Case Study of a Large Scale DFAN Test: Part One

4.3 Combined Control & Data Acquisition System

5.2 Utilization of Novel Operational Assets

Two key operational assets were utilized for this test campaign. The first was a Microphone Support System (MSS), an asset that utilizes the acoustic device as a base for attaching critical test instrumentation. Often the floor of the test area becomes cluttered with cabling and stands required to place microphones in the field. This clutter can cause trip hazards, tip-over concerns, and increases the complexity of spatial navigation for technicians working around sensitive hardware. To reduce these risks, the MSS was designed and deployed with the goal of eliminating the floor footprint required for supporting the microphones used to conduct a DFAN test.

Figure 4: MSS installed for a demonstration.

Each MSS element can be pre-attached to the acoustic devices on the ground prior to the system being stacked. This means that columns of acoustic devices can be built with hardware already safely mounted, bolts torqued to spec, and all microphones pre-positioned to their correct location. This significantly reduces the amount of time required to configure the acoustic test system and reduces the number of trips technicians are required to make into the test circle for microphone installation and calibration.

5.3 Detailed Procedural Documentation for Stacking

The second key operational asset was a detailed procedural document for stacking. The preparation of a detailed procedure for a complex, critical environmental test is not a new topic. However, given the scope of this test campaign and the number of parties collaborating on its successful outcome, more attention needed to be paid to certain portions of the test configuration. This would help to ensure that the risks, safety requirements, and correct steps for the system to be deployed were all clearly understood.

Figure 5: Example from test procedure illustrating detailed stacking operations and signoffs.

Shown in Figure 5 is an example from one portion of the procedure, which illustrates the importance of detailing each step in the assembly process. It is critical to ensure from the very onset of the setup process that not only are all physical connections made safely and correctly, but there are error proofing and traceability steps created along the way. This way, when the system is powered up and checked out, the risk of incorrect connections can be significantly mitigated, if not eliminated, making a notable contribution to the success of the test as well as the timeliness of its safe execution.

5.4 Implementation of Systemwide Grounding Plan

With the implementation of a large-scale data acquisition system – providing a potential connection from flight hardware to DFAN system hardware – running simultaneously with the acoustic test control system, proper grounding was identified as a critically important topic early in the design process. There are several elements in play that need to be considered when designing a plan to safely ground and acoustic test system.

Figure 6: Simplified diagram of system grounding plan.

Each of these subsystems needs to be reviewed regarding where they connect with each other, where they ground to, and how that grounding interacts with the grounding configuration of the test article. For the test described here, the teams were able to collaborate on a plan that ensured the test article, attached support equipment, sensors, and the connected elements of the DFAN system were all grounded correctly. This successfully eliminated potential sources of noise, interference, and other electrical phenomena with no issues seen during testing.

6. TEST EXECUTION & SAFETY

A test campaign was scheduled in collaboration between the DFAN system vendor, spacecraft prime contractor, and host facility support teams. A conservative, nine-day schedule was targeted, which is illustrated in Error! Reference source not found. There can be additional complexity in the establishment of a DFAN test schedule when compared to other environmental tests, as the portable nature of a DFAN system can require the scheduling of trucking, key support staff, and facility spaces that might not typically be occupied for multi-day test campaigns. These additional scheduling considerations also might need to be looked at in the context of other operations in the test campaign such as test data analysis, program approval meetings, or other facility operations that might conflict with resources needed for the DFAN test.

Table 2: Example multi-day test schedule.

A key consideration for DFAN test scheduling is the amount of time required to perform post-test data analysis by the required parties. DFAN is not unique when it comes to the balancing of program schedule goals with the analysis time required to review post-test results and provide safe-to-proceed guidance. For DFAN, there typically needs to be additional consideration for the scheduling of support personnel, and cumulative delays to the program can adversely impact the scheduling of support services such as equipment transportation. As illustrated above (Error! Reference source not found.), the actual time required to conduct a 60-second test run at a given level is typically less than 3 minutes. However, this means that a great deal of the schedule between test runs can be dictated by factors outside of the control of the DFAN test program itself.

6.1 Test Safety & Facility Impact

On the topic of test safety, from the standpoint of executing the test campaign, several key areas were addressed:

If an area of the host facility was noted to likely require personal protective equipment (PPE) during an acoustic test run, then the individuals working in that area would need to be warned of an impending test, given time to apply PPE, and any impacts to critical operations such as lifting actives would need to be noted and resolved prior to the start of an acoustic test run.

6.2 Safety Abort Interlocks

The final item of note from an operational safety standpoint is the implementation of multi-stage, redundant abort capabilities. For the test described here, the following interlocks were utilized to mitigate the risk of anomalous acoustical emissions from the DFAN system during all phases of test:

• Digital mutes in audio processor
• Digital mutes in audio amplifiers
• Acoustic controller abort
• Amplifier safety abort interlock

The amplifier safety abort interlock is a new capability integrated for this test. It comprises of two key parts. First, each amplifier in the system comes with a GPIO (General Purpose Input Output) interface [9], the GPIO’s of all amplifiers in the system can be wired together as part of a pre-designed rack interface system. Second, a handheld controller featuring an integrated power supply was designed and outfitted with a connection interface to provide an uninterruptible connection to the GPIO circuitry of all the amplifiers in the system. Through the application of a constant voltage via the handheld controller to the GPIO of every amplifier, a circuit on each amplifier is activated that shuts down the amplification stage of the amplifier, immediately ceasing the output of any signal. This capability provides an additional level of safety beyond digital audio mutes and controller aborts as it is a physical, analog connection to all devices.

Figure 8: Handheld safety abort switch. 

It ensures that before a test, during any time personnel are present, or during a test run with sensitive hardware, if at any point an abort that cuts the output of the audio system can be called and triggered almost instantaneously. This circuit can also be used preemptively. Whenever personnel are working in the test circle and the system is in its powered state, the risk of damage from unplanned transients or operator errors can be mitigated by placing the system in an “abort active” state until the work is completed, and the circle is again clear of personnel.

7. DFAN TEST CAMPAIGN RESULTS

To conduct the test, a system of 48 purpose-built, full-range ARS Neutron™ acoustic devices were deployed in an 8×6 configuration – eight columns around and six devices high – on a 6m diameter circle. The system was driven with 24 uncorrelated drive signals, each drive fed digitally to two multi-channel amplifiers each via the Dante™ Audio-Over-IP protocol. A 1/3Oct MIMO control scheme was used with a one-second control loop time. The use of uncorrelated noise outputs fed to full-range acoustic devices has been demonstrated in previously published material to provide acoustic field uniformity and diffusivity characteristics that are ideal for acoustic testing. [4] The system utilized 24 control microphones, and 21 monitor microphones were deployed at key locations around the test setup. All microphones were PCB 378A12’s. Note that three monitor microphones were deployed in enclosed cavities within and around the test article, labeled M12 through M14, they are included in the plots below for the sake of completeness but were not utilized for field evaluation.

Figure 9: DFAN system in bare chamber configuration.

Following the installation and connection of the DFAN system, a loop check was performed utilizing the acoustic controller. This provides verification of each individual control loop. Each drive output is fed with a short noise signal, its routing to amplifiers can be verified, and noise emits at a low level (~100dB) from each acoustic device. The controller receives the output of the loop check signal via individual control microphones assigned to each pair of acoustic devices and provides a confirmation of the correlation of drive outputs to control microphone inputs. This process takes approximately 5 seconds per output, for a total of 120 seconds of time to complete a loop check.

The appropriate procedural steps to ensure readiness of the test system and facility were taken, and clearance was given to perform the bare chamber runs. These runs are critical to ensuring that the DFAN system can perform the specified acoustic test levels, and they offer an opportunity to verify power, signal routing, and facility acoustic levels. The first bare chamber run was performed at a -20dB level (~120dB), which enabled staff to traverse the test facility and take handheld measurements of sound pressure levels at key areas while the run was active. This information was used primarily for the purpose of validating the predicted facility sound pressure levels from early in the design process. For the sake of this case study, it can be noted that measured sound pressure levels outside the test area were typically within ~1dB of predicted sound levels. This verification of levels helped to inform decisions regarding where personnel could work, what PPE would be required, and what times that runs could be conducted to ensure the lowest impact to other programs working in the facility.

Figure 10: Simplified diagram of measured A-weighted OASPL levels at key facility locations.

7.1 Combined Bare Chamber Runs

The next set of bare chamber runs were conducted to verify achievement of test level and field uniformity. An approach was employed in which all bare chamber levels (-9dB, -6dB, -3dB, and 0dB) were verified in a single run. Following initial low-level equalization of the DFAN system, the system was scheduled to run for 30 seconds at each level prior to 0dB, and then for the full 120 seconds at 0dB required by the test program. Because the system can rise in level linearly and arrive at the scheduled level without needing additional equalization, there is no risk of under- or over-shoot at the start of each test level. This means that the analysts reviewing the bare chamber data could have 30 seconds of level-correct data for each preliminary level, and then the full 120 seconds of time history data at the 0dB level to review and verify uniformity, diffusivity, and stability. This approach also saved significant time vs. legacy methods, as the entire bare chamber series could be completed safely and effectively in under five minutes. The use of quick-look plots enabled the required test personnel to make an expedient decision regarding the approval to proceed (ATP), and once ATP was given the configuration was broken to install the flight hardware for test.

7.2 ConOps For Intermediary Configuration

The process of moving the acoustic devices and making the test setup ready for installation of flight hardware took approximately 90 minutes, and the restoration of the DFAN system following installation of flight hardware also took 90 minutes. It should be noted, as discussed previously, that calculations had been made to determine the stability not only the fully constructed DFAN system, but also the various partial configurations that were encountered as part of the hardware installation process. These calculations determined that the individual columns of acoustic devices were stable on their own, due in large part to their inherently stable design and the low risk of seismic activity in the region of the test. No counterweights were required for columns, and only one crane was required.

Following the installation of flight hardware, acoustic testing commenced. Due to the proprietary nature of the unit under test, the details of the structural response under test cannot be fully elucidated in this case study. However, the 0dB run will be discussed from the standpoint of DFAN system performance.

The requisite procedural steps to ensure that the unit under test was in a ready state for testing were undertaken, then room clears were given, and safety notifications were made in the facility regarding the use of PPE. The full-level test was then ready to proceed. As with all test runs, a low-level loop check was performed to ensure DFAN system readiness. When the test commenced, the DFAN system came on and equalized at a -15dB level for ~15 seconds. It then ramped in level to -3dB and equalized for ~10 seconds. Then the controller ramped to 0dB and within ~5 seconds began the 0dB run. The 0dB level was held for 120 seconds.

7.3 Power Monitoring Results

During the 0dB run, an Amprobe DM-5 Power Quality Monitor was used to monitor the generator powering the DFAN system. A 150kW generator was chosen for this test as suitable facility power was not conveniently co-located to the test. During the spacecraft 0dB run the DFAN system used an average of 25 amps per line. Utilizing a 3-phase 415/240V mains power supply, this means that the DFAN system needed approximately 20kVa of power to achieve the ~141dB specification. The following plots show power usage and quality captured during the spacecraft 0dB run.

Figure 16: Spacecraft 0dB run, control mics, PSD.

Figure 17: Spacecraft 0dB run, monitor mics, PSD. Note M12-M14 outside of acoustic field.

Figure 18: Spacecraft 0dB run, OASPL of all 24 control mics shown over test time.

7.5 Comparison of Bare Chamber & Spacecraft Runs

For the measurements that follow, the bare chamber 0dB run is compared to the spacecraft 0dB run to observe any potential differences in acoustic field quality between the two runs. Previously published literature tends to endorse the use of a realistic mass simulator [10] to ensure that narrowband DFAN control systems can achieve successful test conditions. The comparisons shown here are intended to illustrate that when utilizing a well-qualified control strategy with 1/3Oct-band MIMO control, the usage of a realistic mass simulator does not appear to be a requirement to achieve acoustic field uniformity and diffusivity goals, as one was not utilized for the test described here. The differences in acoustic field quality between the bare chamber and spacecraft 0dB runs were negligible, and while anecdotal, it can be said that the analysts responsible for qualifying each test run were satisfied with the microphone measurements during both bare chamber and spacecraft runs.

Figure 19: Plot of microphone locations, top view.

Figure 20: Coherence between mics for bare chamber.

Figure 21: Coherence between mics for spacecraft.

Figure 22: Deviation per 1/3Oct band, 0dB runs for control mics comparing bare chamber to spacecraft.

Figure 23: Deviation per 1/3Oct band 0dB runs, for monitor mics comparing bare chamber to spacecraft.

CONCLUSION & FORWARD WORK

Following the completion of the test presented here, it can be stated that all structural responses were analyzed and found to be within expected ranges, especially when responses were compared to those the same driven by diffuse acoustic field conditions in simulation. All acoustic test success criteria were met, and the campaign was completed without incident. The flight hardware was qualified for the selected launch vehicle. Due to the proprietary nature of the hardware under test, more information is not available currently. However, it is the authors hope that in time more detailed structural response data can be released for the sake of enhancing this publication in the future.

This case study of a DFAN system comprising 48 acoustic devices also served as a key step towards the future deployment of a larger configuration of 90 acoustic devices, which will support the testing of a 10m tall payload. Following the completion of that significantly larger test campaign, the publication of an additional case study is planned.

While DFAN has been a topic of interest since the 1990’s, material published on DFAN test operations and safety has been limited despite DFAN’s unique, portable nature. Acoustic field generation and qualification are of the utmost importance to assuring a successful DFAN test campaign, but if a system cannot be designed, constructed, and operated in a safe and efficient manner, its acoustical performance is effectively a moot point. If the larger industry does not widen its field of view to see DFAN testing and DFAN test systems as more than simply microphone measurements and control algorithms, then the benefits of the approach will continue to be weighed down by unforeseen operational challenges, personnel and facility safety concerns, and a lack of clear guidance on non-acoustical topics.

The authors would like to give special thanks to the following individuals for their contributions to the test campaign and case study: Cassandra Dutt, Dann Hayes, Josh Morales, Kevin Hopkins, RJ Kenny, Zachary Crawford.

Reference

1. T. Scharton, D. Anthony and A. Leccese, “Direct Acoustic Test of Quikscat Spacecraft,” in 6th International Congressional Sound & Vibration, Copenhagen, 1999.
2. M. McNelis, W. Hughes, J. Larko, S. Bittinger, C. Le-Plenier, V. Fogt, I. Ngan, A. Thirkettle, M. Skinner and P. Larkin, “Examination of the Structural Response of the Orion European Service Module to Reverberant and Direct Field Acoustic Testing,” in 23th International Congress on Sound and Vibration, London, 2017.
3. R. Roe, “NASA-HDBK-7010,” NASA, 2016.
4. E. Friedlander, E. Kinsella, D. Schick and T. Hoffman, “Qualification of DFAN Field & Improved Acoustic Control Methods,” in 32nd ATS Conference, Los Angeles, 2023.
5. Acoustic Research Systems, “Structural Design Criteria,” Shepherdstown, 2024.
6. P. Larkin, “Developments in Direct-Field Acoustic Testing,” Sound & Vibration, pp. 6-10, November 2014.
7. E. Friedlander, E. Kinsella, D. Schick and T. Hoffmann, “Verifying a Novel Approach to Acoustic Testing: Acoustic Devices & Control Systems,” in 32nd Aerospace Testing Seminar, Virtual, 2022.
8. C. Dutt, “Structural and Practical Implications of Direct Field Acoustic Testing on Host Facilities: A Comprehensive Overview,” in European Conference on Spacecraft Structures, Materials and Environmental Testing, Noordwijk, 2024.
9. Powersoft S.p.A, X Series Manual, Florence, 2019.
10. D. J. A. Hargreaves, “Literature Review of Direct Field Acoustic Noise Testing,” Salford, 2022.