Reports 1-1 of 1 Clear search Modify search
Optical characterization (Optical characterization)
Allocca, Chiummo, Yamamoto, Gouaty, Bonnand - 20:37 Thursday 24 May 2018 (41555) Print this report
Arm losses measurements

Today's shift was devoted to the arm losses measurements. We did it as usual by comparying the power reflected by the arm cavities when these are locked and unlocked, respectively.

Some preparatory work was performed to make sure that no ghost beams were falling on the photodiodes, and that the beams were well centered on each PD. In order to do that, we changed the misaligned position of the ITMs by ~50urad. Any SR misalignment was uneffective. The PR was left in its usual "misaligned" postion, since its misalignments did not have an effect on the power level of the PDs of the central area. Eventually, we made a scan of the TX and TY setpoint of both SPRB and SDB2 in order to make sure about the PD centering.

At the very beginning of the shift, Romain G. zeroed the offsets on all the PDs but B2_PD1 and B5_PD2, where it was not possible to close the shutters.

In order to have more signals available as backup, we looked also at the total power on the quadrants (namely the *QD*Sum* channels) and on the cameras, after making sure to have the right gain and integration time for it not to saturate.

First of all, the coherence between photodiodes was checked in single bounce from each ITM. In particular, it was checked between PDs at the same port and between PDs at different ports. Between the PDs at the same port, the coherence was good, except for B4, probably because of the faint light on it (figure 2). On the other hand, we observed a lack of coherence between PDs at different ports, even at few Hz. (figure 1)

After these checks, we started our "lock/unlock" measurements, one arm at the time. The reflected power was measured in three different conditions:

- cavity locked

- cavity unlocked (but still aligned)

- cavity unlocked and ETM misaligned (only ITM aligned)

You can find all the GPS (commented) in the attached .txt file.

 

From an online look, it seems that B1p is sensible and at a very first sight not worrying, while the others which seem to be a bit funny, and they deserve a more accurate analysis. (figure 3)

 

SIDE NOTES:

- when the North arm was misaligned, we could still see some light and flashes on the B5 quadrants (figure 4)

- glitches are visible about every 100 seconds on all the quadrants of the central area (see figures 5, 6 )

Images attached to this report
Non-image files attached to this report
Comments to this report:
capocasa - 8:57 Monday 28 May 2018 (41579) Print this report

I have looked into the data of the lock/unlock in order to extract the arm cavity round trip losses.

In the attached pdf it is ploted the power reflected seen from different photodiodes and quadrants during the lock/unlock.

I used the power reflected when the cavity is aligned but not locked as a normalization factor (assuming that the reflectivity is 1 in this case).

With such nomaization, the power reflected when only IM is aligned should correspond to the reflectivity of the IM (about 0.986) while the reflectivity R measured when the cavity is locked should allow to extract the losses using the formula:

L = T/2 (1-R)/(1+R) 

with T transmission of the IM mirror  (north 1.375%  west 1.373% )

I have plotted the reflectivity on the left y axis and the corresponding RTL on the right y axis.

RESULTS

All the photodiodes/quadrants measured a value for the IM reflectivity  which is compatible with the expected one.

North arm

B1 (PDs and quadrants): losses below 70 ppm.  (Pag 1) 

B4 (PDs): large fluctuations in locked state but losses below 70 ppm. (Pag 2) 

B5 (PDs and quadrants): losses between 200 and 360 ppm. Different sensors measure different RTL levels. (Pag 3) 

West arm

B1 (PDs and quadrants): losses below 70 ppm. B1p_QD2 sees higher power when the cavity is locked(!) ( Pag 4)

B4 (PDs): large fluctuations in locked state but losses below 70 ppm. (Pag 5) 

 
OFFSET
 
Since Annalisa et al. reported that it was not possible to zero B5_PD2 offset I have estimated it using a IMC unlock. (Pag 6) 
It seems that there are some residual offsets on other PDs and quadrants that maybe shoud be taken into accout.
Non-image files attached to this comment
Allocca, chiummo, Ruggi - 19:02 Thursday 31 May 2018 (41633) Print this report

Follow up of losses measurement.

It seems that B1p_PD1 is the sensor with less spread of the outcomes (it is also the one with more power impinging onto it when we have a single arm locked on resonance), so we had a closer look to it during the stretch of measurements for both arms. On figure 1 histograms with the arm reflectivities (losses) distribution during the lock segments both for North and West arms.

Data  (sampled at 50Hz) were normalized by the reflectivity of the related arm when off-resonance and by input power fluctuations. This normalization confirms what already reported by Eleonora about the rtl for the arm cavities, both within 50-70ppm and very close to each other. This is also in agreement with asymmetries estimated from CMRF in 41634 and the upper limit derived from carrier recycled gain measured in 41344.

A curiosity about the West arm distribution: there are two peaks. Those peaks are also there when we consider each lock segment on its own instead of the cumulative distribution. On figure 2 the distributions for each lock segment is reported with a different color.  A bit of skewness is also present on the distribution of the North arm reflectivity, but less evident than for the West one. On figure 3, a plot of one of the stretches of locked arm for West cavity is reported: the reason for the two peaks is most probably the regular oscillation on the recorded power (fringes with unnoticed ghost beam?!?).

Overall, based on the B1p_PD1 outcomes, it does not seem that losses in the arms have changed with respect to pre-monolithic fiber installation.

Images attached to this comment
Search Help
×

Warning

×