home  >>  engineering information  >>  Shaker System Controllers

General
Several factors are present in vibration test systems that make the use of a vibration controller desirable. One of the most compelling factors is the effects of the complex transfer function between amplifier input voltage and the acceleration response of a control accelerometer. Most vibration tests involve testing at a specified vibration level: acceleration, displacement, etc. The transfer characteristics for these levels vary with frequency due to the complexity of the system components.

Complex System Impedance
If the system power amplifier is used in its voltage source mode, the signal to the shaker will be a relatively good representation of the input signal voltage with some constant gain factor dependent upon the setting of the amplifier front panel gain control.
 
The transfer function from shaker drive voltage to armature coil current is a complex result of both electrical and reflected mechanical impedance characteristics. Low frequency impedance is dominated by back EMF voltage. Mid frequency impedance is typically dominated by the shaker’s AC resistive component, which is not constant. High frequency impedance is usually controlled by armature coil inductance and reflected mechanical reactances.
 
The transfer characteristics between armature coil current and coil force are relatively linear at all but the lowest, high displacement frequency regions. However, the acceleration at any particular point on the shaker armature or the test article is dependent upon the combined armature, fixture and test article mechanical compliance and stiffness matrix. Resonances, damping, etc. produce transfer functions that are not only frequency dependent but are also accelerometer position dependent.
 
This extremely complex combination of transfer characteristics make it impractical to operate vibration systems open loop. Some repetitive testing with constant or nearly identical loads can be done after characterization with a closed loop system, but this is usually not practical. Also, if the desired output level is not acceleration, but displacement or velocity, then there is another layer of complication involving non-linearities in the shaker suspension system, as well as the normal unity or squared relation of velocity or displacement relative to acceleration vs. frequency.

Feed Back
All of the complications mentioned so far can be made relatively transparent if some form of vibration level feed back is utilized. Although some systems are controlled using displacement, velocity or force monitoring for feed back, most systems utilize acceleration for their feed back information. Light weight, electronic accelerometers are available from many manufacturers with a full range of sizes, sensitivities, bandwidths and configurations that make getting a high quality acceleration signal relatively easy.
 
Accelerometers can be placed in critical positions on the shaker or test article and multiple accelerometers can be used to monitor different locations on complex systems.

Controllers
The simplest types of controllers depend on the operator to read and evaluate the feedback signal and adjust the amplifier signal input voltage accordingly. This type of system can be as simple as a sine wave signal generator and an accelerometer monitored by a voltmeter. It is left to the operator to manually make the necessary gain compensation for changes in frequency or desired level specifications.
 
Since most modern accelerometers require a constant current source/buffer amplifier, and most voltmeters read in either average or RMS voltage for AC signals, it can be difficult to read and adjust for peak acceleration with this setup. If the accelerometer has a sensitivity that is not convenient for conversion to voltage, mistakes are easy to make. Random acceleration can be monitored in this fashion more directly because of the RMS nature of most random acceleration specifications, however, an average reading, RMS calibrated meter will inject another error when monitoring Gaussian signals. If the vibration specification involves displacement (pk-pk), it becomes virtually impossible to use this method.
 
“Shaker controllers” vary in sophistication, but usually provide feed back calibrated in acceleration and displacement units useful to vibration testing. Simple manual units are available that provide for frequency and gain adjustment while providing a calibrated acceleration signal in g’s peak. More complex units will feature automatic servo controlled levels with programming and frequency sweep capabilities. Top end controllers utilizing computer technology are available that can control to almost any specification with multiple accelerometers, etc.
 
Click to e-mail Labworks