Gas & Process Monitoring for Additive Manufacturing

Why Calibration Is Non-Negotiable in Additive Manufacturing

Additive manufacturing, whether selective laser melting (SLM), direct metal laser sintering (DMLS), binder jetting, or fused deposition modelling (FDM). Is fundamentally a precision thermal and atmospheric process. Unlike machining, where errors can often be corrected in a subsequent pass, AM builds layer upon layer. A process condition that drifts out of tolerance in the first twenty layers contaminates every layer above it. By the time a part comes out of the machine, the damage is invisible and the only way to detect it is with destructive testing or expensive CT scanning.

Calibrated instruments are the only reliable way to prove that your build environment was within specification for the full duration of a print run. For facilities supplying aerospace structures, medical implants, oil-and-gas components, or defence parts, that proof is not optional. It is a contractual and regulatory requirement. Singapore manufacturers seeking to supply into these verticals must hold calibration certificates that demonstrate metrological traceability to recognised national standards, and SAC-SINGLAS accreditation under ISO/IEC 17025 is the locally recognised route to do exactly that.

Even facilities producing industrial tooling, jigs, or consumer products benefit from calibrated process monitoring. Consistent conditions mean consistent mechanical properties, consistent dimensional accuracy, and lower scrap rates. Outcomes that directly affect material costs and delivery schedules in a city where industrial floorspace and skilled technician time are expensive.

The Six Instrument Classes Every AM Facility Should Calibrate

1. Oxygen Analysers

Oxygen concentration inside the build chamber is the single most important process parameter in metal AM. Titanium, aluminium, and reactive alloys oxidise rapidly in the presence of even trace oxygen, forming oxide inclusions that create brittle zones and cause fatigue failures in service. Most SLM and DMLS machine manufacturers specify oxygen levels below 500 ppm (0.05%) as a build start condition, and aerospace-grade specifications may demand sub-100 ppm. The machine's built-in oxygen sensor is almost always the only barrier between a good part and a scrapped one.

Electrochemical oxygen sensors degrade over time. The sensing cell has a finite life, and readings drift as the electrolyte is consumed. Zirconia sensors used in high-temperature applications also drift with thermal cycling. Without periodic calibration against a certified reference gas of known concentration, you cannot know whether the displayed reading of 200 ppm is accurate or whether the true value is 800 ppm. Calibration against a NIST- or NMC-traceable reference gas, performed by a SAC-SINGLAS accredited laboratory, provides a documented correction factor and measurement uncertainty that your QMS can reference.

2. Dew-Point and Humidity Transmitters

Metal powders (particularly titanium, aluminium, stainless steel, and Inconel), are hygroscopic. Storage rooms, glove boxes, and powder-handling stations require monitored and controlled humidity, typically below 40% RH (and often below 25% RH or a specific dew point for reactive alloys). Polymer filaments used in FDM are equally sensitive: nylon absorbs enough moisture overnight in Singapore's ambient humidity to cause significant surface blistering and strength reduction.

Capacitive humidity transmitters and chilled-mirror hygrometers are the instruments of choice. Both drift over time. The capacitive sensor's polymer sensing element ages, while the chilled-mirror optics accumulate contamination. Annual calibration against a traceable humidity reference is standard practice, with more frequent calibration for instruments in continuous service or environments where contamination is likely.

3. Build-Chamber Temperature Sensors and Substrate-Preheat Thermocouples

Many metal AM machines preheat the build platform to 100–200°C (and some high-temperature systems operate above 500°C) to reduce thermal gradients, residual stress, and cracking. Particularly important for tool steels and some nickel superalloys. The thermocouples and RTDs that control these preheat stages must be calibrated to confirm their readings are accurate across the intended operating range.

Type K thermocouples are standard in most metal AM machines, but they are susceptible to drift through oxidation of the sensing wires at elevated temperatures. Type N thermocouples offer better stability at high temperatures. Either way, calibration every 6–12 months (with calibration intervals shortened for sensors exposed to repeated thermal cycling above 300°C), is the industry norm. The calibration certificate should state the expanded uncertainty at a coverage factor of k=2, which is what quality auditors look for.

4. Pressure Gauges on Inert-Gas Supply Lines

Argon and nitrogen are the inert gases most commonly used to purge and maintain AM build chambers. Both are costly, and their supply lines operate at pressures ranging from 1 bar (purge delivery) to 200+ bar (cylinder storage). Pressure gauges on these lines serve two functions: they tell operators whether the chamber is properly pressurised for a leak-free build, and they trigger alarms if supply pressure falls below safe operating levels mid-build.

A pressure gauge that reads 2 bar when the true pressure is 1.5 bar will mask a developing leak. The build may complete, but with elevated oxygen ingress that compromises part integrity. Bourdon-tube gauges are calibrated against a traceable dead-weight tester or digital reference standard. Annual calibration is standard for gauges in continuous service; gauges on cylinders changed frequently may warrant more frequent checks.

5. Gas-Flow Meters

Mass flow meters and controllers on argon/nitrogen purge systems govern the rate at which inert gas is introduced into the chamber and the rate at which it is recirculated through filters. Incorrect flow rates reduce the effectiveness of the gas curtain that protects the melt pool from spatter re-deposition, and they also affect the efficiency of the filtration system. Thermal mass flow meters and Coriolis meters used in these applications are calibrated against traceable volumetric or gravimetric flow standards.

6. Laser Power and Optical Emission Monitors

In-process monitoring systems (including melt-pool cameras, photodiodes, and pyrometers used for thermal emission monitoring), are becoming standard on high-end metal AM machines. These instruments feed data into quality-assurance algorithms that flag anomalies layer by layer. If the detector sensitivity drifts, the algorithm produces false passes or false alarms. Calibration of these optical and radiometric instruments is a specialist scope that should be confirmed against the laboratory's accredited calibration capabilities before booking.

Calibration Intervals: What Determines the Right Frequency?

There is no single regulatory body that mandates a universal calibration interval for AM instruments. The correct interval depends on instrument type, usage intensity, criticality of the parts being produced, and the quality management standard under which your facility operates. The guidance below reflects common industry practice and the recommendations we give Singapore clients across aerospace, medical, and advanced manufacturing sectors.

For oxygen analysers, a 6-month interval is appropriate for facilities running production shifts five or more days per week, or those supplying parts to aerospace primes under AS9100. A 12-month interval may be acceptable for R&D labs or lower-volume prototype shops where the analyser is powered down between builds. The key question to ask: what is the consequence of a false reading? If the answer involves scrapped titanium builds or a failed batch of implants, lean towards the shorter interval.

For temperature sensors, the thermal history matters. A Type K thermocouple that has been repeatedly cycled from room temperature to 300°C for substrate preheating will drift faster than one used only for chamber-temperature monitoring at modest temperatures. Our article on how often to calibrate instruments covers the general framework for setting intervals, including the risk-based approach recommended by ISO/IEC 17025 laboratories.

SAC-SINGLAS Accreditation and Compliance Evidence

When an auditor conducting an AS9100 or ISO 13485 surveillance audit asks for your calibration records, they are looking for two things: proof that the instrument was calibrated by a competent laboratory, and proof that the calibration result is traceable to recognised national or international measurement standards. SAC-SINGLAS accreditation under ISO/IEC 17025 provides both.

The SAC-SINGLAS certificate from Unitest Instruments (Acc. No. LA-2023-0845-C) carries the SAC mark, states the measurement uncertainty at k=2 for each calibration point, and documents the traceability chain back to Singapore's National Metrology Centre (NMC), which in turn participates in the BIPM Key Comparison framework connecting to SI units. This chain of traceability is exactly what ISO/IEC 17025 requires, and it is what separates an accredited certificate from an in-house check or a non-accredited service provider's report.

Our guide on accredited vs non-accredited calibration explains why the distinction matters and what specific risks arise from relying on non-accredited certificates in a regulated manufacturing environment. The short version: non-accredited calibration cannot be independently verified, may not include stated uncertainty, and can be challenged during audits. Creating compliance gaps that are expensive to remedy retroactively.

What a QA Team in Singapore Actually Needs to Do

For a facilities or QA manager running an AM facility in Singapore, the practical compliance checklist comes down to six actions. First, conduct an instrument inventory: list every sensor, analyser, gauge, and meter that influences build quality or safety, including instruments embedded in the AM machine itself. Machine manufacturers often exclude built-in sensors from their service contracts, leaving calibration to the facility.

Second, assign criticality and intervals. Not all instruments carry equal risk. Oxygen analysers on a titanium SLM machine are critical; a room-temperature thermometer used only for ambient monitoring is low-risk. Assign calibration intervals accordingly, and document your rationale in the QMS. Third, schedule calibration before each instrument's due date. Building in a 30-day lead time to allow for booking, collection, and any instrument turnaround time.

Fourth, review each calibration certificate when it is returned. Check that the as-found readings (before adjustment) were within your acceptance limits. If an instrument was found significantly out of specification, initiate a non-conformance investigation to assess whether any builds performed since the last calibration may have been affected. This is a requirement under AS9100 and ISO 13485 and is good engineering practice regardless of whether you are formally certified.

Fifth, store calibration records in a retrievable format linked to each instrument's asset record in your QMS. Auditors will ask to see the complete calibration history for critical instruments, not just the most recent certificate. Sixth, when an instrument is repaired or its sensing element is replaced, treat it as a new calibration event. Do not carry over the previous certificate.

Polymer AM: A Different but Overlapping Instrument Set

Facilities operating FDM, SLA, DLP, or PolyJet printers face a different but related calibration requirement. The critical instruments in polymer AM are nozzle and bed temperature controllers, UV-intensity meters or radiometers (for SLA/DLP curing), and ambient humidity sensors protecting moisture-sensitive filaments and resins. Dimensional accuracy also depends on calibrated calipers and CMMs used in post-print inspection.

Temperature calibration for FDM extruder heaters is often neglected because the machine's built-in PID controller maintains temperature, but if the thermocouple feeding the PID has drifted, the controller will faithfully maintain the wrong temperature. For engineering polymers such as PEEK, ULTEM, and polycarbonate, even a 10°C error in nozzle temperature affects layer adhesion and mechanical strength. Annual calibration of the thermocouple is prudent; checking against a calibrated reference thermometer at the nozzle tip can be done in-house as an interim verification between formal calibrations.

Detailed guidance on reading and interpreting calibration certificates (including what measurement uncertainty means for your process decisions), is covered in our article on how to read a calibration certificate. Understanding what the numbers on a certificate actually tell you is a practical skill for any QA team managing AM equipment.