Summary:
* ended with 0.7% losses between OMCs due to polarization mismatch, and 0.17% polarization mismatch between beam from Faraday and OMC1
* floated and re-balanced SDB1
* next steps tomorrow morning is to close SDB1 suspension control and in parallel relock ITF at PRITF half fringe without using B1p, and then move to scattered light hunting
Found that the most practical working configuration is to put SDB1 on its fits with the screws in the holes. Then use the beam splitter to aim at the bench. Let the fog accumulate on the cryotrap window and diffract >90% of the power. The beam on the bench remains circular. With the nitrogen flow enabled the turbulence is too high and the beam jitters a lot and changes shape on sub-second time scale.
Screwed OMC1 with 15cNm torque on the PZT. This lead to 4% P-pol in the OMC1 resonating mode and 10% polarization losses between the two OMCs.
Figure 1 shows an OMC1 scan where the 4% on B1s2P can be seen compared to 12*B1s2
Iteratively tightened by steps of ~1cNm a side screw to correct the polarization, the central pzt screw to counteract, and finally added a second side screw. The end result is 17cNm on the central PZT screw, 13cNm on one side screw and 4cNm on the second side screw. This resulted in 0.4% of P-pol losses between the two OMCs.
Then checked the tuning of the waveplate in front of the OMC. The day before it was tuned to grossly maximize the s-polarization on B1s1 within a few %.
Figure 2 shows the starting point. The little peak at 46m00 on B1s2P is the P-pol light resonance in OMC1. It is much lower than the P-pol content of the main peak, which means that by luck the waveplate was very well tuned with only 0.02% of polarization mismatch between the beam and the OMC.
Figure 3. we then removed the waveplate with all of its mount, which indeed was fine tuning the polarization. Without it the P-pol content is the same for both peaks and corresponds to 0.3% polarization mismatch loss.
Figure 4, we put back the waveplate, in the meantime the P-pol content of OMC1 seems to have drifted, and the polarization mismatch loss increased to 0.17%
Figure 5, we then locked both OMCs to checked the polarization losses between the OMC. The B1s2P/B1_PD2 ratio is 0.014, as B1_PD2 sees only half of the power it corresponds to 0.7% losses. The mismatch increased, probably due to the OMC1 resonant mode polarization mode drift seen also with the waveplate dismounting/mounting.
Note that we haven't looked at astigmatism losses, as misalignment for both OMCs is dominant in these conditions and hard to tune. OMC#7 has some astigmatism so this might not be negligible, it will be assessed once everything is in vacuum.
To enable the slow shutter functionality from the rotator the following updates have been made on the corresponding PyAgilis software package (version v2r1):
1 - Addition of the speed (actually step amplitude) parameter on the MOVEREL command
2 - Addition of the STOP command
3 - Addition of the axis status information on the data provided to DAQ via Sms
Functionalities 1 and 2 could be tested fine. Some problems have been encountered while trying to obtaining meaningful information from the rotator driver regarding the moving status. After some investigation the problems seems being solved by resetting the rotator driver via unplug/plug of the USB cable.
This version of the software will be kept running over the week-end and Sms connection to the DAQ via FbsDet will be attempted again next Monday. Some problem were observed on DAQ side during first test connection probably because of spurious characters contained in the message back from the rotator controller. Those characters should be currently safely filtered out.
Actually, the problem on the DAQ (trend frames not properly build for some time) was due to the use of two different collection rates of the Agilis AG-UC8 channels (first every 5 seconds, then every seconds). The spurious characters were probably not the culprits .
This version is now running for both SDB1_Rot and EIBAgilisRot. The VPM setting have been accordingly updated for EIBAgilisRot including the addition of the speed parameter on the MOVEREL command.
SDB1_Rot has been also added on FbsDet and the position data are now available on the DAQ (Fig. 1)
