[For the commissioning team]
We did a quick measurement of the loss in the x-arm after improving the pointing into the cavity using IM4 and PR2.
It is important to note that there were no angular loops closed when doing this measurement, so the power build up wasn't optimized ( LSC-TR_X_QPD_B_SUM ~ 0.95). We were running short on time and verified that the DC angular alignment was good enough so that the locked vs unlocked state gave us a big enough dip to estimate a visibility.
| Channel | Locked Power(Cts) | Unlocked Power(Cts) | Visibility | Loss PPM (calc'd) |
| AS_A_DC_NSUM | 286 | 307 | 93% | 302 |
Comparing the above numbers to alog-38493. They are close, but I can't say anything definitive about the change in loss between the two ITMXs (before and after swap), we can redo this measurement when we have more time to close angular loops (and maybe try a more sophisticated loss measurement). Also important to note, mode-matching contributes to the loss as well and this is not taken into account with this measurement.
Sheila commented that the visibility/loss is a bit concerning, so I trended some other channels to see if they gave any other estimates, unfortunately, the ones I could think of (REFL and AS) were not of much use because they didn't show an obvious dip (First and second attachment) . It is important to also note that the dark offsets for ASC-AS_A_DC_NSUM were on and roughly correct.
That got us thinking, if there is alignment or modal mismatch, will we be over or under estimating the total loss with this measurement? And how much does the modulation depth from the sidebands cause a difference in the power ratio between unlocked and locked compared to the losses of the ITM?
Using Finesse and the as-built Nebula page, I modeled the single x-arm with input power of 1 W. The modulation depths for 9 and 45 MHz was measured by Kiwamu in aLOG-8867. Also, setting the arm cavity eigenmodes as the basis, I estimated 15% mode-mismatch and 15% alignment mismatch by injecting 01 and 02 modes alongside the 00 into the cavity. Another parameter is of course the loss at the input coupler which for display purposes I used 5ppm. (Third attachment)
Conclusion:
The loss of the input-coupler dominates over the modulation depth from the sidebands, and the mismatch either from modal or misalignment will cause us to underestimate the total losses.
