Cont'd from 16565. K. Tanaka, Akutsu w/ help of Mirapro
Abstract
Sealed the K400 flange for IMC_REFL (by Mirapro), installed the high-quality viewport there, and relocate the beam dump for the ghost beams called K, L or MCF_GB5,6 (referred in 16565) so that this beam dump can catch both the ghost beams.
Details
In the morning, thanks to the Mirapro folks (Nakamura-san and Itoh-san), the K400 flange for IMC_REFL returned onto the MCF chamber (Fig. 1). Note that this flange had detached by us as reported in 16102. Then I attached the high-quality viewport (category A; see here for the dimension, and here for the adapter) for extacting the IMC_REFL beam to the optics table outside the chamber. Note that the screws to fix the viewport were left loosely, as I needed to do smoe tweaking as follows.
I checked the ghost beams created at this viewport window. They are called K, L or MCF_GB5,6 as discussed in 16565. With 30-70(?)mW P-pol input, they were clearly visible with a sensor card. The difficulty was that the ghost beams had different height each other. The ghost "L" was flying in the mostly same height as that of the input beam; this would be due to that the ghost L was created from the IMC_REFL beam via reflection at the non-wedge surface of the viewport window. The ghost "K", which was created in the same manner but at the wedged surface was not this case. And unfortunately, the height of K was slightly lower than the under lim of the aperture of the beam dump dedicated for catching these two ghost (Fig. 2).
To resolve this situation, there would be mainly two options. The first option is to move the beam dump toward the viewport window so that the vertical separation of K from L is small enough to be caught at once with this beam dump. But the existence of the ghost beam "J" (see 16565) will interrupt this way if the beam dump comes to close to the viewport window. So I rejected this idea.
The second option is to rotate the viewport window when fixing; due to there are 16 bolts to fix the viewport window, meaning 15-angle choices. After a few trials, we found a nice angle, but this was not match the marking that indicates the wedged side (corresponding to 2-bolts angle in clock wise). Actually, when I assembled this window to the viewport adapter in 2015, I might mistake the direction. Anyway, with this nice rotation angle,
- The height of IMC_REFL, which was monitored with the usual GigE camera for IMC_REFL, did not change so much but only shifted in horizontal. And for some reason, a small spot of ghost(?) disappeared (or clipped anywhere between the viewport window and the camera)
- The ghost beams K and L came the same height (Fig. 3)!
So, I somehow slightly relocated the beam dump to catch both K and L. Phew.
