ISO 5348: Mechanical Mounting of Accelerometers
ISO 5348 — “Mechanical vibration and shock — Mechanical mounting of accelerometers” — is one of the most quietly important standards a vibration analyst will ever use. It tackles a factor that silently governs data quality: how the accelerometer is physically attached to the machine. The standard sets out the practical mounting methods, explains how each one shapes the frequency response of the measurement, and shows why the wrong choice can throw away the very high-frequency information you went looking for. Following its guidance is essential for accurate, repeatable readings — above all when chasing high-frequency faults in bearings and gears.
1. Why the Mount Is Part of the Sensor
The single thesis running through ISO 5348 is that the mounting method is not an accessory to the measurement — it is part of the measurement system. An accelerometer bolted to a surface forms a tiny spring-mass system with the structure beneath it, and that system has its own resonance, the mounted resonant frequency. Above that resonance the reading is no longer trustworthy. A stiff, light, well-prepared mount pushes the resonance high, opening up a wide usable band; a soft or heavy mount drags the resonance down and acts as a mechanical low-pass filter, attenuating or damping high-frequency vibration before it ever reaches the crystal. You can estimate where that boundary falls for a given configuration with an accelerometer mounting resonance calculator, which makes the trade-off concrete before you collect a single point. The standard introduces the methods it evaluates in detail — stud, adhesive, and magnetic mounting, plus hand-held probes — and frames everything in terms of stiffness, mass, surface preparation, and the highest frequency at which data remains reliable.
2. Stud Mounting — the Reference Method
Stud mounting is presented as the optimal, reference-grade technique. A hole is drilled into the machine structure, tapped with a thread, and the accelerometer’s mounting stud is screwed directly into it. The standard specifies that the mounting surface must be clean, flat, and smooth — with a spot face machined where necessary to achieve it — and that a thin film of silicone grease or a similar coupling fluid be applied to the sensor base. That film fills microscopic surface voids, maximises the real contact area, and improves the transmission of high-frequency energy.
The payoff is the highest possible mounting stiffness and therefore the highest mounted resonant frequency, which in turn gives the widest reliable measurement band, free of corruption by the mount’s own resonance. Stud mounting is the benchmark against which every other method is judged, and it is the only acceptable choice for permanent monitoring installations, demanding high-frequency diagnostics such as bearing and gear analysis, and for sensor calibration.
3. Adhesive Mounting — a Strong Semi-Permanent Option
Where drilling into the machine is impractical or forbidden, adhesives offer a semi-permanent alternative. ISO 5348 distinguishes between adhesive types. For the best result it recommends a hard, rigid adhesive — a cyanoacrylate (“super glue”) or a two-part epoxy — applied as a minimal, very thin, rigid bond line between the sensor base and the machine surface. The governing principle is rigidity: a thick or soft adhesive such as silicone rubber behaves as a damper and severely curtails the high-frequency response.
Done correctly on a properly prepared surface, a rigid adhesive mount reaches a usable frequency range nearly as high as a stud mount, making it a credible substitute for many diagnostic tasks. The standard also covers adhesive-mounted bases — small metal pads glued to the machine that provide a repeatable location for attaching a stud-mount sensor, marrying the convenience of bonding with the repeatability prized for trend analysis.
4. Magnetic Mounting — Convenience at a Cost
Magnetic bases are ubiquitous in portable, route-based data collection because they are so quick to use, but ISO 5348 is blunt that this convenience comes at a real cost to data quality. A magnetic mount is inherently less stiff than a stud or adhesive mount, and the magnet itself adds significant mass to the sensor assembly. Lower stiffness combined with higher mass drives the mounted resonant frequency down sharply, which severely limits the usable upper frequency of the measurement.
The standard makes clear that high-frequency data — typically above about 2,000 Hz — collected through a magnet is often unreliable. It offers practical advice for getting the most from a magnetic mount: use a strong, two-pole magnet, ensure the contact surfaces are perfectly clean and flat, and apply firm pressure when seating the magnet. Even so, the analyst must accept the compromised band; for serious high-frequency bearing or gear work, a stud or adhesive mount is strongly preferred. A magnet is best reserved for lower-frequency surveys such as unbalance and misalignment checks, where the frequencies of interest sit comfortably below the lowered resonance.
5. Hand-Held Probes (“Stingers”)
The standard addresses hand-held probes — often called stingers — sometimes used for quick checks or in hard-to-reach spots, and strongly discourages them for any serious diagnostic work. The human body is a highly effective low-pass filter and damper, and it is impossible to hold a probe with consistent pressure or at a perfectly perpendicular angle. The result is poor repeatability and a frequency response often limited to less than 1,000 Hz. A probe may confirm a large, low-frequency vibration such as a severe unbalance, but it is unsuitable for reliable trend analysis or for detecting high-frequency bearing and gear defects.
6. Surface Preparation and Cabling
A closing section gives practical advice that applies regardless of method. The mounting surface must be properly prepared: as flat and smooth as possible, with paint, rust, and dirt removed so that there is direct metal-to-metal contact (or metal-to-adhesive-to-metal). For stud mounting, a spot face should be machined wherever the surface is not perfectly flat.
The standard is equally firm about cabling. The cable should be firmly tied down to the structure a short distance from the sensor. This provides strain relief for the connector and, more importantly, prevents cable motion: a cable left to whip during measurement can generate a spurious low-frequency electrical signal through the triboelectric effect, contaminating the true vibration signal and producing erroneous data.
7. The Four Key Concepts to Carry Away
- Frequency response is everything: the mount acts as a mechanical filter. A poor mount — a magnet, say — adds mass and removes stiffness, forming a low-pass filter that cuts off high-frequency vibration before it reaches the sensor.
- Stiffness is paramount: to transmit high frequencies faithfully, the connection between sensor and machine must be as stiff and as light as possible — which is exactly why a direct stud mount outperforms every alternative.
- Convenience trades against accuracy: magnetic mounts are fast for route work, but the usable band shrinks. For high-frequency bearing or gear analysis, choose stud or adhesive.
- Repeatability protects trending: using fixed mounting pads for repeatable sensor placement ensures that changes in the data reflect the machine’s condition, not variations in measurement technique.
8. ISO 5348 in Practice with a Portable Analyser
These principles are not academic — they decide whether a field measurement means anything. A portable two-channel analyser such as the Balancet-1A is used both for diagnostics and for field balancing, and the same mounting discipline applies on every job. For routine kusawazisha the dominant signal is the once-per-revolution running-speed component — a low frequency that even a clean magnetic mount captures faithfully, which is why magnets remain perfectly acceptable for balancing surveys. But the moment the question turns to a suspected bearing or gear fault — where the diagnostic energy lives at high frequency — ISO 5348 dictates a stud or rigid-adhesive mount on a properly prepared surface, with the cable tied off, so the high-frequency content is not lost to a soft interface. Choosing the mount to match the frequencies you are hunting is the practical heart of the standard, and it pairs naturally with sensible sensor mounting practice and consistent baseline data for reliable long-term trending.
9. Where ISO 5348 Sits Among Related Standards
ISO 5348 governs how you attach the sensor; companion standards govern how you judge what it reads. Vibration-severity assessment that was historically split across ISO 10816 and the older ISO 2372 now lives in the modern ISO 20816-1 series, with industrial-machine limits in ISO 20816-3. The data those evaluations rely on is only as trustworthy as the mount that gathered it — which is precisely why ISO 5348, unglamorous as it is, sits at the foundation of credible condition monitoring.
