Classical Shock Testing Control and Analysis Techniques

Back to: Shock Testing

Now that we’ve discussed classical shock pulses and the machinery required to run a shock test, here are several additional techniques that may be useful during testing.

Using Filtering to Determine Velocity and Displacement

Often, an engineer will want to determine the velocity and displacement a shock test generated. If acceleration is the base measurement, additional filters can integrate the acceleration signal into velocity and double integrate the signal into displacement. However, this filtering method can negatively affect the velocity and displacement values.

When gathering this data, the engineer will likely need to change the filter settings to minimize the extremely low-frequency content included in the calculation. The low-frequency content will cause variations in the velocity and displacement calculations. When this occurs, the calculations will appear as “wander” in the time graphs.

This low-frequency content is especially visible when the test runs multiple pulses. The control trace in the displacement versus time graph will have the same shape as the demand line and will begin at the same point but will look offset at some angle. The angle will change between pulses even though the overall shape will remain the same. This indicates that the high pass integration filter needs to be adjusted to remove more of the extremely low-frequency content.

Increasing the cutoff frequency of the integration filter will remove low-frequency content. This will provide a more realistic representation of the velocity and displacement without the effects of the DC offset or other spectral “bleeding” issues. The ability to manipulate the integration filters (and differentiation if using velocity or displacement sensors) is important to consider when investigating different tools for shock analysis.

Shock Tests and High-Frequency Noise

Shock tests can generate high-frequency noise, which can affect waveform analysis. In many cases, the noise is inconsequential to the relative damage of the product as high-frequency content is often associated with low velocity or low displacement.

However, when the sampling rate is too low, high-frequency noise can show up as an aliased noise. When this occurs, a reflection of the high-frequency noise is visible at the lower-frequency multiples. Vibration Research hardware always applies anti-alias filters to the signal in the form of a fixed analog anti-alias filter and a variable digital anti-alias filter proportional to the sample rate. The anti-alias filter will adjust appropriately if the engineer uses a very low sample rate.

Conversely, high-frequency data may need to be removed when the sample rate is too high. The high-frequency data may cause noise, rattling, banging, or the relative damage caused by the high frequencies may be inconsequential. Removing this content can be accomplished through the FIR filter (applied after the abort limits are checked) or an IIR filter (applied before the abort limits are checked).

Engineers often use FIR and IIR filters to clean up graphs and make them more visually appealing. These filters also help to meet test tolerance requirements when sampling at a high enough sample rate to observe high-frequency content.

In the VibrationVIEW Shock software, the engineer has the choice of an FIR filter or a series of different IIR filters. These filters have unique characteristics with different effects that can remove frequency content outside frequencies of interest. Not all shock tests require additional filtering, but, depending on the product, having a system that can apply filtering to the live signal is important for easy analysis.

Conclusion

There are several potential issues—such as spectral bleeding and aliased noise—to keep in mind when running a shock test. It is important to consider these issues when selecting software and other tools for testing.

To access a free demo of Vibration Research’s shock module in VibrationVIEW, click here.