improved low frequency performance of a...
TRANSCRIPT
1AGU Spring ‘98
Improved Low Frequency Performance of
a Geophone
S32A-19
2AGU Spring ‘98
Aaron Barzilai1, Tom VanZandt2,
Tom Pike2, Steve Manion2,
Tom Kenny1
1Dept. of Mechanical EngineeringStanford University
2Center for Space Microelectronics TechnologyJet Propulsion Laboratory
Contact: [email protected]
3AGU Spring ‘98
Acknowledgements
This work was supported by the Center for Space Microelectronics Technology, Jet Propulsion Laboratory, California Institute of Technology, and is sponsored by the NASA Office of Space Access and Technology. We also acknowledge the NSF Career Award(ECS-9502046), the Charles Lee Powell Foundation, and the Terman Fellowship.
Special thanks to Marcos Alvarez at PASSCAL for his advice and assistance.
4AGU Spring ‘98
Poster Map
Title Introduction
Conventional Geophone
Capacitive Geophone
Conclusions
5AGU Spring ‘98
Objective
• Develop an Affordable, Robust Broadband Seismometer with Resolution Comparable to the Earth’s Seismic Noise
• Enhance the Ability of Seismometer Arrays to Detect Low Frequency Signals
6AGU Spring ‘98
Seismometer Information Flow
AccelerationInput Mechanical
System
RelativeMotion Electrical
System
OutputVoltage
7AGU Spring ‘98
Cross-sectionSchematic
A Conventional Geophone:OYO Geospace 4.5 Hz GS-11D
Magnet
Cylinder
Coil
GeophoneHousing
LeafSpring
8AGU Spring ‘98
Mechanical Sensitivity
• Acceleration Causes Relative Motion Between the Coil and the Housing
• Constant Sensitivity Below the Resonant Frequency
10-6
10-5
10-4
10-3
10-2
10-2 10-1 100 101 102
Mec
hani
cal S
ensi
tivity
[m/(
m/s
^2)]
Frequency [Hz]
9AGU Spring ‘98
Electrical Sensitivity
• Inductively Measure Motion of the Coil Relative to the Magnetic Field
• Output Voltage Proportional to the Proof Mass Velocity
102
103
104
105
106
107
10-2 10-1 100 101 102
Ele
ctric
al S
ensi
tivity
[V/m
]
Frequency [Hz]
10AGU Spring ‘98
Total Sensitivity
• At Low Frequency, Measurement of Proof Mass Velocity Reduces Sensitivity
• At High Frequency, Mechanical System Reduces Sensitivity
100
101
102
103
104
10-2 10-1 100 101 102
Geo
phon
e S
ensi
tivity
[V/g
]
Frequency [Hz]
11AGU Spring ‘98
Circuitry NoiseOutput Voltage
Noise Spectral Density
Geo100Ω
10kΩ
VO
Typical Circuit
100x Ampifier10-7
10-6
10-5
10-4
10-2 10-1 100 101 102
Out
put V
olta
ge N
SD
[V/√
Hz]
Frequency [Hz]
Peak CausedBy Resonance
12AGU Spring ‘98
Conventional Geophone Resolution
• Poorer Resolution at Low Frequency caused by Reduced Sensitivity
• Resolution worse than Fundamental Limit
Resolution gHz[ ] =
Noise VHz[ ]
Sensitivity Vg[ ]
Reference: Barzilai et al., “Technique for Measurement of the Noise of a Sensor in the Presence of Large Background Signals,” Rev. Sci. Instrum., Accepted for Publication
10-10
10-9
10-8
10-7
10-6
10-5
10-4
10-2 10-1 100 101 102
Res
olut
ion
[g/√
Hz]
Frequency [Hz]
Thermomechanical Limit
PredictedCircuitry Limit
MeasuredCircuitry Limit
13AGU Spring ‘98
An Improved Seismometer: A Capacitive Geophone
• Use a Commercial, Off The Shelf Geophone as the Mechanical System
• Improve Low Frequency Sensitivity by Capacitively Measuring Proof Mass Displacement with only Simple, External Modifications
14AGU Spring ‘98
Photos
Guralp CMG-40T and Capacitive Geophone
Capacitive Geophone
15AGU Spring ‘98
Capacitive HardwareAdditionalHousing
MovingElectrode
FixedElectrodes
Insulation
33.37 mm39.37 mm
y
CircuitModel
a = Balanced Gap ≈ 250µm
A = Area = 3.4 ×10-4 m2
CNOMINAL = 12.1pF
C =εε0 A
a − y
C =εε0 A
a + y
16AGU Spring ‘98
Electrical System Overview
VBR is a Sine Wave at the Same Frequency as VSIN with Amplitude Modulated by y. The Lock-In AmplifierDemodulates the Signal to Produce an Output that isProportional to the Displacement of the Proof Mass.
y
+VSIN
x11
-VSIN
Lock-InAmplifier
VOUTVBR
Reference
InputVIN
17AGU Spring ‘98
Preamplifier Circuit
fSIN = 100kHz
VSIN = 3.8Vpk
fc = 12πRC
≈ 600Hz
10 M
1 k
10 k
ModulatedSignal
VSIN
VBR
-VSIN
VINCA
CB TLC 2274
18AGU Spring ‘98
Preamplifier Circuit Sensitivity
VIN = VSIN − −VSIN[ ]( )1
sCA1
sCA+ 1
sCB
− VSIN
1+10kΩ1kΩ
VIN = VSINsCB − sCA
sCB + sCA
11( )
VIN = VSIN
1a −y − 1
a+y1
a −y + 1a+y
11( )
VIN = VSINy
a11( ) = 1.7 ×105 V
m
y m[ ]
Exact Solution: Voltage Amplitude is a Linear Function of Displacement
19AGU Spring ‘98
Demodulator
• Demodulator Output is Directly Proportional to the Amplitude of the Input at the Reference Frequency.
• In the Frequency Domain, Low Frequency Amplitude Variations Appear as Signals Near the Reference Frequency. The Demodulator Shifts These Signals Back to Low Frequency.
20AGU Spring ‘98
Electrical Sensitivity
Demodulator Sensitivity
VOUT =10V
5 10−3( )V1
2VIN pk = 1414 VIN pk
VSIN
SRS 510Lock-In
Amplifier
Reference
Input
Output
VIN
VOUT = 1414( )1.7 ×105 V
m
y m[ ] = 2.3 108( ) V
m
y m[ ]
21AGU Spring ‘98
Total Sensitivity
• Constant Sensitivity At Low Frequency since Output is Proportional to Proof Mass Displacement
100
101
102
103
104
105
106
107
10-2 10-1 100 101 102
Geo
phon
e S
ensi
tivity
[V/g
]
Frequency [Hz]
Capacitive
Conventional
22AGU Spring ‘98
Preamplifier Circuitry Noise
• Noise is Constant Amplitude vs. Frequency
• Lock-In Amplification Produces .18 mV/√Hz Output Voltage NSD at Low Frequencies
Noise Near the Reference Frequency is Mapped to Low Frequencies by Demodulator
10-7
10-6
99.9 100 100.1
Pre
amp
Out
put N
SD
[V/√
Hz]
Frequency [kHz]
ReferenceFrequency
23AGU Spring ‘98
Capacitive Geophone Resolution
Resolution gHz[ ] =
Noise VHz[ ]
Sensitivity Vg[ ]
• At Low Frequencies, Resolution is Limited by Thermomechanical Noise, not Circuitry
• Better Resolution at Low Frequencies as a Result of Constant Sensitivity
10-11
10-10
10-9
10-8
10-7
10-2 10-1 100 101 102
Res
olut
ion
[g/√
Hz]
Frequency [Hz]
Thermomechanical Limit
Circuitry Limit
24AGU Spring ‘98
Clip Level
• Demodulator Output Range ±10V
• Corresponds to .040 µm Displacement
10-6
10-5
10-4
10-3
10-2
10-2 10-1 100 101 102
Clip
Lev
el [g
]
Frequency [Hz]
25AGU Spring ‘98
Resolution Comparison
USGS LNM:Seismic Noise At Quietest Sites On Earth
10-11
10-10
10-9
10-8
10-7
10-6
10-2 10-1 100 101 102
Res
olut
ion
[g/√
Hz]
Frequency [Hz]
Thermomechanical Limit
Circuitry Limit
ConventionalGeophone
USGS LowNoise Model
CMG-40T
26AGU Spring ‘98
Resolution Comparison
• Low Frequency Resolution of a Geophone is Improved by using Capacitive Detection
• Circuitry Noise does not Limit Capacitive Geophone at Low Frequency
• Limits on Resolution of a Capacitive Geophone are Better than the Resolution of a Guralp CMG-40T Broadband Seismometer
27AGU Spring ‘98
Performance ComparisonConventional
Geophone
100 ng/√Hz
90 mg
120 dB
$50
Resolution
Clip Level
DynamicRange
Estimated Cost
CapacitiveGeophone
.1 ng/√Hz
5 µg
90 dB
$500
GuralpCMG-40T
.5 ng/√Hz
1 mg
130 dB
$10,000
28AGU Spring ‘98
Conclusions
• The Low Frequency Resolution of a Geophone can be Improved by Adding Capacitive Detection
• Capacitive Detection Does Not Improve High Frequency Resolution
• Thermomechanical Noise Sets a Resolution Limit ≈ 0.1 ng/√Hz on all Geophone Based Seismometers
29AGU Spring ‘98
Future Work
• Experimentally Validate the Predicted Resolution of a Capacitve Geophone
• Reduce Size by Integrating Electronics onto a Single Printed Circuit Board
• Operate as a Closed Loop Sensor to Increase the Dynamic Range and Tune the Frequency Response