There are a lot of things to consider when selecting and using a shaker table for environmental vibration (i.e., durability) testing.
Shakers used in vibration testing (Figure 1) have a variety of sizes and performance ratings.
Figure 1: Different shaker types include Multi-Axis (left), Vertical (middle), and Horizontal (right).
Commercially available electrodynamic (ED) shakers come in force ratings from 2 lbf (9 N) up to 80,000 lbf (356 kN) and hydraulic shakers can also produce larger forces and displacements than their ED counterparts. This article primarily focused on the larger (>100 lbf (450 N)) ED shakers used to replicate vibration environments. Smaller (modal-type, transfer path, inertial, etc.) shakers are not covered in this article.
Contents: 1. Shaker Overview 2. Performance Curve 3. Force Ratings 4. Average Control 5. Shaker Health
1. Shaker Overview
A typical large shaker consists of the following components (Figure 2):
Figure 2: Cutaway view of shaker.
Components include:
Armature (black): The moving component of the shaker that the test article is mounted upon.
Voice Coil (orange dots): Copper wire or tube wrapped around armature that has generates a dynamic magnetic field when energized with a time-varying voltage.
Shaker Body (blue): Metallic ferrous material that holds everything together.
Field Coil (light orange): Energized copper coil to create static magnetic field. This is a permanent magnet in smaller shakers
Flexures (magenta): Flexible J-shaped connector between armature and shaker body that allow relative movement between the two
Guide Bearing (purple): Post in middle that armature slides up and down upon, prevents armature moments (tilting). This is only one such design option; designs vary between manufacturers.
Pneumatic Load Support (dark blue): Bears test article weight, keeps armature centered.
Isolation Mount (green): Isolates shaker induced vibration from surrounding environment (test hall, etc). These are typically airbags.
During shaker operation, the amplifier applies a constant voltage to create a static magnetic field in the field coils. Voltage to the voice coil is varied, creating a dynamic magnetic field that moves the armature relative to the field coils.
Equation 1 is the force that the shaker can produce.
Equation 1: Force a shaker can produce.
Where:
F: Force that shaker can supply
L: Length of copper wire
I: Current
B: Magnetic field or sometimes called the “B-field”
The available force to the shaker is dictated by the length of the voice coil winding, the current passing through the voice coil, and the strength of the magnetic field.
2. Performance Curve
The force a shaker can generate is not uniform per unit voltage across the frequency range of operation. A shaker's performance is further limited by its maximum displacement and maximum velocity, especially in the lower frequencies. The frequency range of operation of a shaker is generally defined as 1.5 times the first armature resonance (per ISO 5344). To illustrate a shaker's performance, most manufacturers will provide a performance curve of their shaker before shipment.
Every shaker can be characterized by its performance curve (Figure 3).
Figure 3: Transfer functions vs. drive voltage (top graph) vs. measurement accelerometer spectra (bottom graph).
Shaker performance curves are typically displayed in units of g/V. This is a simple transfer function between the drive voltage going to the shaker amplifier and the response of the bare armature accelerometer. Performing the test in this way yields a performance curve that can be directly compared two previous curves regardless of the input voltage to the shaker amplifier.
As seen in the curves above, despite varying drive output levels of 50, 100, and 500 mV, reflected in the blue, magenta and cyan response curves in the bottom plot, the top plot of the transfer functions remains consistent.
Shaker performance curves have the following sections (Figure 4).
Figure 4: Performance curve of a shaker table is output acceleration (g’s) over voltage output (V) versus frequency.
The performance curve has several distinct features:
A: Maximum stroke at low frequencies (light blue)
B: Maximum velocity at mid-frequencies (light green)
C: Max force (acceleration) at high frequencies (light orange)
D: Frequency range is determined by the first resonance of the Armature x 1.5 by ISO 5344 (grey)
The performance curve is measured on a “bare” shaker armature. A “bare” armature has no fixture or test object on it, only a measurement accelerometer. More on the specifics of performing a test to calculate the shaker performance curve is in the “Shaker Health” section below.
3. Force Ratings
Need to run a 6000 lb force test? Don’t buy a 6000 lb rated shaker! The maximum rating of the shaker needs to be higher than the minimum force level required for a test. Shaker manufacturers typically recommend a 15-20% margin when sizing a shaker for vibration testing.
There are a few reasons why the maximum rating of the shaker should be higher than the maximum level of the test to run:
Shaker force ratings are specified without the mass of the test article and fixture.
Vibration tests are dynamic. Need to account for resonance and anti-resonance behavior.
The type of excitation is important. Force rating limits for the same shaker are different for a random vibration test than a sine vibration test.
The entire moving mass of the system needs to be considered when planning a shaker test. This includes not just the mass of the Device Under Test (DUT), but also all fixturing, fasteners, electrical harnesses, etc. The maximum acceleration available on the shaker is limited proportionally to the mass of the test article due to Newton’s second law. Since force remains constant, heavier DUTs reduce the maximum acceleration a shaker can produce.
Temperature plays an important role in ED shaker performance as well. For every increase of 100°C, resistance increases 40% in copper wire. This can affect the force rating due to the relationship between voltage, current, and resistance. Based on Equation 1, ED shaker force is directly proportional to the current supplied by the shaker amp. If resistance increases, the voltage supply would need to increase proportionally as well to maintain the same force. This is why large shakers always have some type of cooling. Air cooling, water cooling, and other novel solutions are available depending on the manufacturer.
For random vibration, the force rating of a shaker is specified over a finite frequency band, typically up to 2,000 Hz. Testing inside that band reduces the available force of a shaker as seen in Equation 2.
Equation 2: Narrow band force reduction.
4. Average Control
It is not typically recommended to run a shaker test by coupling a DUT directly to the shaker armature. Standard practice is to bolt a very stiff head expander to the armature and couple the DUT to the head expander. The point of this is to try and create a surface with a uniform vibration input at the points where the DUT interfaces with the head expander. However, everything has structural modes, even the best designed fixtures. If the vibration input to the DUT varies due to the structure of the test setup, average control can be used to mitigate this to whatever extent possible.
Large slip tables can sometimes benefit from average control. One situation is if there are different vibration levels along the edge of slip table, as indicated in Figure 5. In this case, two or more control accelerometers can be mounted along the edge.
Figure 5: Average control configuration with multiple control points along the edge of a slip table.
Another possibility is that there are higher vibration levels closer to the slip table interface with the shaker (the “fore” side), called a driver bar or bullnose, than the other side further away from the bullnose (the “aft” side). Assuming a linear roll off in force from fore to aft, placing fore and aft control accelerometers as shown in Figure 6 and using average control should produce the desired reference vibration level between the two control accelerometers.
Figure 6: Average control configuration with “fore” and “aft” control locations.
One of the most common deployments of average control is used when there are multiple DUT connection points, as seen in Figure 7. This can be useful in both vertical or horizontal orientations.
Figure 7: Average control configuration with control points at each mounting point of the DUT.
In this case, control accelerometers are placed at each point where the DUT interfaces with the expander head (or fixture).
All of this is dependent on the exact setup. If any of this is a concern, best practice is to take data from multiple accelerometers on the table to see if there is any significant variation that needs to be addressed with average control or other options.
Well-built shakers can operate for decades, but just like a car, they require periodic maintenance, and can be subjected to high level testing at the limits of their capabilities. Guide bearings can wear; a shaker armature can become deformed, cracked, or have the voice coil winding delaminate; flexures can fatigue over time; and other potential wear and tear can occur. When shaker components fatigue it can affect the performance of the shaker.
To identify these issues, the shaker “bare table” transfer function can be measured periodically. This transfer function is calculated by dividing the spectrum of g’s of acceleration measured by an accelerometer over the spectrum of drive voltage being sent to the shaker amplifier from source output of the shaker controller. This transfer function, or performance curve, of a shaker should remain the same over the life of a shaker. If this transfer function changes, it may indicate a problem with the shaker or amplifier.
Obtaining a shaker performance curve is a straightforward process. The steps to do so are: A. Setting up the SCADAS controller hardware and measurement accelerometer. B. Configuring Testlab Tracked Sine Dwell (or Swept Sine) C. Shaker Amplifier “Calibration” D. Run a Sine Sweep E. Compare Obtained Transfer Function to Legacy Data In a nutshell, the controller will output a constant voltage sine sweep to the shaker, and an accelerometer mounted on the bare shaker armature will record the g level at each frequency line. This will characterize the shaker response at each frequency to the same voltage output level.
A. Physical Test Setup
First the shaker and measurement system should be set up as shown in Figure 8 below.
Figure 8: Diagram of setup to measure shaker health.
Key setup points:
Tee the signal from the SCADAS hardware source output into both an input channel and the shaker amplifier input. This will be designated as the control channel in Simcenter Testlab software.
Place an accelerometer on the bare shaker armature (e.g. without any head expander, fixture, etc.). This will be designated a measurement channel.
B. Configure Testlab Tracked Sine Dwell
The Simcenter Testlab Tracked Sine Dwell module is used to do the measurement. There are two main steps outside of a typical Sine Dwell setup:
Calibrate the shaker output using the amplifier gain to reach a pre-defined measured acceleration at a single frequency. This is typically 1 or 2 g at 100 Hz, but different settings might be desired. If in doubt, check with the shaker manufacturer for recommendations.
Perform a sweep to characterize the shaker system.
