Which Accelerometer Should I Use?
We recently had a series of questions from one of our customers regarding the selection of accelerometers for vibration testing.
Is there a rule of thumb regarding how to select the optimum type of accelerometer and sensitivity depending on the required test conditions? For example, most tests are 3 G peak sine dwell and a 50G, 11mS classical shock pulse. Which type of sensitivity accelerometer would be recommended for each of these test profiles?
There are a lot of different accelerometers available on the market today. The specifications vary based on the maximum acceleration, frequency range, test environment and much more. In most testing environments, we recommend using an IEPE accelerometer with TEDS. The IEPE constant current powered accelerometer provides a clean signal and removes many of the noise sources that are common with other types of accelerometers. TEDS removes much of the operator error. If the sensitivity of an accelerometer is entered incorrectly it is possible to cause significant damage to a shaker. The TEDS chip on an accelerometer holds all of the important information required for a test (Calibration Date, Manufacturer, Sensitivity, Output Range, Model Number, etc). TEDS ensures that the sensitivity of an accelerometer connected to a particular channel is entered properly.
The most common sensitivities for accelerometers in vibration testing applications are 100 mV/G and 10 mV/G. Common IEPE accelerometers can output a ± 5V excitation signal. This means that the range of a 100 mV/G accelerometer with a ± 5V output will be 50 Gpk (5Vpk / 100mV = 50 Gpk) and a 10 mV/G accelerometer will be 500 Gpk (5Vpk / 10mV = 500 Gpk). These accelerometers are capable of generating a signal greater than ± 5V, but beyond that point the response is no longer linear and the overall measurement uncertainty increases significantly.
For the test profiles defined, a typical IEPE/TEDS 100 mV/G would be a great accelerometer for the low amplitude (3 Gpk) sine dwell test, but would not be the best choice for the shock test as defined (50 Gpk). A 10 mV/G accelerometer would make for a much better choice. The vast majority of laboratories have both 10 mV/G and 100 mV/G accelerometers available for the appropriate application.
Is there a rule of thumb concerning the maximum accelerometer cable length?
Simply put, the shorter the better. The cables in a vibration testing system are one of the most easily damaged components and common sources of noise problems in the system. Cables get strewn across the floor, scraped, kinked and the shielding begins to break down. As the shielding breaks down the radiant electrical noise of the environment is induced into the cable and is measured by the vibration controller. This noise is environmental and therefore not controllable by the controller. This applies to both the Drive output cable between the controller and the amplifier and the accelerometer cables between the accelerometer and controller. Eventually, the ground/shield bond established by the cable will fail and any signal generated or measured will have no real reference and can cause damage to other components in the system.
This one of the reasons why Vibration Research uses a 1000 Base-T Ethernet connection between the controller and the PC. This allows the controller to be installed close to the amplifier and shaker and the operator station to be in a control room up to 300 feet (100m) away. The shorter cables (Drive and accelerometer) are more easily maintained, more cost effective when they need to be replaced, and remove a lot of confusion when trying to determine which accelerometer is connected to a particular channel.
The maximum length is very dependent on the type of accelerometer and cable. For instance, if you are using charge mode accelerometers, the distance between the accelerometer and the charge barrel or charge amplifier needs to be short. These accelerometers output very low voltage levels and any noise that is absorbed from the environment into the cable can cause variation in the signal. For IEPE accelerometers the signal is carried on the same wire as the voltage suppling power to the accelerometer. This means that the signal has much more power and is less subject to variation. As cable length increases the inherent impedance and capacitance will change. Depending on the type of cable and the output of the IEPE power supply (2mA vs. 5mA) the maximum length can be up to several hundred feet for a well-insulated, low capacitance straight cable. Joining multiple cables, running cables near high voltage wires and a variety of other factors will have effects on the signal which may cause issues.