Several people reported that the DARM calibrated signals from the OAF model did not look healthy in the past two or three days. Looking at the data in frequency domain, I confirmed that they have behaved funny. Cosulting with Jeff K, we decided to gradually migrate the functionality from OAF to CAL-CS model (alog 16669) because CAL-CS is going to be the official calibration place in future anyways. We are still in the process of moving from OAF to CAL-CS, but a calibrated DARM signal is now available in CAL-CS. The funny behavior seen in OAF is not seen in CAL-CS so far. The calibration accuracy is not yet examined.
(Bad OAF data)
Looking at the data in frequency domain, indeed they looked funny -- the spectral shape of, for example, OAF-CAL_DARM_DQ was suspiciousely featureless with the noise floor around 100 Hz much higher than it should be by a couple of orders of magnitude. Also it did not show a roll-off shape at 5-ish kHz for some reason. A confusing fact is that it sometimes looks behaving correctly and sometimes doesn't. I did not make a further investigation because we decided to move to CAL-CS.
(CAL-CS)
I copied the DARM-related filters that I had in OAF over to CAL-CS. So it is right now completely a duplication of what we had in OAF. I did not move the DRMI, CARM, or IMC filters yet.
Here is a brief summary of how the current calibration was done.
The current one relies on the reponse of the ALS DIFF VCO. Since we knew the VCO response in Hz/V, we were able to calibrate the ALS DIFF sensor into meters using the dx/ L = d(nu)/nu relatation. Then, by knowing the UGF of ALS DIFF, we calibrated the ESD coefficient in meters/counts. The last step is to calibrate the DARM optical gain by measuring the UGF in the final DARM loop.
I checked in the matlab code, which I used for the calibration, into svn. It is Runs/S7/Common/MatlabTools/LSC_DARM_calibration.m in the calibration SVN. For more details, please refere to the code.
H1:CAL-DELTAL_EXTERNAL_DQ is the new calibrated DARM signal. The signal is digitally whitened by five zeros at 1 Hz and five poles at 100 Hz.