The bottom side template was separated into two pieces and successfully removed from the breadboard. The template was assembled together again and bagged to store it in a cabinet.
We found that the invar block for DCPD(R) was bonded with some air gap (Attachment2 1/2).
The Allen key used as a weight was too small, which caused it to get under one of the screws used as hooks and lift the block.
We've investigated the impact of this tilt.
- Bonding strength: The bonding area is ~60% of the nominal. So this is weak, but we can reinforce the bonding with an aluminum bar.
- Misalignment of the DCPD housing: The tilt will laterally move the position of the DCPD. However, the displacement is small and it can be absorbed by the adjustment range of the DCPD housing.
- Removal: From the experience with the removal of the beam dump glass, this requires a long time of acetone soaking.
- We don't need to remove the invar block.
- Action Item: Reinforcement of the bonding
Qty1 1/2 mounts
Qty2 prism mounts
Qty6 gluing fixures
Qty1 Rotary stage
Qty1 2" AL mirror
Qty1 Base for the AL mirror
=> Handed to Stephen -> Camille on Jan 27, 2023.
We continued to bond two CM mirrors and the other two steering mirrors for QPD2.
Before the bonding work, the FSR and TMSs were checked again.
FSR: 264.7925 MHz
TMS_V: 58.15125 MHz
TMS_H: 58.33375 MHz
Checked the transmission: The OMC loss was 4.3 +/- 0.2 %.
This does not make the HOMs coincidently resonant until the 18th-order (+9MHz). Looks good.
- Applied the bond to CM1 and the UV illuminated.
- Applied the bond to CM2 and the UV illuminated.
==> The cavity bonding is completed.
Removed the micrometer for CM2 to allow us to bond SM2/SM3
- Checked the spot at QPD2: The spot was a couple of mm too left. This was too much off compared to the QPD adjustment range. ==> Decided to shim the SM3 position with a piece of Al foil.
- Otherwise everything looked good. SM2/SM3 were bonded.
Invar block bonding
- There are three tubes of EP30-2 that expires on 2/22, 2023.
- A tube was almost empty. Used this tube to fill/purge the applicator. The 2nd tube was then attached to squeeze out 8g of glue mixture.
- 0.4g of fused silica beads were added to the glue mixture.
- Mixed the bond and a test piece was baked by the oven. (200F=95C, 5min preheat, bakeing 15min).
- The glue test piece was "dry" and crisp. Looked good.
- Applied the glue on the invar blocks. Confirmed that the bonding surfaces were made completely "wet".
- 4-40 screws were inserted to the blocks so that the blocks were pushed toward the template. See Attachments 3 and 4.
CM1: PZT ASSY #8 (M7+PZT11+C11)
CM2: PZT ASSY #11 (M14+PZT13+C13)
We worked on the bonding of the flat mirrors for the OMC cavity with UV epoxy.
- Prepared the UV illumination setup. Cleaned up the table a bit to spare some space for the illuminator.
- Checked the output power of the illuminator. The foot pedal worked fine. The timer was set to be 10s. The UV output from the fiber was nominally 6W. This is after some warming up for ~1min. (Checked the output power continuously with the UV power meter.)
- Checked the cavity alignment / FSR / TMS - it looked good at this moment
- We confirmed that the UV epoxy has an expiration of July 3, 2023. The bond capsule was brought from Downs right before the work started, and thawed at the lab.
- The bottom of FM1 and the breadboard were cleaned. Cleaning with lens cleaning paper + IPA remained a few specks of dust on the surface. We decided to use Vectra swabs to wipe the breadboard surface. This worked pretty well.
- Applied a tap of UV epoxy to FM1 and placed it on the template. The optic was constrained by a retainer clip.
- We found that the spot positions were significantly moved. Probably FM1 was not well touching the template before. We tried to recover the previous optical axis by aligning CM1 and CM2.
- Here is the tip: align the beam on CM1 at the desired spot. Move CM1 to bring the spot on CM2 to the desired spot. CM2 is aligned to have TEM00 as much as possible.
- We recovered reasonable spots on the mirrors. Measured the FSR and TMS (vertical and horizontal) to be 264.73MHz, 58.18MHz, and 58.37MHz, respectively. This makes the 9th-order modes well separated from TEM00. Very good.
- Gave UV illumination 10s x 2. Confirmed that the mirror is rigidly bonded.
- Continued to bond the other flat mirror. The same process was repeated.
- The bottom of FM2 and the breadboard were cleaned.
- Applied a tap of UV epoxy to FM2 and placed it on the template. The optic was constrained by a retainer clip.
- Measured the FSR and TMS (vertical and horizontal) to be 264.7925MHz, 58.15MHz, and 58.3725MHz, respectively. This makes the 9th-order modes well separated from TEM00. Very good.
- Continued to bond some less important mirrors.
- SM1 was placed on the template with the same step as above. BS2 (for QPD) and a dummy QPD housing were also placed just to check if the optical axis has any inconsistency. The good beam alignment on the QPD housing was confirmed.
- Applied a bond to SM1 and blasted the UV (20s)
- Applied a bond to BS2. Checked the alignment on QPD1 again. It looked good. UV illumination was applied.
- Placed BS3 to the cavity transmission. A dummy DCPD housing was placed at the reflection side of BS3. There was no inconsistency with the beam alignment.
- The UV illumination was applied (20s).
CM1: PZT ASSY #8 (M7+PZT11+C11)
CM2: PZT ASSY #11 (M14+PZT13+C13)
[Camille, Koji] Log of the work on Dec 15, 2023
The vertical and horizontal TMSs for OMC #4 were measured with the PZT voltages scanned from 0V to 200V.
We concluded that this alignment nicely avoids the higher-order mode structure up to ~19th order. We are ready for the cavity mirror bonding.
The RF transfer functions to the trans RF PD from the modulation on the BB EOM were taken with the presence of the vertical misalignment of the incident beam and the vertical clipping of the beam on the RFPD.
The typical measurement results and the fitting results are shown in Attachments 1/2.
The TFs were taken with the voltage 0, 50, 100, 150, and 200V applied to PZT1 while PZT2 were left open. The measurement was repeated with the role of PZT1 and PZT2 swapped.
The ratio between the TMS and FSR was evaluated for each PZT voltage setting. (Attachment 3)
When the PZTs are open, the first coincident resonance is the 19th-order mode of the 45MHz lower sideband. (Attachment 4)
When the PZT2 voltage is scanned with PZT1 kept at ~0V, no low-order sidebands come into the resonance (Attachment 5) until the PZT1 voltage is above 100V.
We found that the high voltage on PZT1 misaligns the cavity in yaw and the spot (presumably) moves to an undesirable area regarding the cavity loss.
This does not happen to PZT2. Therefore the recommendation here is that the PZT2 is used as the high voltage PTZ, while PZT1 is for the low voltage actuation.
We replaced CM2 with a PZT mirror subassembly serialized by PZT "13" (Attachment 1).
This made the transmission increase to 96.x%. Therefore the quick measurement of FSR and TSM were done. Also more careful measurement of the transmission was done.
== Alignment ==
== Quick measurement of the transmission ==
Transmission: 20.30 mW
Reflection Voltage (locked): 65.0 mV
Reflection Voltage (unlocked): 3.094 V
Reflection Voltage (dark): -6.5 mV
Incident Power: 21.64 mW
---> Mode matching 1-0.023 / Pcoupled = 21.14 / OMC Transmission 0.96
96% transmission is not the best but OK level. We decided to proceed with this mirror combination.
== Quick measurement of FSR/TMS ==
TMS_V = 58.2105MHz
TMS_H = 58.1080MHz
The HOM structure (with PZT Vs = 0) is shown in Attachment 3. 9th order modes look just fine. The excplicit coincidence is 19th order 45MHz lower sideband. (Looks good)
== Transmission measurement ==
The raw measurements are shown in Attachment 4. The processed result is shown in Attachment 5.
We found that data set 2 has exceptionally low transmission. So we decided to run the 4th measurement excluding the set 2.
Over all OMC loss
Set1: 0.029 +/- 0.014
Set3: 0.041 +/- 0.0014
Set4: 0.038 +/- 0.001
--> 0.036 +/- 0.004
We quickly measured the basic parameters of the OMC as is.
=== FSR ===
Used the technique to find a dip in the transmission transfer function (TF) with offset locking + phase modulation. The FSR was 264.79003MHz = The cavity length of 1.13219 [m] (requirement 1.132+/-0.005 [m])
=== TMS ===
Used the technique to find the peaks in the trans TF with phase modulation + input misalignment + trans PD clipping.
TMS_V: 58.0727 / TMS_H: 58.3070 => TMS/FSR V:0.219316 H:0.220201
This makes the 9th-order modes nicely avoided (Attachment 1). A slightly longer FSR may makes the numbers close to the nominal.
=== Spot positions ===
The image/video capture board turned out not functional with the new Apple silicon mac. We decided to use a small CCD monitor and took a photo of the display.
All the spots are within the acceptable range. The scattering on CM2 was particularly bright on the CCD image and also in the image with the IR viewr.
The spot on FM1/2 are right at the expected location. The spot on CM1 is 0.5mm low and 0.7mm inside (left). The spot on CM2 is ~0.25mm too high and 0.3mm outside.
(Attachment 2, a small grid is 1 mm/div)
== Transmission ==
We made a quick simplified measurement (Attachment 3).
Assuming the reflectivity of the matched beam to be ~0, the mode matching is M=1-(59.2e-3-(-6.5e-3))/(3.074-(-6.5e-3))=0.979
==> The power of the coupled mode is M x 21.28mW = 20.83 mW
The measued transmission was 19.88 mW
==> The OMC transmission (total) was 0.954 (4.5% loss)
This number is not too bad. But the spot on CM2 has too bright scattering. Next week, we want to check if swapping CM2 may improve the situation or not.
We started buikding the OMC #4.
The second trial of the water scrub
A bright scatter is visible on FM1, so I tried water scrub on FM1. This time, both surfaces of FM1 and both surfaces of BS1 were cleaned.
Smaller Vectra swabs were used for the scrub. Then the water was purged by IPA splashed from a syringe. Right after that FC was applied.
This was a bit messy process as the mixture of water/IPA/FC was splattered on the breadboard.
Nevertheless, all the mess was cleaned by FC in the end.
The transmission measurements are shown in Attachment 1, and the analyzed result is shown together with the past results.
The 2nd water scrub didn't improve the transmission and it is equivalent to the one after the two times of deep cleaning.
I concluded that the water scrub didn't change the transmission much (or at all). We reached the cleaning limit.
Another set of FC cleaning was applied to FM1/FM2/CM1/CM2 and SM2. Some FC strings are visible on SM2. So I decided to clean SM2 as well as the cavity mirrors close to SM2 (i.e. FM2 and CM2)
As a result, the bright scattering spot on CM1 is now very dim. And the loss was reduced to 4.0%. This is 0.4% better than the value before the water cleaning.
It'd be interesting to repeat the water cleaning, at least on FM1. FM1 is the closest cavity mirror to the beam dump damaged by the high-power laser pulse.
Maybe we should also clean the AR side of FM1 and BS1, as they were right next to the damaged beam dump. It is not for the loss but for reducing the scattering.
We obtained Regent grade DI water. It was poured into a smaller cup.
FC liquid was also poured into a small beaker.
Wash the mirror with a swab. We should have used a smaller swab that GariLynn has in her lab.
As soon as the mirror was wiped with the water, the FC was applied with a large brush. Don't let the water away!
Then more layer of the FC was added as usual.
The quick painting of FC made a mess around the mirrors due to excess liquid (Attachment 2). So, we decided to remove the FC remnants (on non-optic surfaces) with cotton swabs and then applied FC as usual.
This made the mess removed, however, we found the OMC loss was increased to >10%(!) (Attachment 3). We decided to continue tomorrow (Thu) with more weapons loaded consulting with GariLynn.
Aaron took the set to Cryo lab
OMC Transmission measurement after the 2nd deep cleaning
The 2nd deep cleaning didn't improve the transmission. (See Attachment 2)
The measured loss was 0.044+/-0.002
The four cavity mirrors in OMC #1 were each scrubbed using acetone and a cotton swab.
Then, the four mirrors were painted with First Contact (picture attached). The First Contact was allowed to dry for 20 minutes, then removed while using the top gun.
Conclusion on the cleaning of OMC #001
- After a couple of first contact cleaning trials and deep cleaning, the total loss was measured to be 0.045+/-0.004.
This indicated a slight improvement from the loss measured at LLO before any cleaning (0.064+/-0.004).
However, the number did not improve to the level we marked in 2013 (0.028+/-0.004).
- This loss level of 4.5% is comparable to the loss level of OMC #3, which is currently used at LHO.
Therefore, this OMC #1 is still a useful spare for the site use.
- Some notes / to-do regarding this unit:
1) The beam dump with melted black glass was removed. A new beam dump needs to be bonded on the base.
2) The connector bracket still needs to be replaced with the PEEK version.
3) The PZT of CM1 has been defunct since 2013. Combining LV and HV drivers is necessary upon use at the site. (LLO used to do it).
- Removed the first contact we left on Monday.
- Measured transmission (Set1) Very high loss! Total optical loss of 18.5%! Observation with the IR viewer indicated that CM1 has bright scattering. We suspencted a remnant of FC.
- Applied the second FC on the four cavity mirrors. This made the CM1 sport darker.
- Measured the transmission (Set1~Set3). We had consistent loss of 4.2~5.0%. We concluded that this is the limitation of this OMC even with the cleaning.
Photo of the BS1 AR cleaning process
Attachment 1: Before cleaning. Foggy surface is visible.
Attachment 2: After FC cleaning. The structure of the deposited material is still quite visible.
Attachment 3: Acetone scrubbing. Cotton Q-tip was used so that the stick does not melt with acetone.
Attachment 4: After acetone scrubbing. Nicely clean!
Acetone scrubbing was applied to HR/AR of BS1, FM1, FM2, BS2, and HR of CM1 and CM2. (total 10 surfaces)
Then final FC paint was applied to these 10 surfaces.
We'll come back to the setup on Thu for FC peeling and loss measurement.
The damaged black glass was removed from the OMC breadboard leaving the glass base.
The black glass pieces were bonded very tightly on the FS base with EP30-2. The apparent amount of the bond was not so much but it was such hard that removal by hand was not possible.
We decided to give drips of Acetone on the base hoping the gradual dissolving of EP30-2. Using a knife edge, the "filets" of the bonds were removed, but the BD was still tight.
By wedging the black glass-black glass bonding with the nife edge, the left side (the directly damaged one) was taken off from the structure leaving a tiny fragment of the glass on the base.
The remaining one was even stronger. We patiently kept dripping Acetone on the base and finally, the black glass piece was knocked off and removed from the base.
Attachment 1: The base right after the black glass removal.
Attachment 2: The black glass pieces were stored in a container with Al foil + clean cloth bed. The damaged and fogged surfaces faced up.
Attachment 3: The zoom-in shot of the black glass pieces.
Attachment 4: The base was wiped with Acetone and cleaned with FC. We will bond another BD assembly on the base, presumably using the UV epoxy.
Monday, November 14, 2022
Camille and Koji did a "deep cleaning" of OMC#1:
1) Applied First Contact to the mirror surfaces. Removed first contact after ~10 minutes.
2) Acetone scrub of the mirror surfaces with a cotton swab.
3) Applied First Contact again. Removed after ~10 minutes. We left the FC paint on for the work on Thu.
The foggy spot on the input mirror was unchanged after the first round of First Contact. But the foggy spot came off during the acetone scrub.
The measured total optical loss of the OMC was
1st: 0.015 +/- 0.003
2nd: 0.085 +/- 0.005
3rd: 0.0585+/- 0.0008
4th: 0.047 +/- 0.002
In avegrage the estimated loss is
Loss = 0.055 +/- 0.014
This is unchanged from the measurement at LLO after the FC cleaning
Loss = 0.053 +/- 0.010
Friday, Nov 11th, 2022
Setting up OMC #1 for transmission measurements:
The laser beam was aligned to the OMC cavity. The OMC cavity was locked and the transmission measurements were recorded.
Clean Supply Ordered
Fiber matching: 43.2/56.7 = 76%
S/P-pol ratio 0.7/43.2 = 1.6%
Began setting up fiber assembly for OMC testing:
-Aligned fiber mount to maximize transmission through fiber
-Adjusted polarization at output of fiber to minimize s-polarized output.
fiber input: 56.7 mW
fiber output:43.2 mW
s-polarized output: 700 uW
Upon the LLO work, the current PD arrangement in the cages are:
B1 OMC1 PDT (A1-23)
B2 OMC1 PDR (A1-25)
B3 original (C1-03)
B4 OMC2 PDT (B1-22)
C1 OMC2 PDR (B1-23)
C2 original (C1-08)
C3 original (C1-09)
C4 original (C1-10)
Started July 15, 2022 and finished Aug 30. So it took ~1.5 months (with a couple weeks of break)
Class B special tools
First Contact Kit
Bonding kit (excl EP30-2 bond)
Power meters (excl Power meter controller)
Cable bracket replacement kit
Optics / Optomechanics
=== Action done on Aug 30 ===
Fiber MM setup / Fiber coupler mount
Glass Beamdumps (for optical testing)
Thorlabs fiber coupler tool
General bent nose plier for fiber
Thorlabs collimator tiny allen
Spare High QE PDs
Spare OMC bags / Zip bags
Balance Mass 10g Qty 8 (Different Type D11*** 1.25" dia), 20g Qty 10 / Mass damper D1700301 -04 / Mass damper screws SHCS 1/4-20 x 1.25 Qty 25 / 1" screws and 1 1/8" screws
Shipping request: https://services1.ligo-la.caltech.edu/FRS/show_bug.cgi?id=25002
=== Low supply! ===
Here is the balance mass info for the LLO OMC#001 analyzed from the photographs
If we attach the additional mass, longer 1/4-20 screws (1", 1" 1/8, 1" 1/4) are going to be used.
Inspected the past LLO add-on mass configuration.
There are unknown masses at the DCPD side. It looks like a small SS mass with an estimated mass of 5g. But the DCC number is unknown.
We are going to add 10g on each corner as well as the damping aterial. We should be able to figure out the fastener / mass configuration.
The OMC #002 is ready for shipment.
Attachment 1: Work done on Sept 19, 2022
Other attachments: Putting the OMC in the pelican case.
EP30-2 bonding setup
Inspection of the delaminations in the optics side
Inspection of the bonding on the suspension interface side. All look good.
#1 The Invar bar on the cable bracket (DCPD side)
Added short (frosted) Al bars (Attachment 1) to the short sides of the invar bar. (Attachments 2/3). Some glue was sucked into the delamination gap by capillary action (=good) (Attachment 4)
#2 The Invar bar on the cable bracket (QPD side)
Added short (frosted) Al bars to the short sides of the invar bar. (Attachments 3/5). Maybe some glue was sucked into the delamination gap??? Not so clear. (Attachment 4)
#3 The Invar bar reinforced in 2016
Added a short (frosted) Al bars to a short side of the invar bar (Attachment 6). On both sides of the 2016 reinforcement, rectangular prisms are added (Attachment 6)
Some capillary action is visible beneath the invar bar (Attachment 7)
Leave it as it is for a day
Checked the delamination status:
New cable ties were installed on the cable pegs attached to the long sides of the cable bracket.
The cable bracket was successfully replaced.
- The hole size extension is going forwared now.
- Madeline and Chub are cleaning (sonicating) a drill (29/64=0.4531")
- The parts in a bag were brought to the 40m C&B lab.
- The hole is going to be 11mil=0.28mm larger than the recommendation (0.442").
It's not a D-hole. The connector has a rounded-rectangular flange that fits into the PEEK parts.
So I don't think it's an issue.
- Chub has a proper spanner to fasten the nuts. We want to use it here and LLO.
Now we got the C&Bed parts to continue to work on the cable bracket replacement.
1) Helicoil insertion
1/4-20 Helicoils were inserted into the 6 thread holes of D1300052. It went mostly okay. We witnessed that the Helicoil insertion tool delaminated the plating of the Helicoils upon insertion (Attachment 1). Stephen mentioned that this is not usual, but we didn't find anything further such as increased friction, more debris, etc. So we decided to go forward.
2) EP30-2 Kit
The EP30-2 kit was transferred from the 40m clean room to the OMC lab. The EP30-2 kit tracking was updated via C1900343
3) D1300052 reinstallation -> FAIL
Now resumed to the installation of D1300052 bracket. However, the hole size of the bracket is just a bit too small compared with the size of the mighty mouse connectors. It was already quite tight with the metal version. However, this PEEK version seems to have 0.1 mm further small diameter, and then the connectors do not penetrate the holes. The plan could be
1) Use a razor blade to shave the hole inner circle.
2) Use a cleaned drill bit to make the hole size 0.2mm bigger.
Checking the spare parts
- Conclusion for OMC#2: need PEEK cable ties
- for more OMCs: need more BHCS / PEEK cable ties / Helicoils
- (Attachment 1) The connector nut rings were removed using an angled needle nose plier. The connector shell has a tight dimension relative to the hole on the bracket. But of course, they could be extracted.
- The 4 screws mounting the bracket to the invar blocks were successfully removed. No extra damage to the bonding.
- (Attachment 2) The plan was to remove the cable pegs by unfastening the button head 1/4-20 screws from the bracket and then just replace the bracket with the new one. However, these screws were really tight. The two were successfully removed without cutting the PEEK cable ties. Two cable ties were necessary to be cut to detach the bracket+pegs from the fragile OMC. Then one screw was removed. However, the final one could not be unfastened. This is not a problem as we are not going to recycle the metal cable bracket... as long as we have spare parts for the new bracket.
- (Attachment 3) Right now, the new bracket is waiting for the helicoils to be inserted. So the OMC lid was closed with the cables piled up. Just be careful when the lid is open.
- D1300052-V3 SN001 is going to be used (Attachment 1)
- This is the PEEK version of the cable bracket (Attachment 2). The side thread holes have no Helicoils inserted. This needs to be done!
Connector arrangement check / cable routing check
Attachment 3: Connector Arrangement from the Northside
Attachment 4: Connector Arrangement from the South side
Attachment 5: Cable routing (Northside down)
At this point, the delamination of the V shape beam dumps was visible. This is the subject of bonding reinforcement.
The optical surfaces were coated with FirstContact to keep them clean / somewhat protected during the transportation.
The PD aperture was sealed with FirstContact "caps" (made by Kate in 2016?).
- DLPCA-200 trans-impedance amplifier was calibrated.
Keithley source meter 2450 was connected to the amp. Provide current and read the output voltage with the precision digital voltage meter (Agilent/Keysight).
Gain: 999.7V/A@7mA, 999.6V/A@8mA
- From the power meter spec, Thorlabs S401C seemed the best (+/-3%). So the QEs of the 9 PDs were checked with this power meter again.
- All PDs exhibited the QE of 0.95~0.96. It's all relative as the power meter has a systematic error.
- Tried to clean B1-22 and B1-23 PDs. They didn't show significant improvement after the cleaning. To avoid the unnecessary risk of damaging the PDs, further cleaning was not performed. (Some photos were attached)
- What we can do is use this result as the relative measurements.
- For OMC#2, B1-22 is the DCPD(T) and B1-23 is the DCPD(R). C1-03 and C1-12 are the spares, according to this latest result.
- At LLO, we track down the source of the throughput reduction (-10%). The QEs of the PDs are going to be tested in the same setup at once to compare their PDs and our PDs.
- 4 CLASS A wire clamp obtained from the OMC spare
- 4 more DIRTY wire clamp obtained from WB experiments (they no longer use these)
Once the later ones are C&Bed, we have enough.
- Installed the High QE PDs to OMC #002
Upon the installation, the legs of the PDs were cut by 3mm. Also, the tab of the PD could not be embedded in the DCPD housing. Therefore, the tabs were cut.
The alignment looked just fine. The weak reflections are directed to the black glass beam dumps.
- After the installation, the QEs were measured.
It is so confusing. So I decided to make the QE test setup.
Ophir RM9 with chopper (+/-5%): 8.97mW
Thorlabs S140C integrating sphere (+/-7%): 9.11mW
Thorlabs S130C PD power meter (+/-7%): 9.15mW
Thorlabs S401C thermal power meter (+/-3%): 8.90mW
So there looks ~3% discrepancy between S130C and S401C
Then tried to measure the QE of C1-03@Cage B3 with Ophir RM9
- Initial state: QE=0.95
- First FirstContact application: QE went up to 0.973
- Second FirstContact application: QE = 0.974, basically no change
- Calibrate the trans-impedance amp with Keithley
- Apply FC to B1-22 and B1-23 to see if there is an improvement
- The power should be measured with S401C because the accuracy seems better (+/-3%).
- Take photos of the PD FC process
General To Do:
- Backscatter test 2nd trial
- Start applying the first contact to the optical surfaces
- Beam dump cleaning
- Apply FC cap to the PDs
- Delamination repair (light side)
- Delamination repair (dark side)
- Cable bracket replace (dark side)
More epoxy delamination check:
DCPD R (Attachment 1): Found half delaminated
DCPD T (Attachment 2): Found half delaminated
QPD1/QPD2 (Attachment 3): Looks fine
In total we need to fix bonding of three invar bases (including the one for the cable bracket)