(1) Vibration characteristics
When working on vibration measurement and analysis, the first step is to feel the vibration and sound by touching or tapping the object. If you tap it with a hard object and the sound is clear, you can expect good data measurement. If the sound is cracked or dull, it's troublesome and requires ingenuity. You can generally predict whether the measurement will be successful based on this.
If you have a 1-channel FFT analyzer on hand, you can start by measuring the vibration frequency and amplitude with Accelerometer. However, if you want to focus on the vibration characteristics, you will basically need an FFT analyzer with 2 or more channels.
Performing vibration tests by inputting signals from an impulse hammer and an Accelerometer, or a vibrator and Accelerometer, into a 2-channel FFT analyzer and measuring the frequency response function (transfer function) means examining the vibration characteristics of the object being measured. Conversely, knowing the vibration characteristics allows us to predict how much vibration will occur at what level of excitation input.
Now, when you go to the site, the first thing that catches your eye is the assembled, completed product. Analyzing the vibration characteristics of a finished product by exposing it to vibration is quite difficult. Only someone with extensive experience and a deep understanding of the product would be able to do it. Finished products are heavily influenced by nonlinear elements such as friction and play, resulting in a jumble of various vibrations. Identifying which component is producing which vibration and to what extent is similar to trying to identify ingredients that have been cooked and then disappeared.
As you've probably experienced, trying to find the target vibration and its signal source from measured data is far from straightforward. The challenge lies in how to measure good, or rather, easy-to-understand, data.
Starting with measuring the vibration characteristics of the basic components, and then analyzing each component before analyzing the finished product, might reveal a more accurate picture through the superposition of vibrations. Furthermore, the success of this series of analyses depends on how well the product's behavior is understood and measured from a vibrational perspective, that is, in terms of its three elements: weight, stiffness, and damping.
(2) Support method
Let's consider the method of vibration measurement for the object being measured. Light objects are often suspended by strings, while heavier objects are usually securely fixed. Various considerations are made depending on the purpose of the measurement, such as whether to allow complete free vibration or to fix the object in place. In free vibration by suspension, the vibration of swaying is also measured.
By ensuring that this frequency is about one-tenth or less of the frequency being measured (the first frequency of the natural mode), it will not get mixed in with the mode frequencies during measurement, saving problems in subsequent analysis. The support points should be at the vibration nodes to minimize their influence on the vibration.
When you want to support heavy objects freely, you can sometimes measure them by placing them on tire tubes or similar materials. A resonance point is created by the weight and the air spring, but the air spring is quite soft and produces low frequencies, making it easy to handle. Sponge or other styrofoam materials are a bit harder and can interfere with damping and the vibration itself, so caution is needed.
When fixing components in place, they are often secured with bolts, but at high frequencies, vibrations can occur in the unsecured areas between the bolts.
Because the vibration amplitude is so small that it cannot be seen with the naked eye, it is easily overlooked and requires careful attention. The erratic movement of images in modal analysis is likely due to this effect. Also, if the fixing force is weak or damping conditions are added, the frequency peaks and damping characteristics may be affected and change. In my personal view, if the first resonant frequency is f1Hz, I consider up to 10×f1 as the low frequency band, up to 20×f1 as the medium frequency band, and anything above that as the high frequency band. In the low frequency band, modal analysis and measurements can be performed stably, but in the high frequency band, it is difficult to perform stable measurements because it is easily affected by linearity, sensor characteristics, and support methods. In the medium frequency band, it is possible to obtain good data by paying attention to the support method, sensor characteristics, and excitation method.
You can choose either fixed support or free support, whichever is easier, but the goal is to ensure that the support method does not appear in the measurement data. Free support is used when supporting at frequencies lower than the frequency band of the object (the frequency band you want to measure), and fixed support is used when supporting at higher frequencies.
Once the method for supporting the object being measured is decided, it feels like half the measurement is done.
(Excerpt from the email newsletter issued on May 20, 2004)
