Torque Calibration in Singapore: Torque Wrenches, Screwdrivers, and Transducers

Reading a torque calibration certificate

A torque calibration certificate issued to ISO 6789-2 carries specific data worth understanding rather than skimming for the pass/fail line. At each of the five test points, the certificate records the individual measurements (not just their average), the mean applied torque, the deviation from nominal expressed as a percentage, and the tool's repeatability, the spread between the highest and lowest of the repeat measurements at that point, itself an indicator of mechanism wear separate from the mean deviation. A wrench can show a mean deviation comfortably within the ±4% MPE while also showing wide repeatability, meaning any single use of the tool could land anywhere across that spread; this is a meaningfully different (and often more concerning) finding than a tool with tight repeatability but a small, consistent offset, which a user could in principle learn to compensate for. The certificate should also state the expanded measurement uncertainty of the calibration itself and the direction(s) tested (clockwise, counterclockwise, or both for bidirectional tools), details that a bare pass/fail summary omits entirely.

Which torque tools and instruments require calibration?

Any torque tool whose output is used to control fastener tension in a quality-managed process requires calibration. The industries where this matters most in Singapore include aerospace MRO (Seletar, Changi), oil and gas (Jurong Island), precision manufacturing (electronics, semiconductors), and pharmaceutical manufacturing (equipment assembly, maintenance). The tools requiring regular calibration include:

• Click-type torque wrenches (Type II, ISO 6789). The most common industrial torque tool. Set a target torque; the mechanism "clicks" when reached.
• Beam torque wrenches (Type I). A deflecting beam displays applied torque on a scale. Simple but requires reading skill.
• Dial torque wrenches (Type I). A dial indicator shows applied torque continuously. Higher resolution than beam type.
• Electronic torque wrenches. Digital display, programmable torque targets, data logging capability. Common in aerospace and precision assembly.
• Electronic torque screwdrivers, for low-torque precision assembly (PCBs, connectors, medical devices).
• Torque transducers / sensors. Measurement instruments used in test and calibration rigs. Require calibration as measuring equipment.
• Powered torque tools (pneumatic/electric). Nutrunners and impact wrenches used in high-volume assembly lines.

How torque wrench mechanisms work, and why it affects calibration

The internal mechanism of a torque wrench determines how it should be calibrated and what failure modes calibration is actually looking for.

Click-type (Type II) mechanisms

A calibrated spring, compressed by adjusting the wrench's handle to a target torque scale, releases a cam mechanism at the set torque, producing an audible and tactile "click." The spring's characteristics are what calibration verifies: a spring that has taken a permanent set from being stored under tension, or from repeated over-torque events, will click at a different torque than the scale indicates, even though the wrench "feels" and sounds exactly the same as it always has. This is precisely why click wrenches are the type most often found out of tolerance at calibration despite showing no visible external damage.

Beam and dial (Type I) mechanisms

A beam wrench displays applied torque through the mechanical deflection of a calibrated beam, read against a scale; because there is no spring to fatigue or click mechanism to wear, beam wrenches tend to hold calibration longer, though the scale and pointer alignment can still shift with rough handling. A dial wrench uses an internal deflecting element connected to a geared dial indicator, offering finer resolution than a beam type but introducing gear-train wear as an additional potential source of drift.

Electronic torque wrenches and screwdrivers

These use a strain-gauge or similar electronic sensor to measure the actual deflection or strain in the tool's driving element, converting it to a digital torque reading, often with data logging and programmable target alarms. Because the underlying sensor is electronic rather than a mechanical spring, drift characteristics differ from click-type tools, but the strain-gauge element itself, along with the tool's internal amplifier and display electronics, is still subject to calibration drift and needs the same periodic verification.

ISO 6789-2: the calibration standard for torque wrenches

ISO 6789-2:2017 is the key standard for torque wrench calibration. It defines five calibration test points across the tool's usable range, the number of test measurements at each point, the method for calculating the mean deviation, and the maximum permissible error (MPE). For most torque wrenches, the MPE is ±4% of the applied torque at each test point. An important detail: calibration under ISO 6789-2 requires a minimum of five measurements at each test point (with the tool re-set to the nominal torque between each measurement for Type II tools), not a single measurement. The mean of the five measurements is used to calculate the deviation.

Torque multipliers and powered tools deserve their own check

Torque multipliers, gear-reduction attachments used to generate very high torques (common on large flange and structural bolting) beyond what a hand wrench alone can deliver, introduce their own mechanical losses through gear friction, meaning the input torque and the actual output torque delivered to the fastener are not simply related by the multiplier's marked ratio. A multiplier's efficiency should be verified as part of its calibration, not assumed from the nameplate ratio, because gear wear reduces efficiency over time and use. Pneumatic and electric nutrunners used on high-volume assembly lines present a different challenge: their output torque is influenced by air pressure or supply voltage, drive socket condition, and joint characteristics (a "hard" joint reaching target torque quickly behaves differently from a "soft" joint that compresses gradually), so calibration of powered tools typically needs to be performed under conditions, or against a joint simulator, that reasonably represents the actual production joint, not just a bench-mounted static load cell in isolation.

When to recalibrate a torque wrench outside the normal interval

Beyond the standard 12-month or 5,000-cycle interval, there are specific events that require immediate recalibration:

• The tool was dropped. Even a drop from bench height can shift the calibration of a click-type mechanism. Do not use a dropped torque wrench without recalibration.
• The tool was over-torqued. Applying torque beyond the tool's maximum rated torque can permanently damage the click mechanism or flex the beam beyond its elastic range.
• Improper storage. Click-type torque wrenches should be stored at minimum scale setting (not at a set torque value) to relieve the spring. Storing at a set torque for extended periods fatigues the spring and affects calibration.
• The tool shows unusual behaviour. A click that occurs inconsistently or at unexpected torque levels suggests mechanism wear or damage.
• Before a critical assembly, for safety-critical joints (pressure vessel flanges, structural bolts, aircraft fasteners), calibrate or verify the torque tool before use, not just on schedule.

Torque calibration for aerospace MRO in Singapore

Singapore's aerospace MRO sector (ST Engineering, SIA Engineering, HAECO, and others) operates under AS9100 quality management systems, which require calibrated and traceable torque tools for fastener applications. SAC-SINGLAS accredited calibration is required for torque tools in AS9100 environments. A non-accredited calibration certificate is not acceptable for aerospace customer audits. Unitest provides SAC-SINGLAS accredited torque calibration with full ISO 6789-2 compliance and NMC-traceable certificates suitable for AS9100 records.

How torque calibration is actually performed

A torque calibration bench applies a controlled, measured torque to the wrench under test through a calibrated reference transducer, typically a precision strain-gauge load cell mounted in-line with a lever arm of known, calibrated length, or a rotary torque transducer directly measuring the applied twisting force. For a Type II (setting) wrench such as a click-type, the wrench is set to the target torque, then loaded smoothly until it clicks or the mechanism releases, and the reference transducer's reading at the point of release is recorded; this is repeated (a minimum of five times per ISO 6789-2) with the wrench reset to the nominal setting between each measurement, since resetting captures the tool's real repeatability rather than a single lucky reading. For a Type I (indicating) wrench such as a beam or dial type, torque is applied and increased steadily while both the reference transducer and the wrench's own scale are read simultaneously at each of the five test points, again with repeat measurements. The test points themselves are spread across the tool's usable range as defined by ISO 6789-2, commonly including points near 20%, 60%, and 100% of maximum scale, because a wrench's percentage error is rarely uniform across its full range.

Why the 20-100% usable range rule exists

ISO 6789 defines a torque wrench's usable range as roughly 20% to 100% of its maximum scale value, and using a wrench below that 20% floor is a common but consequential mistake. Near the bottom of a wrench's mechanical range, the ratio of the mechanism's inherent friction and backlash to the actual torque being measured grows large, meaning the tool's percentage error at low settings is disproportionately worse than at mid-range or high settings, even though the wrench passed calibration at its tested points. A 200 Nm click wrench set to apply 20 Nm, for example, is being asked to operate right at or below the edge of its usable range, and a calibration certificate showing the tool comfortably within tolerance at 40 Nm, 120 Nm, and 200 Nm says nothing reliable about its accuracy at 20 Nm. The practical fix is straightforward: select a torque tool whose full-scale range comfortably brackets the target torque, rather than reaching for the largest wrench in the toolbox and dialling it down to a fraction of its capacity.

What over-torque and under-torque actually cost

The consequence of an uncalibrated torque tool is not abstract. A fastener joint is engineered to a specific clamping force, delivered through a target torque that accounts for thread friction, washer friction, and the joint's material properties. Under-torquing leaves a joint with insufficient clamping force, which under vibration or cyclic loading can loosen progressively, a mechanism directly implicated in fastener loosening failures across rotating machinery, structural connections, and piping flanges. Over-torquing does the opposite kind of damage: it can stretch a bolt beyond its elastic limit into permanent (plastic) deformation, reducing its remaining clamping capacity and fatigue life even if the joint does not fail immediately, or it can crush a gasket beyond its designed compression, creating a leak path in a pressure joint that looked, at the moment of assembly, like it had been tightened "extra safe." Both failure modes trace back to the same root cause: a torque tool whose indicated value does not match the torque it is actually delivering, which is exactly what calibration exists to catch before the joint is assembled, not after it fails in service.

Industry context: where torque calibration carries the highest stakes in Singapore

Beyond aerospace MRO, torque-controlled fastening is a safety-critical discipline across several of Singapore's industrial sectors. On Jurong Island, flange bolting on pressurised piping and vessels is typically governed by structured bolting procedures aligned to standards such as ASME PCC-1, where torque tool calibration and the bolt-up sequence are both audited elements of a joint integrity programme, an under- or over-torqued flange bolt is a direct leak or blow-out risk. In precision electronics and semiconductor assembly, torque screwdrivers control fastening of small components where over-torque can crack a PCB or strip a miniature thread, damage that may not be visible until the product fails much later in the field. In structural steel and marine construction, torque or tension-control methods on high-strength structural bolts (aligned to standards referencing ASTM or equivalent bolting specifications) directly affect a connection's designed load capacity. In each of these settings, a torque calibration certificate is not a quality-paperwork formality, it is the evidence that the fastening process actually delivered the clamping force the engineering design assumed it would.

What to prepare before sending torque tools for calibration

• State the actual torque values used in service, not just the tool's full-scale range, so calibration points can be concentrated where the tool is actually operated.
• Declare the direction(s) of use. Many click wrenches are bidirectional (clockwise and counterclockwise); if your application uses both directions, both should be calibrated, not just the default direction.
• Flag any drop, over-torque, or unusual-click history, exactly the events that justify sending the tool in ahead of its scheduled date rather than waiting.
• Confirm whether AS9100, ASME, or another specific standard governs the application, so the lab issues a certificate referencing the correct acceptance criteria and test point structure for that context.

Common mistakes in torque tool calibration programmes

• Storing click-type wrenches set at a torque value rather than at their minimum scale setting, which fatigues the internal spring over time and is one of the most preventable causes of premature drift.
• Using a tool below the 20% usable range floor, treating a calibration pass at higher settings as proof of accuracy at a much lower target torque it was never really validated for.
• Continuing to use a dropped or over-torqued tool because "it still clicks normally," when a mechanical shift in the spring or beam is not something a user can feel or hear.
• Not distinguishing torque tools from torque transducers when scoping calibration, a bench-mounted reference transducer used to calibrate other tools needs its own periodic calibration against a higher-level standard, and treating it as a permanently "known good" reference is a traceability gap.
• Skipping bidirectional verification on wrenches used in both directions, when a mechanism's spring and cam geometry can behave slightly differently clockwise versus counterclockwise.