KAGRA Logbook

Komori, Sugioka, Tanaka, Dan

Summary

To address a persistent ~0.6 Hz feature limiting laser power, we manually tuned the TM coil balances. Coil gains were adjusted for both HP and LP for ETMX suspension so that H1/H2 (P) and H3/H4 (Y) responses match. Final checks using length excitation confirmed that the leakage into P/Y was improved.

Motivation

We see a persistent ~0.6 Hz feature, when we increase the laser power at the locked state.
To mitigate this, we decided to improve the TM coil actuator balance (reduce P/Y coupling) for each Type A suspension.

1) First trial: `coilbalance_TypeA_TM.py` (ETMX)

We first tried the standard script: /users/Commissioning/script/VIS/coilbalance_TypeA_TM.py

Procedure (summary)

Bring IFO to DOWN.
Set ETMX suspension GRD to ALIGNED.
Hold output for MN OLDAMP to keep alignment at a good oplev point.
Move to ISOLATED.
Run the python script:
- It injects a sine wave to K1:VIS-ETMX_TM_LKIN_{P,Y}_EXC.
- Then it observe the demodulated TM oplev signals and evaluate gain values of H1–H4 to reduce P/Y coupling.

Result

The script returned an unexpected value: H1 = -0.0013.
This suggested a possible issue (e.g., broken H1).

2) Check if H1 is broken (it is NOT)

Injected a sine wave to the input of TM COILOUTF_HP and checked TM oplev response.
Excited H2 first (known-good) and confirmed a clear pitch response in the spectrum.
Excited H1 in the same way and observed a similar pitch response level.

Conclusion: H1 is not broken. We decided to tune coil balance manually instead of the script.

3) Manual coil balance measurement/tuning (HP)

Measurement script: /users/Commissioning/data/VIS/ETMX/2026/0127/TM_coil_balance_check_HP.xml

Injection settings

Gain = 5000
Frequency = 12 Hz
H1: REF 0–2, H2: REF 3–5, H3: REF 6–8, H4: REF 9–11

Measured values at 12 Hz (used for tuning) (Fig: TM_coil_balance_check_HP.png)

P:
- H1 (REF1): 0.00992818
- H2 (REF4): 0.00908853
Y:
- H3 (REF8): 0.00748049
- H4 (REF11): 0.0073523

Update rule:

new_gain = current_gain × (reference oplev / target oplev)
This is because we did not change the gain values before the measurement.
Update H1 to match H2 in P:
- H1: 0.961 × (0.00908853 / 0.00992818) = 0.8797
Update H4 to match H3 in Y:
- H4: 0.725 × (0.00748049 / 0.0073523) = 0.7376

Applied updates (HP)

K1 ETMX TM COILOUTF_HP FILTERS
H1: -0.961 → -0.8797
H2: no change
H3: no change
H4: 0.725 → 0.7376

Verification: TM_coil_balance_check_HP_after.xml (Fig: TM_coil_balance_check_HP_after.png)

H1 excited → P(REF1): 0.00909664
H2 excited → P(REF4): 0.00907685
H3 excited → Y(REF8): 0.00749889
H4 excited → Y(REF11): 0.00751527

4) Manual coil balance measurement/tuning (LP)

Data: TM_coil_balance_check_LP.xml

Injection settings

Gain = 5000
Frequency = 12 Hz
H1: REF 0–2, H2: REF 3–5, H3: REF 6–8, H4: REF 9–11

Measured responses (Fig: TM_coil_balance_check_LP.png)

H1 excited → P(REF1): 0.00990401
H2 excited → P(REF4): 0.00970738
H3 excited → Y(REF8): 0.00797979
H4 excited → Y(REF11): 0.00782329

Gain update (same rule)

H1: 0.961 × (0.00970738 / 0.00990401) = 0.9419
H4: 0.725 × (0.00797979 / 0.00782329) = 0.7395

Applied updates (LP)

H1: -0.961 → -0.9419
H2: no change
H3: no change
H4: 0.725 → 0.7395

Verification: TM_coil_balance_check_LP_after.xml (TM_coil_balance_check_LP_after.png)

H1 excited → P(REF1): 0.00970023
H2 excited → P(REF4): 0.00969923
H3 excited → Y(REF8): 0.00797811
H4 excited → Y(REF11): 0.0079572

5) Final cross-coupling check: push Length (L) and confirm minimal P/Y leakage

As a final sanity check for both LP and HP, we injected into Length (L) from SUMOUT and confirmed that the leakage on P/Y went down after this gain modification.

LP (Fig: TM_coil_balance_check_LP_push_L.png)

File: TM_coil_balance_check_LP_push_L.xml
Gain = 10000, freq = 12 Hz
REF 0–2 (blue): after gain tuning

HP (Fig: TM_coil_balance_check_HP_push_L.png)

File: TM_coil_balance_check_HP_push_L.xml
Gain = 10000, freq = 12 Hz
REF 0–2 (blue): after gain tuning

The leakage looks better after this coil balance work.