A high-temperature accelerometer build is governed by four independent limits: the piezoelectric or MEMS element's own thermal ceiling, the magnet assembly's irreversible demagnetisation point, the mounting method's mounted resonant frequency, and the certification scope named on the supplier's paperwork. Treat any one of these in isolation and the measurement chain fails at temperature even if every individual spec looks healthy on the data sheet [S1][S2].
The thermal envelope typically spans 150°C to 350°C for industrial turbines, downhole MWD tools, and aerospace engine bays, which is well beyond the working range of any NdFeB magnetic circuit. Above roughly 80°C, standard sintered NdFeB begins irreversible flux loss, and the sensor drift becomes non-recoverable on cool-down; this is the single most common root cause of a "calibrated at room temperature, wrong on the machine" failure [S2].
Sensor Element: Where the First Thermal Wall Lives
Piezoelectric charge-mode accelerometers generally outrun IEPE types on continuous high-temperature exposure because the IEPE integral electronics (FET input stage, gain stage) carry their own derating curve, and the built-in amplifier is usually the part that surrenders first, not the crystal stack. A charge-output device with a remote in-line converter mounted in a cooler zone is the conservative path for sustained service at or above 200°C [S1].
If any of those three numbers is missing, treat the device as unproven for the application; the supplier's high-temperature marketing copy does not substitute for a printed spec [S1].
Magnet Material: Sm2Co17 Above 150°C, Encapsulated
Samarium-cobalt 2:17 (Sm2Co17) is the default magnet chemistry for accelerometers rated for 150°C to 350°C service, with a reversible temperature coefficient of Br (alpha(Br)) of −0.030% to −0.035%/°C, roughly an order of magnitude better than NdFeB and low enough that the resulting gain drift is correctable in post-processing [S2].
Sm2Co17 is also brittle: fracture toughness sits at 2 to 3 MPa·m^0.5, so any high-g shock profile (>50 g peak) demands a non-magnetic stainless sleeve, a potted high-temperature epoxy backfill, or both. Sintered SmCo without encapsulation is a known cracking failure mode in turbine bearing-housing applications where rotor imbalance events excite structural resonances in the magnet body [S2].
AlNiCo and SmCo5 sit between NdFeB and Sm2Co17 on the temperature-stability curve, but neither replaces 2:17 in sustained service above 200°C; AlNiCo's reversible coefficient is favourable but its low coercivity makes it vulnerable to external demagnetising fields from adjacent motor windings, and SmCo5 (1:5) has lower energy product and worse thermal drift than 2:17 at the same geometry. Match the magnet chemistry on paper, then verify on a hot bench at the real operating temperature, not at 25°C [S2].
Mounting Method: Stud Beats Magnetic Base by 5 kHz
Stud mounting with a clean, flat, grease-coupled surface preserves the accelerometer's intrinsic mounted resonant frequency above 10 kHz, while a magnetic base typically drops that figure to 1.5 to 4.5 kHz depending on the contact interface stiffness and added mass [S2]. For high-frequency vibration analysis (gearbox meshing harmonics, bearing outer-race defect frequencies above 5 kHz), this single difference moves the sensor from "useful" to "missing the fault."
Wax and beeswax adhesives are convenient but only valid at or near room temperature; their shear strength collapses well below 100°C and the sensor can shift or decouple under vibration once the wax softens. For high-temperature service, use a stud with a thin film of high-temperature coupling grease (silicone or similar rated for the target temperature), a tapped and machined mounting pad with a surface finish of 1.6 µm Ra or better, and a torque value taken from the sensor data sheet rather than guessed [S1].
Cable routing also matters: route the cable away from heat sources, support it within 50 mm of the connector to limit flex, and avoid sharp bends smaller than the manufacturer's minimum bend radius. A stud-mounted sensor that pulls its cable taut against a hot casing is a sensor that drifts [S1].
Certification Scope: Read the Certificate, Not the Brochure
Common accreditations to check on the supplier paperwork: ATEX 2014/34/EU for explosive atmospheres, IECEx Scheme for international hazardous-area recognition, ISO 9001 for the quality system, and ISO 17025 for any in-house calibration lab issuing the test data. The certificate must name the exact model number and the exact revision of the standard it claims compliance with; a generic "ATEX-certified accelerometer" line on a brochure is not evidence [S1].
For hazardous-area service, the certificate will name a category (1, 2, or 3) and a gas group (IIA, IIB, IIC) or dust group (IIIA, IIIB, IIIC). Confirm the marked ambient range covers the installation's minimum and maximum, and that the certificate is in date. Many "ATEX" markings on imported sensors actually cover a different model variant or a different revision of the housing, which is the single most common audit finding during on-site compliance reviews.
For OEM-level quality audits, request a copy of the manufacturer's ISO 9001 certificate (issued by an accredited certification body, with a unique certificate number traceable to the issuing CB's register), the accelerometer's individual calibration certificate with sensitivity in mV/g or pC/g and the test conditions, and a declaration of conformity covering the specific serial numbers shipped. The accelerometer selection criteria reference is a useful cross-check when comparing two suppliers' claims side by side.
Failure Modes and Common Pitfalls
Five failure patterns dominate field returns on high-temperature accelerometers: (1) magnetic-base mounting in a high-frequency analysis band, where the user assumes "magnetic" means "calibrated"; (2) wax or beeswax adhesive on a hot machine, where the bond softens within hours; (3) IEPE electronics overheating even though the crystal element is rated, because the amplifier is the weak link; (4) NdFeB magnets in a >120°C environment, leading to irreversible flux loss and a sensor that never recovers its calibration; and (5) cable jacket failure where the cable is rated for 80°C but routed over a 200°C surface [S1][S2].
Each of these is preventable at the specification stage, not in service. The cheapest fix is a written qualification protocol that ties the four limits (element, magnet, mounting, certificate) to specific test evidence on the supplier's documentation, and that rejects any data sheet missing one of the four.
When This Checklist Does Not Apply
This walkthrough targets vibration accelerometers (charge and IEPE) for rotating machinery and structural health monitoring in the 150°C to 350°C band. It does not cover shock accelerometers rated for single high-g events, DC MEMS accelerometers for tilt and low-g inertial measurement, or piezoelectric film sensors for impact detection. Those families follow different certification paths and have different thermal trade-offs, and forcing them into the high-temperature checklist above produces a wrong conclusion. [S2]
It also does not substitute for the safety certification reference, which is the broader standard for hazardous-area equipment marking across all instrument types, or for any site-specific electrical classification that may demand additional witness testing by a notified body. Where the installation sits inside a Zone 0 or Zone 1 boundary, treat the certificate review as a separate workstream, not as one line on the sensor PO.
Track two signals as the qualification cycle closes: the supplier's published high-temperature sensitivity drift curve, ideally plotted across the full operating range rather than a single 25°C number, and the certificate revision dates, since standards are revised on multi-year cycles and a 2018-vintage ATEX certificate may not match the 2026 plant's electrical-classification drawings. For related work on adjacent instrument qualifications, the high-voltage tester reference covers a different but structurally similar checklist for insulation-test gear in the same plant envelope.
For related coverage, see Planetary Reducer Spec Map for Steel Mill Drives.