For reasons unknown, the seismic spectra posted above Rosalba has been wrong since ~January when it was first posted. The noise that we were claiming was waaaay lower than is really possible.
Rana and I checked the calibrations, and the numbers in DTT for the Ranger and the Guralp are correct (it's unknown what was being used at the time of the bad plot) - Cal for the Guralp is 3.8e-9 m/s, and for the Ranger is 1.77e-9 m/s.
Something is funny with the accelerometer calibration. Hopefully Kevin's investigation will sort it out. Their calibration used to be 1.2e-7 m/s^2 , but it was changed to be 7e-7 m/s^2 to match the noise level of the accelerometers with the seismometers at ~10Hz. We need to go through the calibration carefully and figure out why this is!
Posted above Rosalba for easy reference, and attached below, is the new seismic spectra. The black trace is when the Ranger's mass is locked down, and the teal circle markers indicate the Guralp Spec-Sheet Noise Floor.
** Rana says> the y-axis in Jenne's plot is (m/s)/sqrt(Hz). The Guralp has a velocity readout bandwidth of 0.03-40 Hz, so we would have to modify the calibration to make it right in those frequencies. I believe the Ranger cal has the correct poles in it. The huge rise at low frequencies is because of the 1/f noise of the SR560. |