NAOJ GW Elog Logbook 3.2
Note to switch on alert light after turning on the laser.
Some tips from Yuefan
1.Remember to switch off the piezo driver of SHG, otherwise it will drift away from the right position. It's located in the right side of the third shelf counting from the bottom.
2.The Laser power we use is around 1.2W, the value can be adjusted by the left knob on the panel.
The procedure of using the control system(the computer is always on, the system is based on Labview)
1.The document is under Project Explorer, the name is Eleonora-e-Manuel.
2.Control and demonstration panel:
BS and PR: under SAS_NM2b, the name is: tele12_control.vi
IM: under SAS_NM1b, the name is: IM1_contol.vi
Em: under SAS_EM2a, the name is: EM_contol.vi
3.Measurement of transfer function:
PR and BS: under SAS_NM2b, the name is: advtransfer_functions_tele.vi
IM: under SAS_NM1b, the name is: transfer_functions.vi
EM: under SAS_EM2a, the name is: transfer_functions2.vi
4.The local control and mode-matching of each mirror:
After opening each document, we can see clearly the mirror we want to control. Just by clicking on the dark green button, we can close the control loop. There is a very important parameter, offset. By setting different value, we can get a different orientation of this mirror. By this way, we can adjust the mirror to the right place. The right place can be checked by confirming the reflecting reference laser is on the right remark. The remark is made by Yuefan. For example, there is a remark on the viewport of BS tank. If the reference laser is reflected to this point, it means that the PR is in the right position.
For BS, we need to adjust its position slowly. We can do it with the help of diaphragm. Because we have several diaphragms in TAMA's arm pipe(between IM and EM).
After adjusting BS, we should see some high order modes in the CCD. Then we need to align the optical layout. There are two kinds of higher order modes. By adjusting BS and PR, we can eliminate the Hermite-Gaussian mode. However, the Laguerre-Gaussian mode is caused by the waist mismatch. We can solve this by moving two lenses behind AOM.
5.Try to find the reason of BS high motion of microrad
(1)The power supply. We usually stabilized voltage supply for the laser. Because PR's control is good enough. We changed their power supply, and then we find nothing changed.
(2)The data acquisition system. By changing the connection port of BS, we find nothing changed. Because we use the same data acquisition for all the data taking. PR doesn't have problems means other ports is fine. So no problem in data acquisition system.
(3)Output connection. We check the difference between using 50 Om and not using. This connection noise comes from the resistance thermal noise, we also call it Johnson noise. Johnson noise is proportional to the resistance. But, there is no obvious difference. Or just 0.6 in motion of microrad. (It's hard to read because the motion of microrad keeps changing.)
So we think the problem comes from the Laser.
6.We check the saturation level of each actuator
By changing the offset, we check the output. Beyond a certain value, the output will keep increasing. This is the saturation level because the actuator cannot offer enough force. Details are shown in the attached photos.
So we need to read the sum value first and type it into our Labview program. This will make sure the right normalization.
7.Taking transfer function
We should note that the transfer function is measured in mechanical part(actuator part). So we insert the noise in port NOISE2, which is white noise. And then we should make sure the loop is closed. And we have a good PSD angle. Finally, we can start measuring.
Besides, there is another influencing factor. It is the driving matrix. We should take care.
8.Checking the PSD angle to reduce the coherence.
By changing PSD angle, we find a good angle to have the least coherence. The result is plotted and attached. But Yuefan said the optimal PSD angle will change overtime.
Measured the power of the probes with and without the surface calibration sample to check the transmission of the sample at different wavelengths
633nm: 2.87 mW -> 1.89 mW Transmission= 65%
1310nm: 24.8 mW -> 16.6 mW Transmission= 67%
----------------------------------------------------------------------------------------
To check linearity of the Si PD, I measured the HeNe power and changed it putting some filters while measuring the PD DC signal
2.87 mW -> 6.58V
2.44 mW -> 6.52V
1.45 mW -> DC=4.28V
1.22mW (2 filters) -> 3.4V
328uW -> DC=0.88V
Result: the PD is saturated, see plot
Manuel, Kuroki
We measured the beam waist of the probe at the detector.
We used a blade and a vertical micrometric stage to move the blade with steps of 0.5mm.
We did this for the 1310nm laser using a power meter for infrared lasers (and repeated the measurement twice).
Kuroki-san made the fit of the data with a gaussian cumulative function erf(x/c) where x is the blade position and c = w/sqrt(2)
Result: w = 2.57±0.05 mm
For the HeNe laser the power is too low, so the resolution of the power meter was not good, so we used the smaller one. The small one was too small for the size of the beam, so after the blade we put a lens (f=19mm) to focus the beam on the power meter.
Result: w = 2.40±0.05 mm
During the last week of September, the Infrared Mode Cleaner has been assembled in clean room using the same screw size used for the Green Mode Cleaner ( a bit different than the design due to availabililty and to avoid breaking).
To differenciate the 2 modes cleaner, one can look to the BMC connector : the blue one corresponds to the infrared and the yellow one to the green.
We might still need to make a holder for the BMC as we did for the green mode cleaner to avoid stress on the wires.
We were able to test this mode cleaner on the infrared path on the table and found a finesse of the order of 280 below the expected value given by the mirrors reflectivity (we expect between 312 and 520).
This might come from power and temporal fluctuations. We could see these fluctuations while testing the green mode cleaner but we were suspecting that they came from the power fluctuations coming from the SHG cavity.
These 2 fluctuations are reduced if we have more sampling points.
They are also reduced if the ramp frequency sent to scan the piezo actuator is reduced ( we could see a better stability by reducing the frequency from 20Hz to 5Hz).
We could also see some temporal shift between the ramp and the mode cleaner transmission.
It seems that this is due to some low pass filtering of the piezo driver because this shift can be seen between the output of the piezo driver and its monitoring output.
From the attached picture, it is visible that there is an excess of noise in the error signals. They look similar to the ones I observed when spikes started to appear.
The spike issue is reportend in entry 428 where I also give some reference values for the rms of the mirror motion in standard condition (where no spikes are present). The rms shown in the picture attached by Yuefan is about 7 microrad which is higher that the normal one (about 1 microrad). I'm not sure if the spikes are the only cause for the smaller range of the PR, anyway they make the local control very unstable and preventing the lock of the cavity.
It is good that PR magnet is back to life but I'm very surprised that it seemed not respondig for a period. Indeed something strange is going on...
Pic 2 (about BS), referred to in the text, is missing.
For the local control,while I was trying to change the matrix things, I tried again to test each magnet. It seems that the third magnet of PR which we thought it was lost, response to the noise very well. Then I test the magnet of the input, there is really one of the magnet lost. I am not sure what causes the no response of the PR magnet, the things I have changed is just the red beam position on all the aluminum mirror.
Yesterday I adjusted the PR local control again, so then I need also change the PSD angle, now the value is -0.045.
Also I rethought about the order we did everything, the reference was put after we realign the cavity by moving the picomotor. And after that we installed the AOM and the lenses, according to the reference we put on the bench and outside the PR chamber, now we already has the same alignment as before.But the range we can move the PR mirror is not large enough to put back the beam on the reference outside the BS chamber, which means also the PR range got smaller. The control of the PR and BS mirror was written in the same VI, so now I am a little bit suspect the program, but I am not able to check it in Labview.
There are something I cannot understand, should discuss with Eleonora later. The PR error signal has been put near 0 in both direction, when put offset of yaw at 0.9, the correction is -2.77, and if put offset at 1, the correction will suddenly saturate. For the BS, also for yaw if the offset is larger than 2, the correction signal looks like pulse.(pic 2)
From the attached picture, it is visible that there is an excess of noise in the error signals. They look similar to the ones I observed when spikes started to appear.
The spike issue is reportend in entry 428 where I also give some reference values for the rms of the mirror motion in standard condition (where no spikes are present). The rms shown in the picture attached by Yuefan is about 7 microrad which is higher that the normal one (about 1 microrad). I'm not sure if the spikes are the only cause for the smaller range of the PR, anyway they make the local control very unstable and preventing the lock of the cavity.
It is good that PR magnet is back to life but I'm very surprised that it seemed not respondig for a period. Indeed something strange is going on...
Pic 2 (about BS), referred to in the text, is missing.
We tried adjusting the distances of the parts in the IU to make the sharp image on the detector of a blade at 18mm after the crossing point.
We can make it a bit sharp. It is blurred because of diffraction.
We found the problem which the probe laser saturates the detector signal at 4V.
So, we put a neutral density filter (optical density = 2) in front of the laser and put a optical damp in order to remove the reflected beam by the filter. We measured a photodetector signal DC=2V, about half of the range (with laser current of 130mA)
We measured optical power ,which is 0.48 mW. (Before putting the filter, the power was 37.5 mW.)
After that, We measured a absorption signal of a reference sample.
We get it, but it have a low signal to noise ratio (AC=40mV). Also the phase shape of the scan is not as expected.
We think that imaging unit(IU) should be correctly aligned
Next step is that we should consider how to better align IU.
Maybe we can try a different position of the blade after the crossing point, and see if we get a better signal.
We did a few checks.
Firstly, we checked all the magnets, they all response when we inject noise.
Then, we put the pitch and yaw back to zero pretty precisely.
According to Eleonora, this problem could be related to the laser noise, so we turned off the laser for a few hours and turned it on again, nothing changed. Since the chamber is closed and it is very hard to see the red beam through the window, it is too risky to move or change the laser, so we did not do that yet.
We tried to get the transfer function when inject noise from yaw and pitch, and also the open loop one, compared it with the Eleonora's entry 481, they look exactly the same. Also in this entry, Eleonora mentioned that BS works better when only sent voltage to two upper coils, so from the begging, the driving matrix only have value for two magnets.
During the check of all the local controls, we found out that in the PR chamber, the red laser came out from the chamber was very weak. So we looked into the chamber and found out, on the last aluminum mirror, the beam was too low, so it missed the mirror and hit the mount. We tried to increase the beam height from the laser side by moving the screw of the mirror, in order to receive the beam on the PSD, we also increased the height of the PSD and mirror before it. After doing this, it seems the PR mirror has larger moving range. When we reached the limit of the PR mirror, the beam reflected from the backside of the BS mirror is almost reach the reference.
But still we cannot align the cavity under this 'better' situation, we turned off all the local control laser, to see if let if off for longer time, the situation will change or not. If not, our idea for the next step now is to move the picomotor.
The red mode cleaner has been assembled and tested, Marc will put the details later.
Members: Manuel, Kuroki
We aligned the imaging unit of the 1310nm probe laser. The IU translation stage is translating parallel to the beam (at 0.1rad with the pump)
The DC signal we got at the detector is about 4V. With a probe laser current of 200mA.
Next step is to adjust the distances of the parts in the IU to make the sharp image on the detector of a blade at the crossing point.
https://www.dropbox.com/s/6pim9uxovu1oium/bb43a17cec87c2681906293ca3bf7889.mp4?dl=0
Since last week, we tried to set up a working telescope to be able to lock the filter cavity with the AOM.
Our design is the following : one lens (L1) before the AOM, two lenses (L2 and L3) after the AOM and before the Faraday isolator.
With the configuration f1=100mm , f2 = 100mm and f3 = 150mm we were able to align the beam into the filter cavity but we saw Laguerre Modes.
By amplifying 5 times the Pound-Drever-Hall signal we were able to lock the cavity but this lock was really unstable.
Indeed, either the input beam in the PR chamber was at a good size but the reflected beam from the IM was too big, either the input beam was too small and the reflected beam was at the good size.
Despite many attents to slightly move the lenses, we couldn't manage to obtain a better beam size.
We also check the beam size on the bench which confirm that this telescop can't work.
We then tried to use a f2=100mm f3=175mm configuration. But this telescop couldn't work neither. We think it was because the beam wasn't collimated at all.
In order to be able to use a bigger value of f3, we decide to move the AOM before the sterring mirror on the bench which allows us to have a bigger range for our new telescop. Indeed we mesure a beam size of 1.3mm just after the beam-splitter close to L1. Because L1 shrinks the size of the beam, we can have a small enough beam size for the AOM between this beam-splitter and the sterring mirror.
Then we choose a f2=100mm and f3=250mm configuration.
With this telescop, we were able to obtain a good beam size on the 2'' mirror and send the beam trough the filter cavity.
We obtain transmission peaks around 1V (better than from the first telescop) and gaussian modes. By amplifying 5 times the Pound-Drever-Hall signal, we could obtain a quite stable lock of the filter cavity on the green with the AO
We check the power exiting the optical table and found out that the AOM first order efficiency is now 67%. We are now trying to improve this value as well as the beam shape ( a bit astimatic) which might come either from the AOM or because the beam size entering the Faraday might be a little big.
We put back a camera on the IR path on the end room optical table.
Because the green power is quite smaller than before, we are able to see the IR modes flashes.
When the filter cavity is locked on the green, it seems that the IR flashes disappear and we can see a lock on some higher modes. We need to check if it is reamaining green or really IR.
https://www.dropbox.com/s/6pim9uxovu1oium/bb43a17cec87c2681906293ca3bf7889.mp4?dl=0
Members: Manuel, Kuroki
- Tanioka-san helped us with the crane moving one clean-booth from the first floor to downstairs.
- We washed the prefilter of the clean booth and wipe the walls of the booth inside and outside.
- We stuck the mats on the floor at the entrances of the pre-clean area and put a hanger for the clean suits
- We cut an aperture to enter in the clean booth from the pre-clean booth
- Using some anti-static sheet, we covered the space between the pre-clean booth and the clean booth
- We installed an in-line filter for the air shower, between the membrane-air-dryer and the pressure regulator.
- We moved the electronic parts of the experiment in the pre-clean area (computer, display, keyboard, mouse, and lock-in amplifier) and reconnected the cables
- We wiped inside the clean-booth using wipers and sticky rollers.
- We cover almost all the unused optical table holes with the anti-static sheet. So when we make an air shower of the experiment, the dust doesn't come from the holes.
1. Taking the PR as an example, magnets 1 and 3 for pitch, 2 and 4 for yaw, we checked this by sending noise to each magnet. Also with the same method, we made sure that the magnet we lost is the 3rd one. The original driving matrix for yaw is [0,1,0,-1], the four numbers are the voltage send to 1~4 magnets, which means when we try to move yaw, we don't send any voltage to the magnets for pitch. So even we lost the 3rd magnets, when we send noise to yaw, we should not see the coherence in pitch we can see now.
2. When I tried to change the matrix from [0,1,0,-1] to [0,1,0.1,-1], the coherence should not change since we lost the 3rd magnet, but now it changes.
This situation now make me really confused, so I am not sure how should I change the matrix, so I just did some random try. The picture shows the least coherence I can get when sending noise through yaw.
So the aperture before the FI, the orders are not seperated enough there, the aperture after the FI, we can open a bit to let the beam go through it, but then the reflection beam will be cut. So finally we used the diaphragm we put in front of the window to do the fitler job, showed in the first picture.
One problem of this configuration is that all the orders pass through all the optics after AOM, actually since all the optics are aligned with the first order, so I think for the beam deformation maybe there will be no problem, but I am not sure if this will affect the FI.
After aligning everything and having a reasonable beam, we tried to lock the cavity again. Using the Stanford to amplify the PDH signal with gain of 10, the cavity can be locked again(pic 2), but not very stable. The alignment seems cannot improve more, so I think it is still have something to do with the mode matching.
Then I tried to change the matrix. The situation is when we send noise in yaw, we have high coherence in pitch. But actually the matrix we use for this situation, the coefficient of two coils for pitch is zero. So it seems a bit strange to have this situation. Then I tried to change the coefficient for the three still-working coils little by little, the coherence of pitch got smaller, but still there.
https://www.dropbox.com/s/nerrc4i4yx3bw8v/VID_1.MOV?dl=0