Physical Environment Monitor (PEM)
1-29-08
At the bottom of this Wiki page is a table of the range and sensitivity requirements for each system of the PEM. Rana made an elog entry about the lack of coherence on the Wilcoxon Model 731A seismometers below 0.3 Hz, and some suggestions for very sensitive 3-axis accelerometers. He suggested the Guralp CMG-3T or CMG-3ESP.
1-24-08
Today Rana, Tobin and I discussed setting up PEM sensors to measure acoustic, seismic, magnetic and RF ambience and find how/if it affects the interferometer in terms of noise. One task related to this project is the fixing/building of a weather station to measure atmospheric status (barometric pressure, humidity, rainfall, wind speed). Alan told me that there was a box on the wall that has serial port output, which was reverse engineered for use on one of the VME crates. It stopped working around five years ago, for reasons unknown.
Once the PEM system is set up, an adaptive filter (that neural network stuff that Keegan did last summer) could be employed to cancel out some of the noise from these sources.
To Do:
- Find sensors suitable for picking up noise (microphones, magnetometers, RF antennas, accelerometers), possibly guided by whatever is at Hanford/Livingston
- Figure out the best places to put the sensors in/on/around the interferometer
- Get data from these sources (i.e., use a DAQ channel for each of them, make some kind of amplifier for the RF antennas or magnetometers)
Sensor Ideas:
- The magnetometer could be a bunch of Hall effect probes. I am not sure yet whether measuring variation in the direction of the magnetic field here will be important--how big would this change be, anyway? What level of sensitivity should I aim for? What level of sensitivity is achievable with a Hall probe?
What we have now:
PEM System |
Model Number |
Noise |
Range |
Cost |
Acoustic |
DYNEX DX-MI101 |
|
50 Hz - 15 kHz |
$30 |
Seismic |
Wilcoxon Model 731A Accelerometers (and P31 Amp) |
|
0.05 Hz - 100/450 Hz |
|
Seismic |
Kinemetrics Ranger SS-1 |
|
|
|
Magnetic |
|
|
|
|
Radio |
|
|
|
|
Sensor possibilities:
Seismic:
Model Number |
Noise |
Range |
Cost (one unit) |
<10^-10 g |
0.033 Hz - 50 Hz |
$11,910 |
|
GURALP CMG-3ESP Compact |
<10^-10 g |
0.0167 Hz - 50 Hz |
$8,000 |
<10^-10 g |
0.008 Hz - 50 Hz |
$14,747 |
|
<10^-9 g |
0.008 Hz - 50 Hz |
$7,108 |
Guralp's vendor in the US gave me the price quotes 3/7/08. The quote said that orders will be shipped150 days after the order is received.This may be a problem. The noise levels for the CMG-3ESP/CMG-3ESPCompact and CMG-3T are almost identical, barely getting above 10^-9 g at 10 Hz and above. The CMG-40T is slightly worse, but they're all pretty close.
These three accelerometers are very close in sensitivity and features. The 40T has a few more options for sensitivity (it actually ranges from 2 x 400 V/m/s to 2 x 1000 V/m/s, the same as the other two) and a slightly higher cross-axis rejection. The electronics noise measurements are slightly different and I haven't converted them yet, but it seems likely that any one of the three will be suitable for the 40m. The PEM design paper linked below says that the 2 x 400 V/m/s translates to a sensitivity of 1e-10 m @ 1 Hz, so that would put the others at a sensitivity of 2.5e-10 m if it's just a simple matter of multiplication.
Acoustic:
Model Number |
Noise Floor |
Range |
Cost |
2e-6 V(RMS) |
8 Hz - 10kHz |
|
|
Bruel&Kjaer 4192 Pressure-field mic |
2.5e-6 V(RMS) |
3.15Hz - 20kHz |
|
3.9e-6 V(RMS) |
8 Hz - 12.5 kHz |
|
|
Bruel&Kjaer 4188 Free-field mic |
|
|
|
The LIGO document http://www.ligo.caltech.edu/docs/T/T970112-00.pdf has a good summary at the end of all the PEM sensor sensitivities and costs in a nice table format.
What the sites use: Bruel & Kjaer 4130 microphone The noise of the sites' microphones is 1e-4 Pa/rHz, based on a Schofield elog entry. According to the internet, this corresponds to 14 dB SPL (a common Microphone unit).
Streckeisen STS-2 seismometer
Guralp CMG-40T seismometer