DFAN Notching: Intro and Initial Thoughts

In dynamic testing, there is always a struggle between conducting the best possible (most flight-representative, cheapest) test and protecting the hardware. The structural margins on the hardware in its testing configuration is often so slim that extra hardware protection is warranted. In this case, test engineers will sometimes implement notching. Per https://vibrationresearch.com/blog/what-is-notching-in-vibration-testing/, “A notch is a minimum spectral value at a natural frequency.” These techniques limit the energy at certain frequencies to avoid hardware resonances or over exposure caused by configuration.

For vibration testing where the UUT is directly connected to the excitation source, notching is widely used to avoid off-nominal resonances from occurring. For an acoustic test, its purpose is to limit energy input into certain hardware while trying to maintain the overall acoustic field tolerances.

Two common types of notching are used for DFAN testing. These are input spectrum notching and response limiting. These can be thought of as manual and automatic notching. Input spectrum notching (manual) involves modifying (reducing) the input spectrum that is then sent to the acoustic devices. Typically, notches will be over specific frequency ranges or One-Third Octave Bands but not broadband. Response limiting (automatic) monitors specified transducers on the Unit Under Test and draws energy from the system to make sure the transducer doesn’t exceed a preset level. Both techniques are widely used for a variety of vibration testing and most notching for acoustic tests has its origins in vibration testing.

These notching techniques can be very useful to avoid hardware damage, but they must be used with care. Input spectrum notching ensures that energy won’t be output into the DFAN acoustic field by stopping it from being sent to the acoustic devices. This is a global notch and will be seen by the entire test article in the DFAN test setup. If the notch needed is big enough, the test’s original acoustic test level requirement potentially won’t be met. This could force the test article manufacturer to seek waivers to test to the modified levels. The ideal scenario for an input spectrum notch is a notch that when applied will still fall within the overall tolerances of the acoustic spectrum.

Response limiting for DFAN is useful for test articles that must be tested in a non-flight like configuration. For example, certain test articles have external fairings which act as absorbers/diffusers to the acoustic energy propagating on the test article. If a test on those test articles must be conducted without the external fairings, certain hardware could be over exposed since the fairing is no longer present to modify the acoustic field. In this case, response limiting can be applied to instrumentation on the suspect hardware. If the hardware exceeds the preset limit, energy will be drawn out so the instrumentation can stay within tolerance. This is a local change as the controller can figure out the combination of acoustic devices that are contributing to the hardware response and draw energy out of those sources at the right frequencies. These changes are not applied globally.

There are numerous drawbacks to response limiting for DFAN, however. The main issue is the applicability of response limiting to the “Test Like You Fly” methodology. By using response limiting you are subjecting the test hardware to an acoustic field that it will never face during flight. A flight like test configuration and acoustic field will always be preferred for qualifying hardware. For hardware that has configuration issues (very large test articles, articles with multiple modules tested unstacked), response limiting is a valid method as long as the physics of the limiting wanted are well understood. We do not recommend response limiting be used unless there is a configuration issue since we feel that not using response limiting provides a more flight-like test.

One final issue with response limiting is with verifying controller functionality before use on flight hardware. A cardinal rule in test engineering is to never validate the testing setup using flight hardware. Ideally every DFAN test would use a full featured volumetric/mass simulator to verify the DFAN system functionality before testing the flight hardware. That is a luxury most programs do not have, so most testing on a DFAN system before the flight hardware is introduced is done with a “bare chamber”. However, if there is no hardware in the DFAN circle during bare chamber runs, there is no possibility to verify response limiting functionality before using it on flight hardware. In these cases, it is highly recommended to introduce a piece of test hardware into the circle during bare chamber runs. This test hardware can be anything from a flat honeycomb panel to a more complex structure. This test hardware is instrumented, and the controller’s response limiting is applied to that instrumentation during bare chamber test runs. Ideally, this test hardware would be simple so the analysts can easily model it and come up with response limits that might be like the ones they need to use on the flight hardware. These bare chamber runs will give you confidence that the response limiting will work as designed on the flight hardware.

Notching encompasses a series of useful techniques that can be applied individually or together to protect hardware. However, a deep understanding of the test setup – flight hardware physics must be had to avoid a test that doesn’t meet overall requirements.

At ECSSMET, Dale Schick will be presenting a paper about the ARS implementation of response limiting.