What ISO 8655 Actually Requires of a Pipette Calibration

ISO 8655 is cited constantly and read rarely. It is the standard that defines what a piston-operated pipette must do and how you prove it. Here is what it actually asks for, in plain terms.

The parts that matter

ISO 8655-1 sets terminology, general requirements, and the maximum permissible errors — the accuracy and precision limits an instrument must meet at a given nominal volume. ISO 8655-2 covers piston-operated pipettes specifically: the test method, the conditions, the number of measurements. ISO 8655-6 defines the gravimetric reference method that both rest on.

Gravimetry is the reference method

You dispense water onto an analytical balance and convert mass to volume. That conversion is not a constant — it depends on the density of water at the measured temperature, on air density, and therefore on barometric pressure and relative humidity. The correction factor that ties them together is the Z-factor.

This is why environmental capture is not optional bookkeeping. A calibration performed without recording temperature and pressure has applied the wrong Z-factor, and the reported volume is wrong by an amount nobody measured.

Measurement counts and volume points

The standard specifies ten measurements at each test volume for a full type-evaluation-grade calibration, at nominal, 50%, and the minimum specified volume. Routine in-service calibration commonly uses a reduced but documented scheme — what matters is that the scheme is defined, justified, and reported honestly.

• Low volume is where relative error is largest and where instruments fail first. Never skip it.
• Mid volume catches linearity problems a two-point test can miss entirely.
• High volume is where seal leakage and piston wear show up most clearly.

A two-point calibration is a legitimate reduced protocol. A two-point calibration reported as if it were a three-point one is a data integrity problem.

Conditions the standard cares about

• Water: distilled or deionized, Grade 3 or better, equilibrated to room temperature.
• Temperature: stable between 15 °C and 30 °C, varying no more than ±0.5 °C during the test, with instrument, water, and tips all equilibrated.
• Balance: resolution appropriate to the test volume — sub-microgram readability for the smallest volumes.
• Evaporation control: an evaporation trap or a humidity-controlled enclosure. At 1 µL, uncontrolled evaporation is a first-order error, not a rounding detail.
• Technique: consistent pre-wetting, immersion depth, dwell time, and dispense angle across the whole measurement series.

Accuracy and precision are different failures

Systematic error — accuracy — is the difference between mean delivered volume and target. It is often adjustable. Random error — precision, reported as a coefficient of variation — is the scatter across the measurement series, and it usually points at a mechanical fault or a technique problem rather than a setting.

An instrument that is off by 2% but tight is a different problem from one that averages perfectly but scatters by 3%. Only the first can be fixed with an adjustment. A calibration report that gives you one number and not both has hidden the more important half.

What to ask your calibration provider

Ask which volumes were tested, how many measurements at each, what water grade was used, whether temperature and pressure were recorded at the point of test, and whether as-found data was captured before adjustment. Those five answers tell you almost everything about whether a calibration is meaningful — and any provider doing the work properly will answer them without hesitating.