JGW-T2011606-v1
- Document #:
- JGW-T2011606-v1
- Document type:
- T
- Submitted by:
- Tak Lun Terrence Tsang
- Updated by:
- Tak Lun Terrence Tsang
- Document Created:
- 11 Apr 2020, 08:42
- Contents Revised:
- 11 Apr 2020, 08:42
- DB Info Revised:
- 11 Apr 2020, 08:42
- The current configuration of the feedback system at the inverted pendulum stage utilizes only LVDTs
with sensor correction. A conservative residual motion requirement of the inverted pendulum was derived from such
configuration under the condition of 50th percentile seismic noise. Under this requirement, the inverted pendulum
residual displacement should be less than 0.0128 µm in terms of band limited integrated RMS within 0.1 Hz to
1 Hz. The total integrated RMS figure is not as representative due to noise amplification of sensor correction and
complementary filter at lower frequencies, but it was included for reference anyway.
Complementary filters and control gain were optimized for a list of different geophone noise level under different
conditions. Noise level was taken from the list as the requrement under the condtion if the optimal filters give a
simulated residual motion meeting the aforementioned requirement.
Without sensor correction, it requires the inertial sensor noise to be impractically low in order to meet the requirement. With sensor correction of 40% inter-calibration mismatch, the geophone noise requirement is the noise
21% lower than the typical geophone noise under 50th percentile seismic noise. It would as well require the sensor
noise to be unrealistically low under 90th percentile seismic noise under this condition. With 10% of mismatch, the
typical geophone noise level should be sufficiently low under ground motion of 50th percentile, whereas the required
geophone noise level is 15% lower than that of the typical geophone noise under 90th percentile seismic motion.
With the typical geophone noise level, the tolerable sensor correction inter-calibration mismatch is 7%. If intercalibration error of less than 7% is achievable, then upgrading inertial sensors is not necessary. This means that it
is possible to withstand 90th percentile ground disturbance with current hardware if the implementation if sensor
correction is good enough. Therefore, it is very crucial to implement sensor correction in the near future.
Under 50th percentile seismic noise, the optimal gain is around 3 times higher than the current gain of the ITMY
inverted pendulum and is around 30 times higher under 90th percentile ground noise. In the latter case, the filter
must be reshaped since the optimal gain exceeds the gain margin. Nevertheless, the drastic changes in optimal gain
strongly encourages adaptive control. This should also be considered.
DocDB Version 8.7.10, contact
Document Database Administrators