As a test for the CARM offset reduction using the arm transmissions, I tested the transition from the ALS IR lock to the arm transmission lock using X arm.
In summary, the transition to the TR lock was succeeded, though I could not automate the process.

1. Transfer function measurement: X arm detuned with the PDH signal
To understand the behavior of the TR signal, I measured the transfer fuction from REFL_45_I to TR CARM while X arm was detuned using RF45_I.
The result is attached. Note that the amount of the detune was such that the transmission power degraded by 25 %.

2. TR X arm lock
I tried the transition from the ALS IR lock to the TR lock.
The settings of the TR lock is attached.
The measured open loop transfer function is also attached.

Typica procedure was the following:
- feedback the ALS X PDH signal to MCL.
- sweep the LO frequency of X PLL to find the main laser resonance.
- offset the LO frequency from the resonance by ~150-200Hz.
- put 0.001 in the matrix element, TR CARM to MCL, and load it with the ramp time of 5 sec
- put 0.002 in TR CARM to MCL, 0.2 in X PDH to MCL (this number is 1 at the beginning), and load it with the ramp time of 5 sec
- put 0 in X PDH to MCL and load it with the ramp time of 5 sec

Note:
- While reducing the gain of ALS IR loop, the lock point was drifted so that the lock was lost sometimes. The amount and direction of the drift seemed random lock by lock, and this was a part of the reason why I could not automate the process.
Probably this is due to the offset in ALS IR signal; the offset just before the matrix. A systematic way to reduce the offset before the transition process seems necessary.