Last night I worked on the several locking configurations:
General preparations / AS table inspection
- The AS beam looked clipped. I went to the AP table and confirmed this is a clipping in the chamber.
This may be fixed by the invacuum PZTs.
Modulation frequency tuning
RFPD Mon of the MC demodulator was check with the RF analyzer. Minimized the 25.8MHz (=55.3-29.5MHz) peak by changing the marconi freq.
This changed the modulation freq from 11.066147MHz to 11.066134MHz. This corresponds to the change of the MC round-trip length from
27.090952m to 27.090984m (32um longer).
- I wonder why I could not see good Michelson signal at REFL ports.
- I roughly aligned the Michelson. On the AP table, the RF analyzer was connected to the REFL11 RF output.
By using "MAX HOLD" function of the analyzer, I determined that the maximum output of the 11.07MHz peak
- I went to the demodboard rack. I injected -61dBm from DS345 into the RFEL11 demodboard. This produced
clean sinusoidal wave with the amplitude of 4 count. The whitening gain was 0dB.
- The output from the PD cable was -64.0dBm. So there is ~2.5dB loss in the cable. Despite this noise, the demodulation
system should be sufficiently low noise. i.e. the issue is optical
- The Michelson was locked with AS55Q. And the REFL11 error signals were checked.Fringe like feature was there.
This suggested the scattering from the misaligned PRM. The PRM was further misaligned. Then some reasonable
(yet still noisy) Michelson signal appeared. (Usual misaligned PRM is not at the right place)
Q. How much scattering noise (spurious cavity between PRM and the input optics) do we have when the PRM is aligned?
Q. Where should we put the glass beam dumps in the input optics?
Q. Can we prepare "safe" misaligned place for the PRM with the beam dump?
- The Michelson was locked with REFL11Q. From the transfer function measurement, the gain difference between AS55Q (whitening gain 24dB)
and REFL11Q was 32dB. The whitening gain was 0dB. In fact I could not lock the Michelson with the whitening gain 33dB (saturation???)
The element in the Input matrix was 1, The gain of the servo was +100. BS was actuated.
Coupled cavity tests
- At least REFL11 is producing reasonable signals. So what about the other REFL ports? The Michelson signals in the other frequencies
were invisible. So I decided to use three-mirror coupled cavity with the loss PRC.
- Aligned X arm, Misaligned ETMX, ITMY. Aligned PRM.
- Locked the PRM-ITMX cavity with REFL11 and REFL33.
- Aligned ETMX. If I use REFL11I for the PRC locking, I could not lock the coupled cavity. But I could with REFL33I.
This is somewhat familiar to me as this is the usual feature of the 3f signal.
- The coupled cavity could be locked "forever". To realize this I needed to tweak the normalization factor from 1.0 to 1.6.
Q. How does the coupled cavity change the response of the cavity? Can we compensate it by something?
Q. Measure open loop transfer functions to check if there is any issue in the servo shapes.
- Transmission during the lock is 3.2 while the nominal TRX with PRM misaligned was 0.93.
This corresponds to power recycling gain of 0.17.
- X arm:
- Source: POX11I, phase 79.5 deg, whitening gain 36dB
- Input matrix: POX11I->1.0->XARM, Normalization TRX*1.60
- XARM servo gain +0.8, actuation ETMX
- XARM trigger 0.25 up, 0.05 down. XARM Filter trigger untouched.
- PRC: (sideband locking)
- Source: REFL33I, phase -34.05 deg, whitening gain 30dB
- Input matrix: REFL33I->1.0->PRCL, Normalization None
- PRCL servo gain +4.0, actuation PRM
- PRCL trigger None
- Same test for the Y arm. At the moment ETMY did not have the OPLEV.
Same level of transmission (~3.3)
- Y arm:
- Source: POY11I, phase -61.00 deg, whitening gain 36dB
- Input matrix: POY11I->1.0->YARM, Normalization TRX*2.1
- YARM servo gain +0.25, actuation ETMX
- YARM trigger 0.25 up, 0.05 down. YARM Filter trigger untouched.
- PRC: (sideband locking)
- same as above
Sideband PRMI attempt
- Now I got some kind of confidence on the REFL33 signal.
- So I tried to get any stable setup for sb PRMI, then to find any reasonable MICH signals anywhere else than AS55Q.
- With REFL33I(PRCL) & AS55Q(MICH), I got maximum ~10sec lock. It regularly locked. It was enough long to check
the spectrum on DTT. But it was not enough long to find anything about the MICH signals at the REFL ports.
- I tried REFL33Q for MICH. The lock was even shorter but could lock for 1~2 sec.
Q. What is the cause of the lock loss? I did not see too much angluar fluctuation. The actuation was also quiet (below 10000).
- PRCL: (sideband locking)
- Same as above except for
- the PRCL servo gain +0.05, No limitter at the servo output.
- Trigger POP22I (low pass filtered by LP10) 20 up, 3 down
- AS55Q -24.125 24dB -> x1.0 -> MICH -0.7, No limitter -> ITMX/Y differential
- REFL33Q -34.05dB -> x2.0 -> MICH same as above
- For both case, trigger POP22I (low pass filtered by LP10) 20 up, 3 down
At this point Jenne came back from dinner. Explained what I did and handed over the IFO.