Augmentation Study of the Naval Research Laboratory Acoustic Chamber

Dale Schick
Dr.-Ing. Thomas Hoffmann
m+p international
Verona, NJ

Eric Friedlander
Acoustic Research Systems
Shepherdstown, WV

Abstract

Previous studies aimed at electro-dynamic chamber augmentation of the Reverberant Chamber Acoustic Test (RFAT) Facility at the Naval Research Laboratory demonstrated functionality of in-chamber devices. This study builds on that scaffolding to explore the possibility of an on-chamber electro- dynamic option that can eliminate the need for nitrogen driven horns and EPT’s. Real-world test data are presented alongside traditional RFAT test results to compare the benefits and limitations of such an endeavor. Furthermore, the existing passive room acoustic properties are surveyed to identify opportunities for improvement.

Introduction

Reverberant Acoustic Testing Facilities (RATF) have been traditionally used to qualify spaceflight hardware for launch by exposing the unit under test (UUT) to similar input spectra as recorded or assumed by the launch vehicle manufacturer. As launch vehicles evolve, the specifications for payloads are also changing to include more high-frequency energy or to envelope several rockets when the payload manufacturer is unsure of which will be used. RATF chambers typically provide excitation between 30 and 3000 Hz, on average, but are unable to control the energy up to 10,000 Hz, where many of the test specifications require input. Lower frequencies (between 20 – 80 Hz) are also typically difficult to control in chambers depending on their size and horn configuration. It is beneficial in these cases for chamber operators to explore supplemental drivers to augment the chambers existing excitation system.

Previously published efforts (Hayes, 2015) to reinforce reverberant chambers with electrodynamic transducers focused around using a quantity of loudspeaker enclosures placed inside a chamber to excite the room alongside the existing GN2 driven horns. The expectation was to use the electrodynamic loudspeakers while controlling with a narrowband MIMO control strategy to enhance the frequency response of the existing electropneumatic horns and 1/3oct-band acoustic control system inside the chamber.

While this approach did offer improvement to frequency response by providing control for certain frequency bands, especially in the frequencies below 40hz, one of the primary issues with this approach was that the large, acoustically damped wooden loudspeaker enclosures made a significant impact on the reverberation time (RT) of the reverberant chamber. This was exacerbated by the need for the DFAN system to utilize multiple different types of enclosures (Mid-High, Sub-bass, etc.) to address the entire active frequency range of the chamber.

The results of utilizing an electrodynamic system to drive the response of the chamber showed promise, but the downsides of this approach – namely the reduced reverb-time and challenges of operating in the chamber with minimal floor-space – outweighed the benefits at the time of this study and the augmentation of the chamber with electrodynamic loudspeaker enclosures was not explored further.

NRL Chamber

The site of the previously referenced study was the Reverberant Chamber Facility at the US Naval Research Laboratory in Washington D.C. as was this paper’s study. The chamber itself is roughly 6.5 meters deep, 5.2 meters wide, and 8.2 meters high and contains a LING A340 shaker for combined vibroacoustic environments. The horns, which include a single low frequency electropneumatic transducer (EPT) and two small high-frequency EPT’s, only provide controllable energy between roughly 45 Hz and 4 kHz and require massive amounts of gaseous nitrogen (GN2) to operate. The benefit to using these GN2 driven EPT’s is that the level of output, though difficult to control, is quite high. Typically, levels are attainable between 145 dB – 160 dB OASPL, depending on the frequency requirements.

NRL Chamber Horn Configuration

Figure 1 – NRL Chamber Horn Configuration

For modern launch vehicles, even enveloped 1/3 octave specifications provided by launch manufacturers rarely exceed 150 dB, even when 3 dB is added for proto-flight or qualification levels. Since many test specifications require energy between 20 and 10 kHz, it was proposed by the authors that a study be conducted to explore the feasibility of not simply augmenting a chamber, but to completely drive the excitation of the chamber with electrodynamic devices.

2021 NRL Chamber Study

One option to replace the GN2 driven EPT system is to simply arrange concert speakers in the chamber, in the typical fashion of a DFAN test. This, however, would serve to reduce the usable space of the chamber as well as effectively diminishing the broadband reverberating features of the room. Additionally, the UUT would need to be placed prior to arranging and stacking the speakers using the small and minimally transferable in-chamber crane.

Power delivery is another concern in the case of using traditional concert systems. Fully populating the chamber with such a system would likely require more current than is readily available from the facility. Though an external generator could be used, this would add yet another complication to completing tests and the space under consideration was not necessarily accessible to a large external power generator.

If an electrodynamic system were to realize the proposal of replacing the nitrogen driven system, the system’s concentration of energy would need to be sufficient that it would require minimal surface area. It was thusly proposed to conduct a study that placed the devices on-chamber rather than in-chamber.

Example Theoretical On-Chamber Electrodynamic Solution

Figure 2 – Example Theoretical On-Chamber Electrodynamic Solution

Theoretically, an on-chamber electrodynamic excitation system could occupy one or more walls of the chamber and utilize the reverberant features the room was originally designed for. Additionally, the Electrodynamic Drivers ability to clad the insides of the electrodynamic devices with acoustically reflective surfaces would help maintain the overall RT of the chamber.

To realize such a system, it was projected that the electrodynamic drivers would need to emanate from a surface area nearly five times larger than the primary low-frequency horn already occupying the lower portion of the rear wall. Therefore, it was proposed that modifying the walls of the chamber to remove the existing horns as well as cutting an opening in the wall where the acoustic devices could be inserted flush to the wall on the existing chamber.

While the previous study showed the potential promise of an electrodynamically driven solution for reverberant chamber electropneumatic system augmentation, a successful replacement would require a more holistically designed electrodynamic solution that offered the same full-range benefits while in an enclosure that would address the negative impacts of the previously discussed loudspeaker enclosures on chamber reverberation time, required floor space, and general efficiency.

During early demonstrations of pre-production NEUTRON acoustic devices, it was noted by some attendees that the hybrid multi-band horn design of the device, combined with its ability to operate as a single full-range acoustic source without requiring summation from neighboring devices (such as in a concert loudspeaker) might make the device well-suited for revisiting the study.

The unique design of this acoustic device offers the following key benefits for reverberant chamber augmentation:

1. The mouth of the acoustic device can be directly coupled to the chamber via cut-outs in the chamber wall, maintaining the open floor-space in the chamber.

2. Directly coupling to the chamber offers the added benefit of an impedance-matched device, like that of a nitrogen-driven horn.

3. The efficiency of the device would allow for the chamber to be operated on available facility power.

4. The acoustic efficiency and output capability of the device ensures that chambers can be operated for long times during full-level tests without concerns for the electrodynamic devices overheating and becoming damaged during a test.

5. It’s hybrid design allows operation either as a full-range single acoustic source, or in various acoustically coupled configurations.

It was proposed that demonstrating this system would be possible by utilizing the existing chamber door, but in the open position. The devices were arranged to occupy most of the open space and the remainder sealed with a surround constructed on-site. The following describes that operation, as well as results, of the study.

Equipment List:

Equipment Name QTY Model Manufacturer

1. Acoustic Device 6 NEUTRON ARS

2. Device Cables 6 ML-16 ARS

3. Amplifier Rack 1 CORERACK ARS

4. Microphone/Cable 16 378A12/003EBAC-0053 PCB

5. Control System (24 in/12 out) 1 VibRunner (VRB/T2S2/AI810)
m+p International

6. Post Analysis Software – m+p Analyzer m+p International

7. Reference Microphones 2 378A12 w/ 426A14 PCB

8. Microphone Calibrator 1 4228 B&K

To begin the study, the doors of the chamber were opened and locked into position. The acoustic devices were loaded into the facility via truck and rolled to the reverberant chamber area and a crane was used to arrange vertical columns of the devices tall enough to fill the height of the open chamber doorway. They were then arranged in the open doorway pointing inward and with their front faces flush to the surface of the chamber wall. While this activity was being completed, a single amplifier rack and distribution system was placed in the hallway outside the chamber, where three-phase (120V/208V), facility mains power was connected and the cables were connected, one from each acoustic device to its corresponding connection point on the amplifier rack. While 240V/415V power was preferred, the facility only had 120V/208V power available locally, and the system was able to operate on the lower voltage.

Once the equipment was in place, a temporary frame was constructed around the gap between the doorway and the block of six acoustic devices. The frame was constructed of plywood cut to fill exactly the space between the acoustic devices and the existing doorway, and it was backed by 2” foam and an air gap. The simplistic design was robust enough to ensure the study could be performed effectively while avoiding excessive sound leakage from the chamber into surrounding working area.

Acoustic Device Arrangement

Figure 3 – Acoustic Device Arrangement

The control system and management software were loaded into the working lab next to the reverberant chamber and connected by patch bays via multi-core cables, which enabled sixteen microphones to span the distance between chamber and control center.

Microphones were installed in the chamber on both microphone stands and on pre-existing hanging lines. Following a microphone calibration and short system check, testing was ready to begin.

Test – Coherent vs. Incoherent Operation

To understand the effect of multiple incoherent sources in the reverberant field, a test was made using both incoherent and coherent control strategies with no changes to setup between the two.

Using microphone configuration #1, as shown in Figure 1Figure 4, which was designed to operate the system similar to an incoherent direct field acoustic system, with a microphone in front of each acoustic device in the direct field, to determine if the chamber benefited from being driven by multiple incoherent sources. This also provided a good reference connection between previously performed direct field acoustic testing studies with the system in a non-reverberant environment, and this first test in the reverberant environment, but using the same control strategy.

Microphone Configuration 1 – Drawing

Figure 4 – Microphone Configuration 1 – Drawing

The above microphone configuration was not intended as a realistic use-case for in-chamber testing, but rather to provide a comparison to direct-field control methods employed in a typical environment. Studying the control arrangement in this way helps characterize the effects of the chamber on the field. The control plots were somewhat as expected, with the characteristic uniformity of a single-output system, but without the presence of standing waves which tend to deviate more significantly in DFAN device arrangements.

Coherent Test - 1/3 Oct Resolution, C1 - C6

Figure 5 – Coherent Test – 1/3 Oct Resolution, C1 – C6

Coherent Test - 1hz Resolution PSD, C1 - C6

Figure 6 – Coherent Test – 1hz Resolution PSD, C1 – C6

In this arrangement, with the control microphones technically in the direct field, the monitor microphones, placed above the devices, are more significantly out-of-family with the rest of the data.

Coherent Test - 1/3 Oct Resolution, M1 - M4

Figure 7 – Coherent Test – 1/3 Oct Resolution, M1 – M4

Coherent Test - 1hz Resolution PSD, M1 - M4

Figure 8 – Coherent Test – 1hz Resolution PSD, M1 – M4

In contrast to the single output test the incoherent test, utilizing a single drive output for each individual acoustic device, controlled with a much tighter grouping in both 1/3 octave and narrowband plots.

Incoherent Test - 1/3 Oct Resolution, C1 - C6

Figure 9 – Incoherent Test – 1/3 Oct Resolution, C1 – C6

Incoherent Test - 1hz Resolution PSD, C1 - C6

Figure 10 – Incoherent Test – 1hz Resolution PSD, C1 – C6

Since the control locations remained in the direct field of the devices for this test the monitor locations exhibited the same lower-than-acceptable levels but maintained the more uniform grouping as the previous test.

Incoherent Test - 1/3 Oct Resolution, M1 - M4

Figure 11 – Incoherent Test – 1/3 Oct Resolution, M1 – M4

Incoherent Test - 1hz Resolution PSD, M1 - M4

Figure 12 – Incoherent Test – 1hz Resolution PSD, M1 – M4

The results of the above configuration provided the desired baseline for comparing direct field acoustic testing to the new approach of reverberant chamber operation with the acoustic device. It was found that the control system response was predictable but was able to utilize a faster compression speed than is typical of reverberant chambers, requiring less ‘build-time’ to achieve a similar level. Early results showed that the chamber was highly uniform from a frequency response standpoint in the low and mid frequencies, however the higher frequencies saw some drop-off in level above 5kHz in the suspended monitor microphones that had been deployed in their standard position for RFAN testing. Furthermore, the control microphones would need to be placed more generally around the space typically occupied by a unit-under-test (UUT) to effectively drive the energy closer to the reference spectrum. Following the observation of attenuation in the monitor microphones, it was determined to run another test, but this time with the suspended microphones acting as the controls for the test.

Test – Radiator Panel

The purpose of the second configuration was to control the chamber a similar way as when testing a UUT with the nitrogen driven EPT’s, with the microphones not in the direct field of the acoustic devices but rather in the more developed reverberant field.

UUT Suspended in Test Chamber

Figure 13 – UUT Suspended in Test Chamber

Microphone Configuration 2 - Drawing

Figure 14 – Microphone Configuration 2 – Drawing

This test was conducted successfully for the full test length. The control was slightly less rapid in response, but the larger number of control microphones led to improved averaging across the entire test spectrum. A typical 148.2dB profile was entered based on previous test requirements which were achieved with the nitrogen driven infrastructure and a single drive channel was utilized to drive all the acoustic devices in the chamber. To simulate a UUT in the chamber, a typical composite flight-panel was suspended in the chamber, as shown in Figure 13 above.

The test ran to a partially successful outcome. From a control standpoint, utilizing a single drive that was controlled by the average of (14) control microphones suspended in the chamber led to a uniform average and positive result from a frequency response standpoint. However, running to such a high level (148.2dB) with only (6) acoustic devices led to the discovery of a few circumstances that would somewhat curtail future max OASPL tests during this study:

1. The power source voltage (120V/208V) was sagging significantly at full output, leading the system to perform less efficiently.

2. The paint in chamber is pitted, adversely affecting high-frequency reverb time, and requiring more HF energy from the ED system.

3. The previously described plywood frame was found to be not very reflective above 4 kHz, further contributing to challenges to HF energy levels in the room.

4. The acoustic devices did not feature an updated high/mid frequency section which would later provide an additional +3dB of output. This caused high frequency limiting at a lower sound pressure level than is typical.

Panel Test - 1/3 Oct Resolution, C1 - C14

Figure 15 – Panel Test – 1/3 Oct Resolution, C1 – C14

Panel Test - 1Hz Resolution PSD, C1 - C14

Figure 16 – Panel Test – 1Hz Resolution PSD, C1 – C14

Test – 1540E Workmanship

To ensure that testing could continue but without running into the challenges noted in the previous 148.2dB test, a lower-level profile was chosen so that the chamber study could continue to characterize the behavior of the room, but without adversely affecting the power supply that was available at the time. The same microphone configuration was used with the 1540E (Perl, 2006) specification was chosen for its lower OASPL (137.84 dB,) broadband excitation, and familiarity to the system operators who frequently use it as a reference for other direct field acoustic testing studies.

This next test ran successfully and without event. Utilizing the single drive output, a slightly larger spread of LF responses was seen in the microphones, as is typical of coherently driven systems in the vertical axis. However, the benefits of testing in the chamber can be clearly seen in the highly diffuse microphone responses overall, especially in the post-processed 1Hz resolution frequency response plot.

1540e Test - 1/3 Oct Resolution, C1 - C14

Figure 17 – 1540e Test – 1/3 Oct Resolution, C1 – C14

1540e Test - 1Hz Resolution PSD, C1 - C14

Figure 18 – 1540e Test – 1Hz Resolution PSD, C1 – C14

Test – ESPA +3 dB

After the successful 1540E test, it was decided to push the limits of the system to help determine the minimum viable number of acoustic devices needed to operate the chamber at a functional level. The test profile chosen was a variation of the Evolved Expendable Launch Vehicle (EELV) Secondary Payload Adaptor (ESPA) recommended profile (Anonymous) with an additional 3 dB for an OASPL of 147.8 dB. Similarly to the Coherent vs. Incoherent test mentioned above, this test experienced power source issues, causing the system to limit in order to protect itself.

ESPA +3 Chamber Results 1/3 octave

Figure 19 – ESPA +3 Chamber Results 1/3 octave

Chamber results Narrowband

Figure 20 – Chamber results Narrowband

High Frequency Attenuation in Chamber

One consistent feature among all the tests run in the chamber was difficulty reaching the targeted reference 1/3Oct OASPL levels above 4-5kHz. To try and understand this phenomenon further, the chamber was excited with a sinusoidal frequency sweep signal generated by Rational Acoustics’ SMAART v8 acoustical analysis software, also called a “pink sweep”. The pink sweep is a logarithmic sinusoidal sweep intended primarily for use in impulse response measurements for room acoustics analysis.

Utilizing a single acoustic device as the output, the chamber was excited by the swept signal and the acoustical response of the chamber was collected as an impulse response. This response was then post- processed to create the included spectrograph and RT60 plots by 1/3Oct, respectively. It was understood that the acoustic device was not a perfectly omnidirectional source with a linear frequency response and was therefore not ideal for this type of excitation. However, given its mostly linear frequency response and predictable characteristics, it was decided that this method would be adequate for a full-range room excitation via sweep.

Impulse Response Spectrograph of the NRL Chamber

Figure 21 – Impulse Response Spectrograph of the NRL Chamber

NRL Chamber – Measured RT60 w/ NEUTRON

The results shown here strongly correlated to the more troublesome frequency bands in the chamber, and the following two conclusions were drawn:

First, the chamber was never designed to offer significant benefit to acoustic excitation above 2kHz, and so signals an octave and higher above that point receive no assistance from the reflective walls of the chamber. The system is effectively driving those frequencies in the direct field, but that energy is quickly dissipating (air loss) or not being well reflected (pitted paint) by the chamber walls.

Second, in order to overcome this potential difficulty driving the chamber to full performance in 1/3Oct bands from 20hz to 10kHz, a hybrid approach might need to be considered where the chamber is excited by the electrodynamic devices broadband, but additional high frequency devices might be placed around the room to ensure the area dedicated to the device under test is receiving adequate coverage in all 1/3Oct bands regardless of whether they are delivered by an electrodynamic system diffused via surface reflections, or diffused via being driven by multiple uncorrelated outputs to achieve the same result.

To test this second point, a brief test was run with more of the hanging chamber-installed control microphones lowered into the direct field of the installed acoustic devices. The test confirmed that the high frequency bands were reaching their target levels in the direct field of the acoustic devices, but only the mid and low frequency bands were benefiting from reflections off the chamber walls. The fact that the test ran more closely to specification in this hybrid direct field + reverberant field configuration and with adequate uniformity and diffusivity added further confidence to the method described in the second point made above.

Conclusions

The results of this study seem to indicate that with adequate chamber coverage, electrodynamic transducers may be used as a direct replacement for EPT’s for lower OASPL levels (148 dB or below) using only facility-supplied electricity. With some modification, the acoustic devices may be usable for levels higher than 150 dB, but this requires further investigation.

Regarding chamber control, using the acoustic control system with 1/3 Octave control provides an easily controlled environment even when the microphones are not directly in the path of the NEUTRON’s, offering “DFAN-like” control to the full test specification but in an RFAN-like environment.

With respect to Coherent vs. Incoherent source operation in a reverberant chamber, it was observed that:

• Low frequencies seemed to benefit in the incoherent configuration but inspecting the monitor microphones reveals that the effect is only in the direct field.

• Incoherent drive required ~30% more energy to achieve the same level

• Incoherent seemed to provide tighter control over the entire spectrum, but, again, this was only found to be true in the direct field. Controlling via the chamber microphones (configuration 2) produced similar results but from coherent output, using less energy

This portion of the study revealed that while in direct field environments the incoherent drive method continues to produce superior results regarding diffusivity and uniformity in both horizontal and vertical axes, utilizing the coherent drive method for the electrodynamic devices in the chamber allowed for significant gains in OASPL output level and system efficiency, while the chamber construction provided adequate diffusion of the single coherent source throughout the entire chamber.

There were some factors that hindered the total output capability of the system for this study. Firstly, the supply power available at the facility was 120V/208V (240V/415V is preferred) and this limited current draw capabilities at higher levels, causing the system to limit output to protect itself. Secondly, the chamber wall surface was pitted and rough, possibly adversely affecting the high-frequency Reverb Time. Thirdly, due to the need to close the gap between the acoustic devices and the door, the installed wooden frame was built from painted plywood, which was non-reflective above 1 kHz and had low- frequency modes around 40-50 Hz. Lastly, the acoustic devices used for this test did not include a newly revised mid/high frequency section, which added roughly 3 dB of overall SPL output.

To mitigate the above factors, 240V power delivery would need to be installed near the amplifier system to avoid power-drops inherent to the 120V supply. Additionally, the chamber walls would require a modest resurfacing along with the inner surface of the acoustic devices being clad in a reflective material such as sheet-metal to further enhance the reverberation time of the room above 4 kHz. A hybrid acoustic device solution could also be employed by using a smaller horn driven electrodynamic device made for only high frequencies, the chamber corners could be utilized to achieve levels above If the above factors are mitigated, the proposed system of (10) acoustic devices in an on-chamber configuration are projected to reach roughly 150 dB OASPL.

References

Anonymous. (n.d.). Secondary Payload Planner’s Guide For Use On The EELV Secondary Payload Adapter. DoD Space Test Program.

Hayes, D. (2015). Reverberant Chamber Enhancement. Journal of the IEST.

Kolaini, A., Doty, B., & Chang, Z. (2012). Impact of Acoustic Standing Waves on Structural Responses: Reverberant Acoustic Testing (RAT) vs. Direct Field Acoustic Testing (DFAT). Pasadena, CA 91109- 8099 : Jet Propulsion Laboratory, California Institute of Technology.

Perl, E. (2006, September 6). Test Requirements for Launch, Upper-stage, and space Vehicles. MIL-STD- 1540E. El Segundo, California, 90246-2808: Military Standard.

About the Authors

Dale Schick is Applications Manager of m+p International Inc. and leads US research projects related to Acoustic and Vibration testing. He has extensive experience in environmental testing as applied to Aerospace and over 15 years of practical experience in the industry.

Thomas Hoffmann is product manager of m+p international and coordinator of the Direct Field Acoustic Noise Testing solution. Before he joined m+p international he worked as a researcher at Institute of Dynamics and Vibration Research in Hannover, Germany. There he received his PhD degree on experimental nonlinear vibration and simulation algorithms.

Eric Friedlander is the CTO and Co-Founder of Acoustic Research Systems. He has a BS in Audio Technology from American University in Washington, DC, and over two decades of practical experience in sound engineering for musical, industrial, and scientific applications. Prior to founding ARS, Eric was a product manager at Harman Professional for the JBL Professional Tour Sound portfolio, and prior to that he worked as an acoustic testing engineer at MSI-DFAT.

Read Full Article

Explore More

╱╲