Instrumentation Testing


The BSL Instrumentation Testing Facility located in the Byerly Vault (BKS) has been busy this past year with the testing of several sensors. In July 2007, the sensitivity and noise performance of the 8-channel Quanterra Q4120 data logger was checked to verify that is operating within the factory specifications. The BSL staff has also been involved in projects to test new STS-1 electronics developed by Metrozet LLC and to test a new temperature, hygrometer, and pressure sensor package that is being developed in-house. We have also tested new and repaired broadband and strong motion sensors on an ad hoc basis prior to deployment in the field.

A new advanced electronics package, the Metrozet STS-1-E300, has been developed and tested as a modern replacement for the original Streckeisen STS-1 very-broadband seismometer ``Feedback Electronics'' boxes. The development and testing is a collaborative effort between Metrozet LLC and the BSL. The new electronics package matches the outstanding analog performance of the original Streckeisen circuitry while providing a number of enhancements to facilitate installation and operation of the STS-1 sensors. The enhancements include: digital control of all sensor parameters; digital control of centering motor operations; monitoring of all major state-of-health parameters, and a complete calibration capability. All control and diagnostic functions can be controlled either locally (via RS-232, USB or Ethernet) or remotely (via Ethernet).

A new temperature, humidity, and pressure (THP) sensor package is being developed in-house at BSL as a replacement for the existing and aging temperature and pressure sensors at the BDSN stations. Measurement of the temperature and pressure at a BDSN station is useful for reducing the components of the seismic background noise that are correlated with temperature and pressure. A hygrometer has also been added to the sensor package to enable measurement of the local atmospheric relative humidity, a parameter which is potentially useful for estimating and correcting for geodetic GPS tropospheric propagation delays.

Instrumentation Test Facility

The BSL sensor testing facility is described in detail in the BSL 2001-2002 Annual Report (available on-line at

Data Logger Calibration

In July 2007, the sensitivity and self-noise PSD for each data channel and inter-channel crosstalk of the 8 channel Quanterra Q4120 data logger was checked using a reference signal applied first simultaneously and then sequentially to all channels, with the non-driven channels resistively terminated. The relative sensitivities of the data logger channels were checked by applying a high-level ($\sim$19.8 V peak-to-peak (P-P)) 1 Hz square wave signal simultaneously to all channels. The signal level on each channel was measured and the relative signal levels were compared to the sensitivities on the factory calibration sheet. The sensitivities of four of the channels have not changed by more than 0.01% from the factory calibration values. Of the remaining four channels, three changed by less than $\pm{0.3}$%, and the fourth changed by -0.8%. Modulo 0.25% sensitivity changes are inherent in Q4120 at boot time because it steps the sensitivity in 0.25% increments and selects the sensitivity that has the lowest internal noise level. Sensitivity changes that are not near modulo 0.25% are likely due to a combination of modulo 0.25% increments and degradation of the components.

The self-noise of the Q4120 channels (with 1 k$\Omega$ resistance termination) was determined via Power Spectral Density (PSD) analysis. A composite plot of the Q4120 self-noise over the 20 microHz to 0.5 Hz band is shown in Figure 3.30, and a more detailed plot of the high-frequency (0.2-80 Hz self-noise PSD is shown in Figure 3.31. The minimum observed self-noise PSD is approximately 6 dB below the factory specification for a thermally stable environment in the 2-20 Hz band.

The inter-channel cross-talk of the data logger was checked by connecting each of the channels in sequence with the high amplitude (20 V P-P) 1 Hz square-wave signal while terminating the other seven channels with 1 k$\Omega$ resistors at the data logger input connectors. The Q4120 data logger contains two 4-channel digitizer modules (HH1-HH4 and HH5-HH8). The observed cross-talk signal level on all 8 channels is below the 2.34 $\mu$V quantization (least significant bit or LSB) level of the Q4120 data logger. The cross-talk signal level is thus more than 138.5 dB below the drive signal level. A check of the coherence between the channels was performed by driving each channel sequentially with a one minute duration 20V P-P 1 Hz square wave while terminating the remaining seven channels with 1 k$\Omega$ resistors across the signal input connectors of the data logger. Spectral phase coherence analysis of the signal between the inter-channel pair combinations did not detect any significant coherence.

Figure 3.30: Self-Noise PSD of Q4120 data logger used in testing. A 1 k$\Omega$ resistance termination was placed across the data logger input. The plot is composite, derived from 1 sps noise data for frequencies below $\sim{0.01}$ Hz and 200 sps noise data at higher frequencies.
...}\epsfig{file=rau08_1_fig_1.eps, width=8cm, bb=72 270 540 540, clip}\end{figure}

Figure 3.31: High-frequency self-noise PSD of Q4120 data logger used in testing. The data logger input was terminated with a 1 k$\Omega$ resistor. The median value of 32 noise sample PSD data are plotted, and the lowest self-noise PSD of $\sim$-137 dB is observed in the 2-30 Hz band. For comparison, the factory noise specification is that the terminated input noise level is typically -134.5 dB and it may exceed -137.5 dB at constant temperature.
...}\epsfig{file=rau08_1_fig_2.eps, width=8cm, bb=72 270 540 540, clip}\end{figure}

Instrumentation Test Bed

The data logger is connected to the seismometer test bed breakout box via a $\sim$6 meter shielded signal cable. The above coherence test was repeated with the drive signal and terminating resistances connected at the breakout box, and some inter-channel pairs exhibited detectable phase coherence at 1 Hz and its odd harmonics and also at the 60 Hz power mains frequency, between adjacent signal pairs in the cable, which is consistent with capacitive coupling in the signal cable and with proximity of the signal cable to 60 Hz power wires in the vault. The largest observed cross-talk signal levels on the terminated channels were $\sim$120 dB below the 20V P-P 1 Hz square wave drive signal level which is sufficiently low that the cross-talk does not interfere with subsequent data analysis.

BDSN Ad Hoc Sensor Testing

During the past year, the following sensors were tested to verify that they were performing within the factory specifications prior to re-deployment in the field:

1) The 3 STS-1's from KCC (STS-1Z s/n 109112, STS-1H s/n 29212, and STS-1H s/n 29201) with Metrozet STS-1-E300-005 electronics,

2) The 3 STS-1's from SAO with their corresponding factory electronics boxes (STS-1Z s/n 109119, STS-1H s/n 48528, and STS-1H s/n 48529).

STS-1 Sensor Electronics Testing

From the perspective of the BSL, the critical task in developing new electronics for the STS-1 seismometers, the E300, is the evaluation of performance and the comparison between new and old systems. Several iterations of the E300 were evaluated during the design and development phase. Objective evaluation of the new electronic subsystem required a stable and repeatable test platform of seismometers, base plates, cables, connectors, and digitizer channels. Only when the system was stable, as evidenced by repeated calibration results, was the platform suitable for evaluation of the new electronics.

Seismometer Acquisition and Alignment

Having characterized and calibrated the data logger (as discussed above), it was necessary to establish a stable seismometer subsystem test bed. Nine different STS-1 instruments (six horizontal, and three vertical instruments) were set up and leveled, and the outputs to the Q4120 data logger were compared quantitatively.

The horizontal instruments were aligned along a single axis allowing comparisons and evaluations of their coherence. Misalignment of less than one degree across the six horizontal instruments caused unacceptable variances in signal coherence, and took a week to resolve by rotating individual instruments. Alignment of the vertical instruments was much easier. Only after all alignment incoherences were resolved, were the new electronics evaluated against the original factory electronics.

Each cable, connector, and base plates combination were marked, color coded, and assigned, as new and original electronics were mated with seismometers. The color coding endeavored to eliminate ambiguities and variables beyond the actual electronics. Each combination of seismometers and electronics was recorded for a minimum of 24 hours. In the end, over 100 combinations of original and new electronics and seismometers were evaluated under conditions that were documented.

When individually labeled and identified combinations of base plates, cables, seismometers, and electronics were characterized and deemed repeatable, the new Metrozet electronics were substituted. Initially, the prototype Metrozet electronics were operated on only one seismometer, with three, four or five other seismometers simultaneously operating on Streckeisen's factory electronics. After verifying correct input and output signal levels, the single Metrozet electronics were rotated and verified amongst six seismometers. A second prototype Metrozet electronics package was likewise rotated through the seismometers, but this time powering two seismometers concurrently.

Sensor/Electronics Testing

Testing of the STS-1 sensor and electronics systems involved several components. STS-1 vertical- and horizontal-component sensors for testing were gathered from among the available BDSN sensors, from surplus sensors on loan from the Gräfenberg Array, and from UC San Diego. The sensors were systematically inspected and checked to ensure that they were operating correctly. The ones that performed best were selected for the testing procedures and installed in the test bed. In total, 16 broadband STS-1 sensors (7 vertical and 9 horizontal components) were utilized during the testing process. Nine of them are owned by the BSL; 5 were on loan from Gräfenberg and 2 from IGPP/UCSD. The horizontal-component sensors were aligned east.

...epsfig{file=rau08_1_fig_4.eps, width=12cm, bb=40 180 580 630, clip}\end{figure*}
Figure 3.32: Results for two vertical component STS-1's (HHZ and HH2) and two horizontal component STS-1's (HHE and HH1) in the presence of a large seismic signal. The event is a $M_{w}$ 8.1 earthquake which occurred 87.9 degrees WSW of Berkeley on 2007/04/01 at 20:39 UT. Shown are the signal PSD (red), the noise PSD (blue), and the coherence (brown) for each sensor.

...epsfig{file=rau08_1_fig_4.eps, width=12cm, bb=40 180 580 630, clip}\end{figure*}
Figure 3.33: Results for four vertical component STS-1's (HHZ, HH4, HH5, and HH7) in the presence of background noise. The traces are the same as in Figure 3.32. The lower and upper frequencies at which the coherence degrades from near unity varies among the sensors. Coherence bandwidth is a measure of the performance of the sensors, and HH5 has the best performance of the four sensors.

Coherence and Power Spectral Density

An algorithm ( $\it {scn\_psd}$) to calculate the signal Power Spectral Density (PSD), the noise PSD, and the coherence between sensors has been developed, in-house at BSL, as a tool for quantifying the performance differences between the seismic sensors under test. Three continuous hours of 200 Hz data are used by the $\it {scn\_psd}$ algorithm. $\it {scn\_psd}$ parses the data into 32 non-overlapping segments, applies and corrects for the effects of a Hanning window, scales the data to ground motion, calculates the Fast Fourier Transform (FFT) and stores the resulting complex spectral values for each segment. At each frequency, the RMS signal PSD is calculated from the average of the complex spectral values, coherence is calculated from the averaged complex spectral cross product, and the RMS noise PSD is then determined from the product of the signal PSD and (1 - coherence). The method is described in detail in Barzilai et al. (1992). In all tests, the BKS STS-1's are used as the reference signals in the analysis. Two sample results of the algorithm are shown. Figure 3.32 shows the results for four seismometers in the case of a large seismic event, and Figure 3.33 shows the results for four seismometers in the case of background noise. In the presence of large seismic signals, the coherence is typically close to unity at all frequencies below the 5 Hz high-frequency corner of the BKS reference STS-1's. Note the relatively high noise PSD level on the horizontal components in the vicinity of 0.1 Hz. This is due to a slight misalignment of the sensitive axes of the horizontal components. Several time-consuming trial and error iterations in aligning the horizontal components are required to lower the horizontal component noise PSD.

Response Calibration

A pair of algorithms were developed in-house to determine the seismometer frequency response as characterized by the seismometer free period ($T_{s}$) and fraction of critical damping ($h_{s}$) and the high-frequency (``galvanometer'') corner frequency ($f_{g}$) and fraction of critical damping ($h_{g}$). These algorithms were developed to take advantage of the step function and low-frequency swept sine and high-frequency stepped sine calibration stimuli generated by the E300 electronics boxes, applied to the STS-1 calibration coils, and also recorded on one channel of the data logger. Thus, we can compare the observed and calculated responses to a known calibration stimulus to precisely determine the frequency response of the seismometer. Note that these algorithms determine only the shape of the frequency response and not the sensitivity of the seismometer. The corresponding flat pass-band sensitivity of the seismometer is determined using the Metrozet-supplied STS-1 Scale Factor Calculator V1.0 Java software applet (available via, which, when given the values of all the feedback components from the factory calibration sheet for a specific STS-1 with factory electronics, determines the corresponding sensitivity with the E300 electronics connected. This method, of course, assumes that the original factory determination of the STS-1 response parameters is accurate. An alternative method to determine the absolute sensitivity is to determine the response of the STS-1 to known micro-tilts, a procedure which is very tedious and not easily done with sufficient accuracy. We prefer to check the accuracy of the STS-1 sensitivity by comparison of large seismic ground motions inferred from the STS-1s, with the corresponding ground motions from co-sited broadband or strong motion sensors. An added advantage of comparison with the signals from the accelerometers is that their responses are flat to DC, and their calibration is easily checked by tilting them $\pm{90}$ degrees along the horizontal sensitive axes to induce a $\pm{1}$ g acceleration step.

The first algorithm, $\it {td\_tfp}$, determines the transfer function parameters using an adaptive migrating grid search methodology to minimize the difference between the observed and calculated time series response to a known stimulus function. This routine determines the calculated response via direct integration of the seismometer equation of motion response to a known stimulus input using a fourth-order Runge-Kutta integration scheme. Also, the $\it {td\_tfp}$ algorithm was specifically developed to determine the seismometer $T_{s}$ and $h_{s}$ using, preferably, the low-frequency calibration stimulus or, alternatively, the step-function stimulus data.

The free period $T_{s}$ and fraction of critical damping were determined via a grid search for the maximum variance reduction between the observed response and the theoretical response to stimulus signal input to the calibration coils. Sample outputs of the $\it {td\_tfp}$ are shown in Figures 3.34 and 3.35 and the calibration results for the BKS reference STS-1 sensors are listed in Table 4.

The second algorithm, $\it {pd\_tfp}$, determines the transfer function parameters using a grid search methodology to minimize the difference between the observed and calculated phase response to a known stimulus function. $\it {pd\_tfp}$ determines the parameters from the frequency and slope of the phase response curve when the measured phase delay between the calculated and observed responses is $\pi{/2}$. Also, $\it {pd\_tfp}$ algorithm was specifically developed to determine $f_{g}$ and $h_{g}$ using high sampling rate data (e.g. $\geq$ 80 sps) and the stepped sine (0.5-40Hz) stimulus input to the calibration coils. A sample of the algorithm output is shown in Figure 3.36.

Both algorithms, $\it {td\_tfp}$ and $\it {pd\_tfp}$, are capable of determining the transfer function parameters ($T_{s}$, $h_{s}$, $f_{g}$ and $h_{g}$) to approximately one part per thousand or better when the response to the stimulus signal is 20+ dB above the seismic background noise level.

Figure 3.34: Calibration of BKS STS-1 seismometer low-frequency responses by time domain analysis of response of the seismometer to a step function calibration stimulus. Shown are the observed (solid lines) and calculated responses (dashed lines) to a 1200 second duration current step (equivalent to an acceleration step) input to the seismometer calibration coils.
\begin{figure}\centering\setlength{\abovecaptionskip}{3pt}\epsfig{file=rau08_1_fig_5.eps, width=8cm, bb=72 162 540 670, clip }\end{figure}

Figure 3.35: Calibration of HOPS STS-1 Z-component seismometer low-frequency response by time domain analysis of response of seismometer to a swept sine function calibration stimulus. Shown are the observed (solid line) and calculated (dashed line) responses to a 40-1100 second swept sine current (equivalent to a swept acceleration) input to the seismometer calibration coil.
\begin{figure}\centering\setlength{\abovecaptionskip}{3pt}\epsfig{file=rau08_1_fig_6.eps, width=8cm, bb=85 240 520 520, clip}\end{figure}

Figure 3.36: Calibration of HOPS STS-1 Z-component seismometer high-frequency response by phase domain analysis of response of seismometer to a stepped sine function calibration stimulus. Shown are the observed (solid line) and calculated (dashed line) responses to a 0.5-40 Hz stepped sine current (equivalent to a stepped acceleration) input to the seismometer calibration coil.
\begin{figure}\centering\setlength{\abovecaptionskip}{3pt}\epsfig{file=rau08_1_fig_7.eps, width=8cm, bb=85 240 520 520, clip}\end{figure}

Temperature, Humidity, Pressure Sensor Testing

During the past year, BSL staff have tested several generations of a new temperature, humidity, pressure (THP) sensor at the Instrumentation Test Facility and on the roof of McCone Hall. The new THP sensor package is being developed and upgraded to replace the temperature and pressure sensors currently installed at bdsn stations, which are aging and failing.

The pressure sensing element is a Honeywell SDX15A2-A which is temperature compensated. The specification sheet says that the sensor range of 0-15 psi in absolute pressure results in a 90mv ($\pm{1}$%) differential change on the outputs when the bridge is excited with 12V. The sensor is operated in a bridge circuit configuration and its sensitivity is:

$P(psi) = ( V + 134.3 ) / 9.995$

where: V is the bridge output in Volts and psi is pounds per square inch (1 psi = 6894.8 Pa).

The resistance thermal detector (RTD) is a Honeywell HEL-700 with a resistance of 1k$\Omega$ at 0$^{\circ}$C. The RTD is operated in a circuit with offset and gain and its sensitivity is:

$T($$^{\circ}$C $) = ( V + 9.09 ) / .3463$

where: V is the output in Volts.

The humidity sensing element is a Honeywell HIH-4602C which is sensitive to the relative humidity. The sensor is operated in a circuit which results in a overall calibrated sensitivity of:

$\%RH = ( ( V + 9.29 ) / 3.168 - Z ) / S$

where: %RH is the percent relative humidity and V is the voltage output. S and Z are given in the factory calibration sheet as Z $\sim$0.826 mV and S $\sim$31.5 mv/%RH. Thus:

$\%RH = ( V + 6.673 ) / 0.09979. $

The factory specification sheet indicates that the response time is $\sim$50 seconds and the accuracy is $\pm{3.5}$%RH. The absolute humidity (AH) is a function of temperature, and, given the temperature, AH can be derived from relative humidity (RH) via:

$AH(g/m^{3}) = ( 0.000002T^{4} + 0.0002T^{3} +
0.0095T^{2} + 0.337T + 4.9034 ) * RH $

where: T is the temperature in $^{\circ}$C.


Development of the sensor test facility and analysis system was a collaborative effort of Bob Uhrhammer, Tom McEvilly, John Friday, and Bill Karavas. IRIS and DTRA provided, in part, funding for and/or incentive to set up and operate the facility, and we thank them for their support. Robert Uhrhammer, John, Friday, Jarrett Gardner, Bill Karavas and Barbara Romanowicz (all from BSL), Tom VanZandt (Metrozet LLC), Charles R. Hutt (Albuquerque Seismological Laboratory) and Erhard Wielandt (Institute of Geophysics, University of Stuttgart) contributed to this chapter. The STS-1 electronics upgrade project was funded by the NSF through the IRIS/GSN program, with complementary support from University funds to the BSL. We thank Jennifer Taggart and Aimin Cao for their help in preparing the figures.


Barzilai, A., T. VanZandt, and T. Kenny, Technique for measurement of the noise of a sensor in the presence of large background signals, Rev. Sci., Instrum., 69, 2767-2772, 1998.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2007, The Regents of the University of California