In practical engineering application, the control system must be low-cost, simple, efficient, and reliable. The system may also have to be standalone, light and small, especially in aircraft engineering. In this section, the experiments of adaptive noise and vibration control are conducted by the C31 standalone DSP controller and the associated circuitry. The power source only needs one 9V battery in noise control and two 9V batteries in vibration control.
Figure 5.1 shows the experimental setup of the noise control headset. The noise signal is detected by a condenser microphone embedded in the headset and then processed by a low-pass pre-amplifier to pass through the controller. A power amplifier as shown in Fig. 5.1(a) is need to drive the speaker in the headset because of impedance match. The controller is an FIR adaptive filter with 200 weights updated by the filtered-X LMS algorithm. The convergence rate is set at 10−12 by trial-and-error with the sampling rate of 4.88 KHz. The dimension of the secondary path dynamics model is 52 with 22 time delay samples. The pure tone noise of 200 Hz is attenuated about 8 dB from 71 dB to 63 dB measured by a sound level meter (ONO SOKKI LA-220). Figure 5.2(a) shows the time response measured by the condenser microphone embedded in the headset. The sensor causes some distortion, but the controller remains effective. The power spectrum is shown in Fig. 5.2(b) where the peak of 200 Hz is attenuated. The noise control headset is also effective in multi-tone noise control as shown in Fig. 5.3. The noise attenuation of 200 Hz ~ 800 Hz multi-tone is about 4 dB, from 70 dB to 66 dB.
The experiment setup of vibration suppression by the DSP controller is shown in Fig. 5.4. Because the piezoelectric sensor is a strain rate sensor, it can be seen as similar to a velocity sensor. In order to generate the displacement sensing signal, the piezoelectric sensing signal is processed by an integrator circuit as shown in Fig. 5.4(a)
and then passes through the controller to generate the piezoelectric actuator input.
The actuator driving circuit is shown in Fig. 5.4(b). The FIR adaptive controller has 56 weights with sampling rate of 2 KHz and convergence rate of 10-11. The ARX model is of 38 autoregressive parameters and 39 moving average parameters with 13 time delay samples. Figure 5.5 shows the free vibration response of the composite beam under an initial displacement. Vibration suppression is significant after one circle of oscillation. The peak at 60 Hz is the measurement noise. Vibration control is also conducted under harmonic excitation as shown in Fig. 5.6. The closed-loop response can also be attenuated after two circles and completely canceled in 2.2 seconds.
The experimental results show this DSP controller development shall significantly facilitate practical engineering applications in active noise and vibration control.
A low cost and efficient standalone DSP controller by using TMS320C3x DSK with only an 8-bit external boot EPROM module is developed. A C31 DSP system is also developed to provide an example of how to develop a low cost and efficient standalone DSP controller for active noise and vibration control. This technology facilitates the transition from laboratory experiments to engineering applications. Some implementation parameters, such as the sampling rate, the dimension of the control filter, the dimension of the secondary path model and the convergence rate, can be effectively obtained and validated by experimental design.
6. Summary
In this three-years project, the technology of practical adaptive noise and vibration
control has been developed and validated by using C3x digital signal processor in
duct and earphone noise experiments and in beam and plate smart structures vibration
experiments. The summaries of this work are listed as following:
1. Conventional active noise and vibration control systems utilize several sensors: at
least one to measure the reference input (noise field or disturbance source) and
another one to measure the attenuated noise or vibration signal. A simple and
yet effective configuration by using only one sensor is developed for adaptive
noise and vibration control. The adaptive controller implemented on a C31 DSP
is developed for active noise and vibration control.
2. In acoustical systems, system identification of the secondary path dynamics by
using either ARX model or FIR model is in very good agreement with the
measured data. In structural systems, identification results indicate that fewer
parameters are needed in the ARX model due to the sharp frequency response
(small damping), especially in beam system. The identification facilitates the
FIR adaptive controller with the filtered-X algorithm to control acoustics noise
and structural vibration effectively.
3. In two-sensor adaptive noise control system, both simulation and experimental
results show that the adaptive controller can attenuate not only the pure tone
noise but also the stationary or nonstationary narrow band noise. However, the
problem of acoustic feedback remains unsolved. In one-sensor adaptive noise
control system, both simulation and experimental results show that the adaptive
controller can attenuate the pure tone noise and the narrow band noise.
However, it is less effective to nonstationary noise.
4. A headset design with a pair of speaker and microphone, a pre-amplifier circuit,
and the filtered-X LMS algorithm is developed. The headset is effective in
attenuating the steady state low-frequency noises. The result shows that the
feedback active noise control by using the feedforward adaptive filter
implemented on a digital signal processor is effective.
5. The results of vibration control experiments show that adaptive controller by
using filtered-X LMS control algorithm can reduce the structure vibration
efficiently. It works both in free vibration or forced vibration. The control
performance of the adaptive controller remains well under system variations.
6. Since the concept of active noise and vibration control was conceived in early
days, the technology has been in transition from a dream to implementation and
from a laboratory experiment to engineering application. An effective and
efficient standalone DSP controller by using TMS320C3x DSK with an 8-bit
external boot EPROM is developed for both active noise and vibration control.
Some implementation parameters, such as the sampling rate, the dimension of the
control filter, and the dimension of the secondary path model, can be determined
by the experimental design method. Engineering implementation of adaptive
controller is becoming a reality.
References
1. Berengier, M., and Roure, A., 1980, “Broadband Active Sound Absorption in a Duct Carrying Uniformly Flowing Fluid,” J. Sound Vib., vol. 68, pp. 437-449.
2. Burgess, J. C., 1981, “Active Adaptive Sound Control in a Duct: A Computer Simulation,” J. Acoust. Soc. Am., vol. 70, no. 3, pp. 715-726.
3. Bai, M. R., Shieh, C., and Lee, D., 1995, “Development of a LQG Active Noise Canceler,” ACTIVE 95, Newport Beach, CA, USA, pp. 911-920.
4. Bai, M. R., and Chang, S., 1996, “Active Noise Control of Enclosed Harmonic Fields by Using BEM-Based Optimization Techniques,” Appl. Acoust., vol. 48, pp. 15-32.
5. Bai, M., and Chen, H., 1996, “A Modified H2 Feedforward Active Control System for Suppressing Broadband Random and Transient Noises,” J. Sound Vib., vol. 198, pp. 81-94.
6. Bai, M. R., and Chen, H-P., 1997, “Development of a Feedforward Active Noise Control System by Using the H2 and H∞ Model Matching Principle,” J.
Sound Vib., vol. 201, pp. 189-204.
7. Carme, C., 1988, “A New Filtering Method by Feedback for ANC at the ear,”
Proc. Inter-noise, pp. 1083-1086.
8. Gauger, D. and Sapiejewski, R., 1987, “Voyager Pilots Avoid Hearing Loss on Historic Flight,” Sound and Vibration, pp. 10-12, May.
9. Hu, J. S., 1995, ”Active Sound Attenuation in Finite-Length Ducts Using Close-Form Transfer Function Models,” J. of Dynamics Systems, Measurement, and Control, vol. 117, pp. 143-154.
10. Jessel, M., and Mangiante, G., 1972, “Active Sound Absorbers in an Air Duct,” J.
Sound Vib., vol. 23, pp. 383-390.
11. Kuo, S. M. and Morgan, D. R., 1996, “Active Noise Control Systems,” Wiley
Interscience, New York.
12. LaFontaine, R. F., and Shepherd, J. C., 1983, “An Experimental Study of Broadband Active Attenuator,” J. Sound Vib., vol. 91, pp. 351-362.
13. Lueg, P., 1936, “Process of Silencing Sound Oscillation”, U.S. Patent 2043416.
14. Maxwell, D. W., Williams, C. E., Robertson, R. M., and Thomas, G. B., 1987,
“Performance Characteristics of Active Hearing Protection Devices,” Sound Vib., vol. 21, no. 5, pp. 14-18.
15. Nelson, P. A. and Elliott, S. J., 1993, “Active Control of Sound,” Acadmic Press.
16. Okamoto, Y., Boden, H., and Abom, M., 1994, “Active Noise Control in Duct via Side-Branch Resonators,” J. Acoust. Soc. Am., vol. 96, no. 3, pp. 1533-1538.
17. Olson H. F., 1956, “Electronic Control of Noise, Vibration, and Reverberation,”
Journal of Acoust. Soc. Am., vol. 28, pp. 966-972.
18. Oppenheim, A. V., Weinstein, E., Zangi, K. C., Feder, M., and Gauger, D., 1994,
“Single-Sensor Active Noise Cancellation,” IEEE Trans. on SAP, vol. 2, no. 2, pp.
285-290.
19. Poole, J. H. B., and Leventhall, H. G., 1976, “An Experimental Study of Swinbanks’ Method of Active Attenuation of Sound in Ducts,” J. Sound Vib., vol.
49, pp. 257-266.
20. Rafaely, B. and Furst, M., 1996, “Audiometric Ear Cannal Probe with Active Ambient Noise Control,” IEEE Trans. on SAP, vol. 4, no. 3, pp. 224-230.
21. Ross, C. F., 1982, “An Adaptive Digital Filter for Broadband Active Sound Control,” J. Sound Vib., vol. 80, pp. 381-388.
22. Roure, A., 1985, “Self-Adaptive Broadband Active Sound Control System,” J.
Sound Vib., vol. 101, no. 3, pp. 429-441.
23. Salloway, A.J. and Twiney, R.C. , 1985, “Earphone Active Noise Reduction Systems ,” Proceedings of the Institute of Acoustics 7, pp. 95.
24. Swinbanks, M. A., 1973, “The Active Control of Sound Propagating in Long Ducts,” J. Sound Vib., vol. 27, pp. 411-436.
25. Stevens, J. C., and Ahuja, K. K., 1991, “Recent Advances in Active Noise Control,” AIAA Journal, vol. 29, no. 7, pp. 1058-1067.
26. Wheeler, P. D., 1986, “Voice Communications in The Cockpit Noise Environment - The Role of Active Noise Reduction,” Ph.D. Thesis, University of Southampton, England.
27. Widrow, B., 1970, “Adaptive Filters,” in Aspects of Network and System Theory, R. E. Kalman and N. De Claris (Eds.). New York: Holt, Rinehart and Winston, pp.
563-587.
Figure 2.1 Adaptive Noise Control System in Two-Sensor Configuration.
Figure 2.2 Adaptive Vibration Control System in Two-Sensor Configuration.
Figure 2.3 Adaptive Noise and Vibration Control Considering the Transfer Function S(z) of the Secondary Path Dynamics.
Figure 2.4 Block Diagram of the Filtered-X LMS Algorithm with Two sensors for SISO Adaptive Noise and Vibration Control.
Figure 2.5 MIMO Adaptive Feedforward Control System.
Figure 2.6 Block Diagram of the Adaptive Noise and Vibration Control System in One-sensor Configuration.
Figure 2.7 A Dynamic System with Input, Output and Disturbance.
Model Dimension
Mean Square Error
n1 n2 L
L2 1
Figure 2.18 Model Error versus Model Dimension in System Identification.
A18
Figure 4.1 Standalone DSP Controller Design by Using TMS320C31 DSK with External Boot EPROM.
(a)
(b)
Figure 4.2 Mass Productions of Standalone DSP Controller Design, (a) Active Noise Control Headset, (b) Active Vibration Control of Smart Structure.
+V
out
+
C4
R1
MK1
MICROPHONE 1 2
R2
C1 D1
13
2 C6
C7
C8 C2
C9 C5
C11 C10
C3
R9 R4
R6 R5
R10
R12
+
-U2 3 2
6 7 14 5
+
-U3 3 2
6 7 14 5 +
-U1 3 2
6 7 14 5
13
2
R8 R7
preamplifier lowpass filter
Figure 4.3 Circuit of Measuring Sound Wave by a Microphone, a Pre-amplifier and a Low-pass Filter.
VCC1
D4 D8D5 D18
D13D12
VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD VDD
VSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSSVSS
(continued on next page) Figure 4.4 Schematic Diagram of the TMS320C31 DSP System,
(a) CPU and Power Supply; (b) SRAM, EPROM, and DSK Interface; (c) A/D, D/A, Amplifier, and Serial Port Interface.
GND GND GND
(continued on next page) Figure 4.4 (continued)
+5V Figure 4.4 (continued)
(a)
(b)
(continued on next page) Figure4.5 The Circuit Layout of the C31 DSP System Board, (a)
Component Side Layout; (b) Solder Side Layout; (c) Silkscreen and (d) Hardware.
(c)
(d) Figure 4.5 (continued)
VCC
C3 C2 C1
R3 R2 R1
+
-U1 3 2
5 6 14 8 7
R4
13
2
C4
LS1 J1
1
2
(a)
(b)
Figure 5.1 Experiment of Noise Control Headset, (a) Speaker Power Amplifier and (b) Experimental Setup.
0.00 0.02 0.04 0.06 0.08 0.10 Time (sec)
-0.8 -0.4 0.0 0.4 0.8
Error Microphone Output (V)
Open-loop Closed-loop
(a)
10 100 1000
Frequency (Hz) -200
-150 -100 -50 0 50
Power Spectrum (dB)
Open-loop Closed-loop
(b)
Figure 5.2 Experiment of the Noise Control Headset under 200Hz Pure Tone Noise, (a) Time Response and (b) Power Spectrum.
0.00 0.02 0.04 0.06 0.08 0.10 Time (sec)
-0.8 -0.4 0.0 0.4 0.8
Error Microphone Output (V) Open-loop
Closed-loop
(a)
10 100 1000
Frequency (Hz) -200
-160 -120 -80 -40 0 40
Power Spectrum (dB)
Open-loop Closed-loop
(b)
Figure 5.3 Experiment of the Noise Control Headset under Multi-Tone Noise, (a) Time Response and (b) Power Spectrum.
VCC+
VCC-GND Input
Output +
-U1 6
1 5
2 3
74
R1 R2
R3
R4 C1
R513 2
(a)
VCC
VEE 0
0
0 VCC
0
0 0
0 0
0 VEE
C1
C4 R7 R2
R1
U1A
3 2
84
1 + - V+
V-OUT
C3 R6 V3
R9
R8 C5 R3
R5 U1B
5 6
84
7 + - V+ V-C2 OUT
R4
(b)
(c)
Figure 5.4 Vibration Control Experiment of a Composite Beam, (a) Circuit of Processing Piezoelectric Sensor, (b) Circuit of Driving Piezoelectric Actuator and (c) Experimental Setup.
0 1 2 3 Time (sec)
-2.0 -1.0 0.0 1.0 2.0
Output (V)
Open-loop Closed-loop
(a)
0 20 40 60 80
Frequency (Hz) -150
-100 -50 0 50
Power Spectrum (dB) Open-loop
Closed-loop
(b)
Figure 5.5 Free Vibration Experiment of the Composite Beam, (a) Time Response and (b) Power Spectrum.
0 1 2 3 4 5
Time (sec) -0.8
-0.4 0.0 0.4 0.8
Output (V)
Open-loop Closed-loop
(a)
0 20 40 60 80
Frequency (Hz) -200
-150 -100 -50 0 50 100
Power Spectrum (dB) Open-loop
Closed-loop
(b)
Figure 5.6 Forced Vibration Experiment of the Composite Beam, (a) Time Response and (b) Power Spectrum.