We performed the measurement which is close to the real case when we use QPD. The difference is only light is adjusted into only one segment of QPD.

We could see the modulation frequency we are interested in can be seen.

I tried to see the transmitted flash with scanning the laser frequency, but I could see nothing.
I decided to remove the input and output mirrors and do the alignment work again.

I also modified the PDH servo.
The schematic of the servo will be uploaded on wiki.

Pengbo and Yuhang

We performed the measurement although the mirror is not very stable. As you can see from the attached figure 2. This makes the low frequency measurement of shot noise(with the corruption of backscattering) much worse.

The measurement of squeezing level is only 2.5dB.

We measured again with 5mW green power, but this time we make the beam smaller to see if it will decrase or not.

The result is that shot noise will decrease if the power density is higher than 47mW/mm2.

There is a discrepancy between this measurement and the one measured last time(entry 2067). I think the reason is that, the beam size is highly related to position. And this time we removed QPD and we may put it back to a slightly different position. Then it leads to this discrepancy.

We need to use a more robust telescope if we want to have a more precise measurement.

Pengbo and Yuhang

We measured shot noise spectrum with 5, 15, 26mW seperately. For each power, we measured also with different beam size ranging from to.

The set-up is using the second segement of QPD2. We checked DC voltage with oscillscope. We also checked RF channel after an amplification of 32dB with spectrum analyzer operates in 1MHz RBW(this time we average for 10 times so that the noise vurve is smooth). The noise spectrum of RF signal is plot for each case.

1. 5mW case(attached as the first picture): the noise floor is the same for all the beam size.

2. 15mW case(attached as the second picture): the noise floor reaches maximum when the power density is below ~25 mW/mm2

3. 26mW case(attached as the third picture): the noise floor reaches maximum when the power density is below ~22mW/mm2

We didn't measure the changing point for 5mW, however, from the beam density we measred, all the measurement we did for 5mW has low power density relative to the threshold (roughly between 22 and 25mW/mm2). We should see the noise floor decrease when the beam size is smaller than 500um.

Conclusion: The QPD response will saturate and decrease if the power density exceeds around 23mW/mm2.

(Notice: the beam size in this entry is diameter)

Simon, Pengbo
 

Attach to this report I show the result on shinkosha #7 sample with different polarization input beam(S-pol and P-pol).

We can see a 1 percent difference from these two maps, which is even smaller than the difference for TAMA-size sample.

Simon, Pengbo

After the birefringence measurement, we change to the absorption system with a controllable polarization of the laser. First, we did two XY-plane absorption measurements on TAMA#1 with different polarization, which is P-pol and S-pol. Then we choose a small area of the mirror and did another measurement under the S-pol incident beam. We can see a very clear structure in the map. Then we did another YZ-plane absorption measurement with an S-pol incident beam. The distribution is quite homogenous.

We change the sample to Shinkosha #7, then follow what we did before, checking whether the polarization might have some influence on the absorption. We already have the result of the absorption map with the S-pol incident beam. The result is almost the same compare with the former result.

We did the same measurement as in logbook entry 1875, the QPD response to different size of the green beam

The measurement was done by using the green beam reflected by the green mode cleaner of the Virgo squeezer(1500W). 

First, we measured the green beam size without any lenses and saved the data file from the beam profiler.  The beam profiler has 1928*1448 pixels corresponding to active are of 7.1mm*5.3mm, we got power value at each data point, and we did a 2-d Gaussian fit to these data points. One of the fit shows in figure 1. From the beam plotted in this figure, it is clear that the beam has astigmatism, so the final beam waist size and position are quite different in two axes. Below is the number (position zero is just a random point we chose easy for the measurement) 

By checking the 2d plot of the raw data, we found out the beam profiler is saturated (fig 2 shows the top cut shape). But in the 2d fit, we were not able to remove these points by substituting them into 'NaN' while using 'lsqcurvefit' function in Matlab, because this function needs 'double' format input.  Then to check the quality of the fit, we plot the difference between the fit the original data, result shows in figure 3. It seems the fit is fine. Since the data is very noisy, I was guessing maybe we actually got the peak of the Gaussian, and the saturation part is just the noise. 

Anyway, after measuring the beam size, we put a 50mm lens and measured the beam with the QPD in different positions. We did two groups of measurements with different power by changing the density. the results show below.

Eleonora and Yuhang

We measured the filter cavity reflection beam height. It is 74.5mm, which is 1mm lower than injection beam measured three months ago.

This entry is log on Mar. 2nd.

What I did

I aligned the output mirror of the folded cavity so that the reflected beam can be picked off at the FI.
The procedure is as follows.

1. Install the output mirror.
2. Check the reflected beam postion by a sensor card.
3. Insert the shim sheet between the mirror holders to adjust the direction of mirror.

Next Step

Align the input mirror as the same procedure.
Then scan the laser frequency and monitor the transmitted beam.

Note that back scattering on Saturday was particulary bad. It seemed the mirrors were more excited than usual.

By the way I remember that before Christmas, whitout the additional Faraday, we managed to measure less then 2 dB above shot noise at 30 Hz. See here. 

What I did

• Replaced two pedestals used in the chamber
• Removed the output fused silica mirror and re-aligned the beam. At this moment, the cavity direction was not good as the beam reflected by the apex mirror could not reach the center of the output mirror. So I tweaked the cavity direction.
• Re-installed the output fused silica mirror and checked the transmitted laser power.

Some results

The transmitted laser power was about 1.03 uW with 6.4 mW input power.
Assuming no loss in the apex mirror , the transmittance of the fused silica mirror is estimated about 0.016%.
In practice, there are some loss in the apex mirror and the transmittance of the fused silica mirror is slightly larger than 0.016%.
According to the spec sheet, the transmittance of the fused silica mirror is 0.017%.
So this result is reasonable and the loss in the fused silica mirror is not too large, about 30ppm with the assumption that the reflectivity of the fused silica is same as shown in the spec sheet (99.98%)
In any case, I need to compose the cavity and see the flash.

Next step

• Tweak the alignment of the output mirror to pick off the reflected beam. In fact, I could see the reflected beam on PD by adjusting the direction of the output mirror. Shim sheet, however, is indispensable to maintain the optimal direction.

[Aritomi, Yuhang, Eleonora]

We measured shot noise and squeezing with additional faraday before homodyne to reduce back scattering (attached picture). 

Difference between black and blue is back scattering with faraday which is much smaller than before. Difference between  blue and red curve might be effect of CC. 

I re-installed a folded cavity inside the chamber with one apex silicon mirror.
Then I aligned the 2 steering mirrors to pick off the reflected beam and I could pick off the beam from the chamber.
A PD was put the transmitted side to check the transmitted power when the cavity is composed.
After that, I installed a fused silica mirror as the output mirror and tried to see the transmitted beam power.
Some amount of stray light, however, prevented me to check the transmittance of the fused silica mirror.

I will install an input fused silica mirror and try to see the flash by scanning the laser frequency.

[note]
There is an optical loss at the viewport, about 10%.

The temperature change inside and outside the bench is monitored. At the same time, the suspended mirrors position is also monitored. 

The comparison is attached.

Also the seperate measurement is attached.

In order to mitigate back reflection noise we plan to install an additional  Faraday isolator (FI-1060-5SC-HP) on the squeezing path in reflection from FC, at the edge of the optical bench. See pic 1 and 2.

To finalize the istallation we need to modify a bit the custom support: drill holes for the screws and reduce the hight of about 2.5 mm. (ATC will do it by next Wednesday afternoon)

The Faraday should bring additional 3% of propagation losses, the associated squeezing degradation (simulated in pic 3) doesn't seem too bad.

The drawback is that the faraday will prevent to close one side of the bench cover. But so far we have never closed it.

The beam waist is ~1mm and FI aperture is 5 mm in diameter, so clipping losses shoud be negligible.

Aritomi and Yuhang

We measured backscattering noise with different kinds of filters added to the servo we had. The added filters contain corner frequency ranging from 0.03Hz to 10Hz. The gain is also adjusted for different corner frequencies. All the measurement is plot together in the attached figure.

According to the measurement done in last Friday, we have a variation region of back scattering noise. It is shown in the attached figure as a grey area.

We could see no matter how we set the filters, the back scattering noise is always in the variation region.

I measured backscattering noise of homodyne in two configurations. The first case is that BHD's signal port opens to filter cavity with the path between filter cavity and OPO blocked. The second case is measuring squeezing case.

The measurement is done with network anylizer with bandwidth of 400Hz, averaage is 100 times. The measurement starting time is put as legend.

The result is attached and we could see that this noise varies between -127 to -120 dBVrms/sqrtHz at 10Hz.

As we reported in elog 2046 and 2049, we had issue of filter cavity correction singal drifts. According to the experience of Raffaele, this is due to the tide and can be corrected if we send portion of correction signal to the filter cavity end mirror.

Today I tried to do this, I fed portion of correction signal back to filter cavity end mirror at UTC time 2020-02-20-05:23 and then the correction signal becomes very stable. Up to now, the lock has already lasted for one hour.

Aritomi and Yuhang

We set up three temperature monitors and one humidity monitor around our bench. The position for the moment is shown in the attached first three pictures.(One is localted on the wall of cover and is beisde AMC, another is under the step and on the ground inside clean room, the last one is just beside PR chamber)

We have monitored these parameter for two days already. The conclusion is

Temeprature inside the bench cover is changing between 25deg to 26deg. Temperature on the ground(inside clean room) is changing between 24.3deg to 25.3 deg. Temperature beside PR vacuum chamber is changing between 23.3deg to 24.3deg. All of these temperature measurement go up and down with daytime and nighttime. The minimum of temperature is around 8am and the maximum is around 6 or 9pm. The difference of maximum inside and outside the cleanroom maybe related to manual work.

The humidity changes as well but it doesn't follow the day or night time.