MAIN MENU

An introduction to shaker vibration testing

Simulating the everyday life of any product is an integral part of its test and validation process. Shaker testing allows you to predict the failure modes of an equipment or component and monitor their dynamic behaviour to improve their reliability.
Contact us
image

HBK Webinar: Introduction to Shaker Vibration Testing

Brian Zielinski-Smith

Webinar presenter

BRIAN ZIELINSKI-SMITH 

Has an Bachelor of Science with Honours in Engineering Design.
He joined Hottinger Brüel & Kjær in 2016 (based in Royston, UK) and is our Product Manager for Shakers & Amplifiers and Shaker Systems.

Shaker Testing Introduction and FAQ

Simulating the everyday life of any product is an integral part of its test and validation process. Shaker testing allows you to predict the failure modes of an equipment or component and monitor their dynamic behaviour to improve their reliability.

Things to consider are the frequency range, the displacement and velocity limits and accuracy of control. These are all published and available for your own comparison on Shakers and exciters

Repeatability, statistical DOF improvement, reliability, measurement, resonant accuracy and control can be an advantage over impact hammers. Nevertheless, hammers are cheaper, quicker, easier to use and often accurate enough if you are an experienced user.

Shakers can be very helpful when validating tyre-only models, whereby reference measurements are dynamic stiffness transfer functions. These measurements can’t be achieved on a rolling tyre test, which is best suited to the classical cleat test.

Yes. This is all part of the control software and the demand is flexible. You can manually or automatically set a desired frequency (within the range of the shaker).

Actual displacement/motion of the armature differs from system to system and is dependent on the test profile being used. For information on displacement, please view the data sheets, which can be found on the website here: Shakers and Exciters 

The displacement is high at low frequencies (typically less than 20 Hz) as the acceleration amplitude will demand more movement to achieve the rate of change of velocity. This can be done through calculation at any frequency and any amplitude. If 3 sigma is used for random tests, then the theoretical peak value will be 3 × the RMS.

This depends entirely on the product and its material properties. There are typical burn-in tests to find weaknesses in circuit boards, solder joints and electrical components that normally soak a product at extreme temperatures and can also include random vibration.

An example of these temperature soak tests could be –40 to +70°C for commercial parts, or –54 to +93°C for military components. The time at temperature is also variable and will depend on the thermal mass of the test item. The typical minimum is 2 hours with a maximum of 16 hours.

The digital control system provides the adaptive control to the mechanical non-linear response of any shaker. The electrodynamic LDS shaker has protection to prevent overtravel and the PA will also have limits to prevent current and voltage if you try to run an impossible test. The transfer function is achieved by the controller and, if this is not possible, the controller will stop the test using various protective parameters and tolerances. The shaker limits are all held in the controller and will report any parameter failures before, during and after a test.

The ambient temperature, in situations where the amplifier might be in a hotter than normal environment, without climatic control, causes an amplifier to run hot.

You need to protect the shaker with thermal barriers. You will also need temperature-resistant accelerometers and cables. Moreover, the calibration of the accelerometers will require temperature inclusion. And you must use an isolated stud for the attachment of accelerometers. When finishing the test, you must avoid temperature shock and control the environment back to ambient temperature before entering the enclosure.

There are many factors to design your fixture and should be dealt with by an experienced design engineer. Some things to consider are:
  • Material preference is aluminium or magnesium alloys for strength to mass ratio and good damping properties
  • Low mass, high stiffness and a minimum thickness of 20 mm
  • Specified flatness tolerance and a good balance/C of G
  • All interface hole patterns should incorporate counterbored holes, which allow for a flat washer and cap head screw/bolt
  • Steel inserts should be used for interface to Jig
  • 100 mm minimum hole spacing, in a square pattern where possible and consider how to control and where to position the accelerometers using 10/32 threaded holes

The calculation for this is proprietary and does not include a dynamic setup with fixtures and products. This makes it difficult to accurately calculate without first taking a measurement. The voltage will vary with velocity and the current will vary with the acceleration and force requirement. However, the velocity and acceleration will also change with the characteristic of the overall setup and varies with frequency.
It is best to measure during a ‘low level’ test and confirm with your reference by using a proportion factor for future tests.

Yes, via the vibration controller. PSD refers to a spectrum of random signals and is used as a control profile into the shaker. If the shaker has the frequency range of the test, it will respond to the drive signal and provide the overall spectrum requested by the PSD profile. The only restrictions will be the peak limits of your shaker in terms of D, V, A and Force against the demand PSD and overall GRMS calculation.

Generally, the accelerometer mass is insignificant to a beam and will only become an issue if the accelerometer used approaches the mass of the beam. If this ratio is better than 10:1, the error is well within the uncertainty of measurement stated on the overall test results.

A linear sweep is in Hz/sec, the log sweep option is typically Oct/minute but can also be Dec/min. Therefore, a linear sweep test time is the range multiplied by the number of Hz every second. A log sweep uses a log calculation:

FAQ22 shaker testing

The actual control is still completed by the accelerometer measurement feedback by calculating the acceleration amplitude equivalent to displacement at each frequency. The transition frequency can be calculated by comparing the increase in acceleration with the displacement constant:

FAQ23 Shaker testing
G = Constant 9.80665 m/s² or 386.0885826772 in/s²
D = Displacement peak to peak

This is a performance versus cost question, as they are completely different. The cooling method is more complex with water cooled but similar in operation. The main difference is armature and field coils either incorporate water flow, or they have air flow, to remove excess heat. Our largest air-cooled shaker is 80 kN, but for better performance, LDS increases the cooling efficiency with a water chilling method.

This varies from system to system. All details can be found in the data sheets: Shakers and exciters

This varies from system to system, all details can be found in the data sheets: Shakers and Exciters 

For clarification, RoR and SoR are described in most test specifications:
  • Random-on-Random: narrow bands of noise sitting on a broader band of random energy
  • Sine-on-Random: single tracked pure sine frequencies sitting on a broader band of random energy

The lowest frequencies are for solid trunnion 1 Hz and for air isolated 5 Hz, but it will also depend on the test load. The highest achievable displacement depends on the size of the shaker, but typically an air-cooled medium range is ±25 mm between 5 Hz and 10 Hz.

Yes, but only for a very short period, otherwise damage may occur and reduce the life of the shaker components. For the best results, you should always stay around the 80% rated capacity. This is the same for all shakers on the market.
In addition to this, we would state that you should size the shaker to be above this rating to prevent issues caused by measurement errors, dynamic responses, control position tolerances and overall calculation errors of practice vs theory. 100% is achieved at the factory under strict conditions but we do not expect our customers to try this. The maximum values are there as a performance guide of a perfect setup to provide limits for the 80% calculation. There is typically a 10% overall uncertainty of measurement, which means that 90% limitations are already restricting the actual measured capability.

A drop test normally describes a physical drop onto a surface plate to observe impact damage. This is not a classical shock test or possible with a shaker. The shaker shock can be compared to a free fall shock machine and will successfully control various classical pulses.

Sine sweeps are often used to find the detail of resonances and can be used to test endurance or durability of a product as part of the initial development test.

Support Content