is exactly what a caliper cannot say.
A constant-force device	The ratchet or friction thimble exists to make contact force repeatable. The word is repeatable rather than correct. It removes an operator variable rather than an error.
A 25 mm span	Each instrument covers one 25 mm step. Short spans mean less structure between the scale and the part, and less structure means less to deflect.

The third one is the trade. Covering 0–150 mm takes six micrometers where it takes one caliper, and each of the six needs its own setting standard and its own calibration. The accuracy is bought with inventory rather than with cleverness.

III — What the specification governs

The micrometer standard specifies more than a single accuracy number. It separates maximum permissible error for full surface contact from partial surface contact, and it also bounds repeatability and measuring force, which is the standard acknowledging that force is part of the instrument rather than part of the operator.

Two characteristics matter more in practice than the headline figure:

Characteristic	What good looks like, and how it is checked
Face flatness	0.6 µm or better on a working instrument. Checked against an optical flat by counting interference fringes, each worth about 0.32 µm — so the check is done by eye, in fringes, and needs no electronics at all.
Face parallelism	Checked with a staggered optical parallel set, so the spindle is examined at four different rotational phases. Parallelism error is a function of screw angle, and an instrument that is parallel at one thimble position and not another has a bent spindle or a worn thread.

On published accuracy grades. Grade tables circulate widely in secondary compilations, and the ones we could find were not traceable to the standards themselves. The shape is dependable (roughly ±2 µm at 25 mm for the best grade, loosening by about a micron per 50 mm of range) but for any specific instrument, use the manufacturer’s stated figure and the instrument’s own certificate rather than a table off the internet, including this one.

IV — Frame sag, and the rule that follows from it

A large micrometer frame deflects under its own weight and under the way it is held. One manufacturer publishes an attitude and support error of 0 to −40 µm across 325–1025 mm measuring lengths, depending on how the frame is supported.

Read that against the instrument’s own accuracy class and the consequence is immediate: the manner of holding can matter more than the purchase price. This is why large micrometers carry a defined support condition (supported at Airy or Bessel points, positions chosen to minimize deflection) and why the hard rule is:

zero it in the configuration of use

Held horizontally, zeroed vertically, measured horizontally. That sequence puts an error into every reading that no amount of care at the thimble will recover, and nothing in the data will show it: the numbers are simply offset, consistently, in one direction.

V — Variants worth knowing

Type	What it is for
Outside (standard)	Flat parallel faces. General external dimensions.
Blade	Thin blades and a non-rotating spindle, for narrow grooves, O-ring grooves and keyways. The non-rotating spindle is the point — a rotating one would twist the blade.
Ball anvil	One or both faces spherical, for tube wall thickness and curved surfaces.
Disc	Large flat discs for gear tooth span, paper, soft sheet and thin fins.
Screw thread	Vee anvil and conical spindle matched to the thread angle, reading pitch diameter directly.
V-anvil	60°, 108° or flute-matched geometry, for odd-fluted taps, reamers and end mills.
Point	Sharp conical points for thread roots, small grooves and web thickness.
Tube	Cylindrical anvil perpendicular to the spindle, for tube wall from the bore side.
Indicating	The anvil floats against a dial or electronic display, giving comparative measurement at controlled force.
Bench	Rigid frame, usually indicating and constant-force. Master-level external measurement.
Depth	A head driving interchangeable rods through a flat base. Compliant along the rod axis, sensitive to base rock.
Non-rotating spindle	The spindle translates without turning, which stops it scuffing soft parts.

VI — Technique that moves the numbers

Clean both faces by closing on a sheet of paper and pulling it through. Cleans and confirms contact in one movement.
Zero in the orientation of use, with the ratchet or friction device only. Above 25 mm, zero on the supplied setting standard.
Hold the setting standard with an insulated grip or tongs. Hand heat on a steel standard is a live error rather than a theoretical one.
Approach with the thimble; finish with the constant-force device, one to three clicks. Spinning the ratchet is a large contributor to appraiser variation, and it is the single most common bad habit on this tool.
Rock the frame to find the minimum. Finds the true diameter and removes chord and cosine error.
Mount it in a stand for repeat work. Removes hand heat from the frame and frees the operator’s hands for the part.
Lock before removing if the reading has to be taken away from the setup.

VII — Getting data out of it

A digital micrometer with SPC output emits the same Digimatic frame a caliper does: a value, a sign, a decimal position and a unit. Six BCD digits carry 0.001 mm comfortably. What it does not carry is which micrometer sent it, and with micrometers that omission bites harder than it does with calipers, because a 0–150 mm range is six instruments rather than one, and “which of the six” is precisely the question a recall analysis asks.

For a study, the balance is different from a caliper’s. The constant-force device removes one of the two big operator variables, so a Gage R&R on a micrometer usually shows lower appraiser variation, unless somebody is spinning the ratchet, in which case it shows up clearly as reproducibility and is worth finding, because it is a training fix rather than an instrument one.

VIII — Sources and boundaries

Metrological characteristics (full and partial surface contact treated separately, plus repeatability and measuring force) follow the micrometer standard, ISO 3611.
Face flatness and parallelism practice, the fringe value, the staggered optical parallel set and the attitude/support figure are drawn from manufacturer and national-laboratory published guidance.
Accuracy grade tables are deliberately not reproduced; see section III for why.
The Digimatic frame content is verified against an independent open-source decoder; the protocol page carries the byte-level detail and its sources.

Calipers → Shared error physics → Free Gage R&R calculator →