When selecting accelerometers for your test, a lot of attention goes to matching the sensor sensitivity and frequency response with your expected measurement parameters. What we tend to forget, however, is that the manufacturer specification sheet refers to the sensor response under the ideal test and mounting condition and my be quite difficult to replicate under rel world testing conditions.
Sensor performance is normally determined with the sensor mounted with ideal surface flatness and roughness, mounting hole perpendicularity, recommended mounting torque and often also using a coupling fluid/grease (ultrasonic couplant) between the sensor and the mounting surface.
The sensor natural/mounted resonance frequency and amplitude response across the bandwidth are then characterised under these condition and published in a specification sheet. The sensor mounted resonance frequency is always of great importance, as this determines the useful measurement bandwidth for the sensor. As rule of thumb, the useful bandwidth of an accelerometer is 1/5 x Mounted Resonance Frequency.
“Why is this then so important?” – you may ask. Reality is that we can often not replicate these ideal mounting conditions during our tests. This is an important factor, as improper mounting can affect the mounted resonance frequency of your sensor due to the change in damping and stiffness caused by your mounting technique. This directly affects the measurement bandwidth of the sensor, as well as the amplitude response. It is therefore very important to understand the effect that your mounting method will have on the response of your sensor.
Accelerometer manufacturers have published several guidelines and white papers to illustrate the effect of mounting method on their accelerometers. Herewith a short extract to illustrate the effect in terms of frequency and amplitude response when comparing stud mounting, adhesive mounting and magnet mounting of accelerometers.