Bullion Weighing Scales: Readability, Accuracy and Calibration

A bullion scale can show several decimal places and still give a result that is unsuitable for the decision you need to make. The display tells you how finely the instrument reports a reading. It does not, by itself, establish how close that reading is to the bar’s actual mass.

For buyers checking deliveries and businesses choosing equipment, the useful question is not simply “How many decimals does it show?” It is “Can this instrument support this measurement, at this load, under these conditions?”

Readability is the display step

Readability is the smallest increment shown on a digital balance. A scale with 0.01 g readability displays steps of one hundredth of a gram. That does not mean every result is correct within 0.01 g.

Accuracy concerns agreement with the quantity being measured. Precision concerns the agreement among repeated measurements under specified conditions. A scale can repeatedly show very similar numbers while retaining a systematic error. The international metrology vocabulary explicitly distinguishes these concepts. BIPM: measurement accuracy.

As a hypothetical example, a reference weight with an assigned value of 100.000 g might produce readings clustered around 100.040 g. The cluster could demonstrate good repeatability while the comparison reveals an offset. Adding another decimal place would not remove that offset. These numbers illustrate the distinction; they are not test results for a particular scale.

Capacity is only one end of the range

Maximum capacity tells you the largest permitted load, not the smallest quantity the instrument can measure reliably for your purpose. A balance may display a small object while its measurement uncertainty is too large relative to that object’s mass.

For a purely arithmetic illustration, an uncertainty of 0.02 g represents 2% of 1 g but 0.02% of 100 g. The same absolute uncertainty matters much more for the smaller quantity. This is not a claim that any particular bullion scale has that uncertainty.

A practical minimum weighing amount depends on instrument performance, environment and the required tolerance. It cannot be established from the final display digit alone. METTLER TOLEDO: analytical balance uncertainty and minimum weight.

Calibration, adjustment and checking do different jobs

Calibration establishes the relationship between an instrument’s indications and reference values, with the associated uncertainties. Adjustment changes the instrument so that it provides the intended indications. A routine check compares performance against an acceptance criterion between calibrations.

These activities can belong to the same maintenance programme, but they are not interchangeable. A button labelled “CAL” does not tell you which activity occurred or what was documented. The BIPM specifically warns against confusing calibration with adjustment. BIPM: calibration.

When reviewing a calibration certificate, ask:

• Does the instrument identification match the scale being used?
• Which loads and test conditions were covered?
• What errors and uncertainties were reported?
• Were results recorded before or after adjustment?
• What criteria determine whether performance is acceptable for this task?

A certificate should answer a measurement question, not merely provide a reassuring date or logo.

The weighing location matters

Air movement, vibration and temperature instability can interfere with weighing. A busy counter beside an air-conditioning outlet is a different setting from a stable bench. Off-centre placement can also affect indications, which is why balance assessments include eccentric-loading tests.

Follow the instrument’s instructions for installation, levelling and warm-up. Keep the pan clean, position the load consistently and allow the indication to settle. Handle reference weights carefully and protect them from contamination and damage. NPL’s mass-measurement guide addresses these environmental and handling issues, together with repeatability and scale-error assessment. NPL: practical mass measurement, April 2026 edition.

For a receiving procedure, record the instrument identity, unit, item reference and observed result. Note any packaging present. If a repeat reading changes unexpectedly, investigate before using it to accept or reject a delivery.

Commercial suitability needs a separate check

An instrument intended for stock monitoring is not automatically suitable for a transaction determined by weight. Ask the relevant local weights-and-measures authority or qualified supplier what approval, verification and operating requirements apply to your location and use.

Instrument markings may distinguish the actual display interval, often written as d, from the verification interval, e. These can differ. NIST’s explanation shows why a fine display digit should not be treated as a universal commercial tolerance. Its discussion concerns particular metrology frameworks, not blanket permission to use a device in another country. NIST: scale division versus verification scale division.

What a scale cannot settle

A mass result does not establish metal composition, authenticity or the contents of sealed packaging. Nor does a small disagreement between two scales, without their performance information, identify which result is better.

For the separate question of composition testing, see our guide to XRF testing and its limits for bullion buyers.

The strongest weighing record combines suitable equipment, documented performance and a consistent method. Choose the scale around the smallest and largest loads you actually handle, the tolerance your decision requires and the evidence needed when a result is questioned. More display digits are useful only when the measurement behind them is dependable.