A vibration sensor can only report what the mounting lets through to it. How you attach an accelerometer to a machine sets the upper frequency limit of the data you get, and a careless mount silently throws away the high-frequency information that bearing and gear faults live in. The difference between a solid stud mount and a magnet stuck on paint can be the difference between catching a failing bearing and missing it entirely. This guide walks through the main mounting methods, the usable frequency ceiling each one imposes, why a loose or soft mount corrupts high-frequency data, and how to place a sensor so measurements are repeatable on both route and permanent installations.
Vibration sensor mounting in one line: Mounting a vibration sensor correctly means attaching it to the machine in a way that stays rigid up to the highest frequency you need to measure. The main methods, from best to worst high-frequency performance, are stud mounting, adhesive mounting, magnetic mounting, and a handheld probe, and each has a lower usable frequency ceiling than the one before it. A soft or loose mount introduces a mounting resonance that amplifies and distorts high frequencies and cuts off the range above it, so the mount must be chosen for the frequencies of interest and placed at a rigid point close to the bearing being monitored.
The best method is a stud mount, where the sensor is screwed directly into a threaded, flat, clean spot on the machine, often with a thin film of coupling grease to fill microscopic gaps. This makes the sensor and the machine act as one rigid body, so vibration is transmitted faithfully to the highest frequencies the sensor can measure. Stud mounting gives the widest usable frequency range of any method and is the standard for permanent installations and for any measurement that must reach into the high-frequency bearing range. Its cost is that it requires drilling and tapping a proper mounting location.
Adhesive mounting bonds the sensor to the surface with an epoxy or a threaded mounting pad glued in place, and it comes close to stud performance when done well with a stiff adhesive on a clean surface. It is the practical choice where drilling is not allowed or not feasible, and it is common on permanent installations. Its usable frequency range is high but somewhat below a true stud mount, because the adhesive layer, if thick or soft, adds a little compliance that pulls the upper limit down. A thin, hard bond line preserves the most range.
Magnetic mounting attaches the sensor through a magnet, which is fast and needs no surface preparation, making it the mainstay of walk-around route collection. The trade is a markedly lower frequency ceiling, because the magnet-to-surface interface is not as stiff as a stud or a good bond and it resonates at a lower frequency, cutting off the usable range well below what a stud allows. A flat-faced magnet on bare, clean metal performs better than a two-pole magnet on a curved or painted surface. The handheld probe, a sensor pressed against the machine on a stick, is the least rigid of all and usable only for low frequencies and rough checks, because the hand and probe introduce large, unrepeatable compliance.
Every mounting method turns the sensor and its attachment into a small spring-mass system with its own natural frequency, called the mounting resonance. Below that resonance the sensor faithfully reports the machine's vibration. As the measured frequency approaches the mounting resonance, the mount begins to amplify the signal, so the sensor reports more vibration than is really there. Above the resonance the mount stops transmitting faithfully at all and the data becomes unusable. The mounting resonance is therefore the ceiling on the usable frequency range, and stiffer mounts push that ceiling higher.
This matters most for bearing and gear diagnostics, because those faults show up as high-frequency vibration, exactly the region a soft mount cannot deliver. A magnet or a probe with a low mounting resonance may cut off below the frequencies where an early bearing defect first appears, so a technician using a soft mount can look at clean-looking data and conclude the bearing is fine when in fact the mount simply never passed the frequencies that would have shown the defect. Worse, the resonance itself can add a false peak that is mistaken for a real fault. Either way, a poor mount produces misleading high-frequency data.
The practical consequence is to match the mount to the measurement. For overall levels and low-frequency faults like unbalance and misalignment, a magnet on a route is entirely adequate. For bearing and gear analysis that needs high frequencies, the mount must be stiff enough that its resonance sits well above the frequencies of interest, which usually means a stud or a good adhesive pad rather than a magnet. Knowing the mounting resonance of the method being used, and staying comfortably below it, is what keeps the data honest.
Where the sensor goes matters as much as how it is attached. The sensor should sit on a rigid part of the machine in the direct load path of the bearing being monitored, as close to the bearing as practical, on the bearing housing rather than on a thin cover, a fan guard, or a piece of sheet metal that flexes on its own. Mounting on a flexible part measures the part's own motion rather than the machine's, and adds resonances that have nothing to do with the machine's health. A solid location near the bearing gives the truest picture of what the bearing is doing.
Repeatability is what makes trending possible, and repeatability depends on measuring the same point, in the same direction, with the same mount every time. Vibration amplitude changes with position and with the axis of measurement, so a reading taken at a slightly different spot or angle each visit produces a trend full of noise that hides real change. On walk-around routes this is handled by marking fixed measurement points, often with a small mounting pad or a spot-faced target, so every technician places the sensor identically and the horizontal, vertical, and axial readings are always taken the same way. Consistency of location beats precision of a single reading when the goal is to see change over time.
For permanent online monitoring, such as sensors feeding a cloud platform like Merobix, mounting is done once and done well, which removes the repeatability problem entirely: the sensor never moves, so every reading comes from the same point with the same mount, and the trend is clean by construction. That is a major advantage of permanent installation over routes. Because the mount is permanent it is worth doing as a stud or bonded pad with a high frequency ceiling, so the online data reaches the bearing frequencies. Getting the mounting right up front is what lets the platform trend a machine's vibration continuously and reliably without anyone returning to the sensor.
Stud mounting gives the widest usable frequency range, because screwing the sensor directly into a clean, flat, threaded spot makes it act as one rigid body with the machine and transmits vibration faithfully to high frequencies. Adhesive mounting with a stiff bond comes close. Magnetic mounting has a noticeably lower ceiling because the magnet interface resonates at a lower frequency, and a handheld probe is the most limited of all. Choose the method whose mounting resonance sits well above the highest frequency you need.
A magnet mount has a lower mounting resonance than a stud or bonded pad, which sets a lower ceiling on the usable frequency range. Early bearing defects show up as high-frequency vibration, and if the magnet's ceiling falls below those frequencies, the sensor simply never passes them, so the data looks clean while the defect goes undetected. The magnet interface can also add a false resonant peak. For bearing analysis a stiffer mount is needed so its resonance sits above the bearing frequencies.
Place it on a rigid part of the machine in the direct load path of the bearing being monitored, as close to that bearing as practical, on the solid bearing housing rather than on a flexible cover, guard, or sheet-metal panel that flexes on its own. Mounting on a flexible part measures the part's motion instead of the machine's and adds unrelated resonances. For repeatable trending, use the same point and the same measurement direction every time, marking fixed points on walk-around routes.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.