KAGRA Logbook

MIF (ITF Control)

yutaro.enomoto - 22:51 Thursday 22 November 2018 (7098)

Trying to find the noise sources

[Yokogawa, Sugimoto, Enomoto]

We made efforts to find the sources of the noise that binds ALS as a frequency sensor for PSL (see 7071 and comments).
We tried the followings:
-- shake mirrors in ALS-related optics in PSL room.
-- shake the fiber that carries the green light to POP table.
-- took coherence between signal of an accelerometer put on some mirror holder and the control signal of PDHX.
-- shake POP table.
Most significant was the last one, shaking the POP table.

= Optics in PSL room =
Even though we put the accelerometer on the holders for some optics, including steering mirrors for beat signal of PLL, steering mirrors for green, iris for rejecting undesired beam from AOM, and the fiber coupler, we did not see significant coherence.
See the following plots, that show the coherences and spectra when the acceleromter was put on steering mirrors for IR beam from main path, IR beam from PROMETHEUS, and green beam, respectively. And the fiber coupler.

And here is our feeling when we shook the mirrors or the fiber.
Mirrors: if we hit the mirror holder a bit strongly, some peaks apeared around several hundred Hz in the control signal of PDHX, but the structure does not similar to that of the quiet-time control signal.
Fiber: if we shake the fiber so strongly that the motion of the fiber is well visible, stucture less noise floor goes up above the quiet-time controls signal, but obviously the motion of the fiber when we shook was more than hundred times larger than that in a quiet condition.

= POP table =
When we tap (or hit) the POP table on its side or back, we saw some structures in the control signal spectrum of PDHX grow.
This plot shows the spectrum when we hit the side of the table.
The peak at 33 Hz grew.

This plot shows the spectrum when we hit the back of the table, close to the edge (-X side) of the table.
Almost all the peaks around a few hundred Hz went up.

Last, this shows the spectrum when we hit the back of the table, center of the table.
The shape is different, but this also excited many peaks.

== Note ==
During this experiment, it seemed that we mistakenly misaligned the fiber coupler on PSL room or the beam to it. I will restore the alignment this weekend.

== Comments ==
Jitter coupling in X arm green PDH? scattering? If so, how can we deal with them? \(' o `)/

Images attached to this report

Comments to this report:

yutaro.enomoto - 22:48 Friday 23 November 2018 (7107)

= Trying to find the noise sources, Cont'd =

Today I tried several things in a random way.

0. recovered the coupling of green beam to the fiber coupler

1. rotate the polarization to EOM => No change observed
We thought RFAM might be a coupling path from jitter to the sensing of PDH, so rotated the polarization to EOM to see if any change in PDHX control signal occurs or not.
I put a half-wave plate and rotate it with 2 deg steps, but did not see any changes. After this test, I removed the half-wave plate.
See the photo.

2. switch off "Noise Eater" of PROMETHEUS => No change observed

3. put an iris on POP table to reject some junk light => No change observed
I put an iris in the main path and the path from the FI to RFPD, and neither made changes.

4. tap optics on POP table => confirmed that the structures around a few hundred Hz come from mirror mounts on POP table

5. try to damp the mode of a mirror mount => failed
I tried with sticky tape to damp the mode, but I could not make it. Probably we need a better tool like rubber?

6. tap fiber on POP table => broad noise floor from ~100 Hz to 1 kHz appeared
I feel that it is easier to excite the noise floor by tapping the fiber lying on the table than hung part of the fiber.
The following plot shows the behavior of the green transmission when I tapped the fiber on the table
I start suspecting that fiber phase noise might be ~1Hz/sqrt(Hz) or more in a broad frequency region.

Images attached to this comment

yutaro.enomoto - 21:45 Saturday 24 November 2018 (7115)

= Some trial and errors =
[Yokozawa, Enomoto]

- damp the resonance of a mirror holder on POP table.
--> it seemed that we succeded to damp the resonance of one specific mirror holder, but we still see the struncture that is excited when we hit the mirror or its pedestal post.
This indicates that this structure comes from the resonance of optical post.

- vibration isolation of the fiber from POP table
--> no change in control signal spectrum observed

- coherence with signal of the accelerometer on POP table.
--> found coherence where ACC singal has spetral peaks

Images attached to this comment

kiwamu.izumi - 16:17 Monday 26 November 2018 (7119)

Here is a quick note.

As for the longitudinal phase noise in the input light introduced on the POP table (a.k.a. the Doppler shift), here is a derivation of the coupling coefficient.

The optical phase that is accumulated through the propagation in a distance L(t) can be given by

phi (t) = 2 pi nu_0 L(t) /c

where nu_0 is the laser frequency and assumed to be constant in this consideration, and where c is the speed of light.

Taking the time derivative, one obtains an equation for the effective laser frequency as

nu(t) = nu_0/ c * d L/ dt,

where we have used d (phi) / dt = 2* pi* nu(t). Fourier transforming the both sides of the equation, one gets

nu(f) = i* 2*pi* f* nu_0 /c * L(f).

Plugging the actual number i.e., nu_0 = 563 THz, and evaluating the above at 100 Hz, one can finally obtain

| nu(f)/L(f) |= 1.2x10⁹ x ( f /100 Hz) [Hz/m]