NAOJ GW Elog Logbook 3.2
Participants : Marc, Yuefan, Yuhang
Here is presented the FC scan.
Compared to entries 771 and 775 few corrections have been implemented.
There was a mistake in the first estimation of the mode-mismatch (2 order 0 have been taken into account as well as the 2 sidebands).
There was another mistake done while doing the fitting : Now we used 1 Airy function defined by the 2 order 0 and then only changed the normalization value (T0).
We obtain the following results :
r : 0.9992
FSR : 125 s (the AOM frequency modulation have been changed but we still have FSR=500 kHz)
T0 (normalization factor) : 3.6374 0.0164 0.0604 0.0274 0.0119 0.0262 3.3774
x0 : 65.62 89.52 115.155 115.208 164.78 164.88 190.62
The mode-mismatch was evaluated using the ratio sum of higher order modes T0/ fundamental mode T0.
This leads to a mode-mismatch of 3.9%.
Few things to notice :
The position of the sidebands needs to be confirmed
While modulating the AOM frequency to realize the scan, the beam exiting the AOM is tilted which leads to the different heights of the fundamental order peak (?)
According to the signal we send to DDS board, the modulation frequency of we give to EOM is 15.2MHz. The FSR is 0.5MHz.
15.2/0.5=30.4
So we will have additional 0.4 of FSR. This exactly explains the result we have on the oscilloscope.
The total amount of power not coupled inside the FC is composed of the mode-mismatch and the sidebands.
The sidebands maximum are respectively : 0.1462 and 0.1422 mV (this takes into account the background value).
This means that 11.84% of the light is not coupled inside the FC.
Participants : Marc, Yuefan, Yuhang
We did another FC scan.
Using 2 differents scales to see the fundamental mode and the other higher order modes we could have a quite good precision on every order peak.
By fitting every resolvable peak by an Airy function and using the ratio Area of fundamental mode / Area of fundamental mode + higher order mode,
We found that 95.24% of the light was coupled into the fundamental mode of the FC.
Participants : Marc, Yuefan, Yuhang
Today we continue the alignment of the collimator and the fiber.
Because last time wasn't successful, we decided to use a 1550 nm fiber laser and try to align it at the opposite of its use (inject the laser with the fiber and try to align the output of the collimator).
If we could align it using an IR card, it was not possible to use a power meter nor the beam profiler due to this wavelength.
We decided to go back the first way of aligning this collimator (laser at the input of the collimator and check the power at the output of the fiber).
We decided to use 2 sterring mirrors (actually one mirror and one 98/2 BS).
We got the following power
output of the Auxiliairy laser : 0.873A -> 29.7mW
After 1 mirror + BS (98/2) : 28.7mW
After collimator : 28.7mW
After the fiber 4.5 uW
We can see some light at the output of the fiber (on IR card, power meter and fiber photodiode).
One problem is that as soon as we touch the fiber or the collimator, we see power fluctuations.
This means that between each time we screw we need to wait for few seconds...
It should be useful to find a way to fix the fiber.
Participants : Marc, Yuefan, Yuhang
In order to study the coherence between the input laser IR power and the reflection from the filter cavity, we installed a PBS on the IR injection path and a photodiode at its reflection.
We studied 3 cases :
1) Both IR and Green resonants inside the FC (greenir.png)
2) No beam resonant inside the FC by misaligning the EM and making sure that there was no attempt to lock on the servo (no.png)
3) Only Green resonant inside the FC by detuning the EOM frequency (green.png)
We were expecting to find similar results between the cases 2 and 3 which is not the case.
We were expecting to find coherence at high frequencies due to power fluctuations only but it only appears in the only green resonant case (maybe the coherent length of IR is too short?)
The no beam case shows clearly a 200Hz peak (ventilation of the clean booth) and a 600Hz peak (turbo pump of BS).
The error signals for green and infrared account for the closed loop laser frequency noise filtered by the pole of the cavity (which is different for green and IR).
In the fist attached plot, the error spectra has been divided for the corrispondig pole in order to go back to the close loop laser frequency noise.
freq nois = err sig * ( sqrt( 1+(f/f0)^2) with f0 = 55 Hz for IR and 1.45 kHz for green
The two curves obtained shoud be coincindents. The discrepancy (about a factor 2.5) suggests that there is maybe an issue with the calibration.
Participants : Marc, Yuefan, Yuhang
Last Friday we did a scan of the FC for IR.
[ In order to obtain a good beam position, the BS pitch correction is close to saturation (9V). It might be useful to move some picomotors in order to have a better beam position and a lower correction]
The scan was performed by modulating the frequency sent to the AOM.
We used the following parameters : 1 MHz for the half peak-to-peak value over a half-period of 2 mHz (the lowest value permitted).
[The "half" values are used as they are what we provided to the AOM frequency generator]
We obtained the result presented in the first picture (FC_scan.png).
Few points to notice :
there is a symmetry at the second peak (if we zoom in we can distinguish 2 peaks)
The vertical scale was not good enough to have a good resolution of the small peaks -> on Monday we will do another scan with 2 scales : 1 to resolve the fundamental mode and another one for the higher order modes.
Anyway, as a preliminary analysis of this FC scan, we fitted the first 2 peaks with an Airy function (fig. FCscan_fit) which gives us a FSR of 62.57 s.
By doing : 1MHz / 2mHz * FSR[s] = FSR[Hz] we find a FSR = 500.560 kHz which is in good agreement with the logbook entry #668.
Another preliminary analysis was trying to estimate the losses due to coupling with higher order modes.
The fundamental peaks were fitted with a gaussian function while the (poor resolved) small peaks were fitted by triangles.
By comparing the ratio between the areas of fundamental mode / fundamental mode + higher order modes, we found a mode-mismatch of roughly 10%.
This analysis will be more precisely performed on Monday.
We are still investigating the reason for the calibration problem in the bulk absorption that gives a factor of 3 between the measurements done with our experiment and at LMA or Caltech on the same samples.
In order to understand how the interaction length (the crossing area of probe and pump) affects the measurement, I made some simulations of the scan changing the thickness of the samples and keeping constant the absorption/cm rate.
Before measuring the coating absorption of LMA samples and crystalline coating I wanted to test the surface reference sample in order to check if it had any damage. So I made a map of it.
The map shows a regular pattern of absorption that oscillates by a factor of 2.
I made some checks to understand the nature of this strange behavior. I report the steps in the attached pdf slides.
I'm going to look for a better alignment
Participants : Marc, Yuefan
To have a lower beam divergency (so easier working conditions) we installed a lens f=200mm 6 cm after the output of the auxiliary laser 1.
We also increased the laser power to 44.3 mW (0.9 A for the pump).
By carefully aligning the lens and the collimator vertical and transversal position (X and Y) we obtain a good aligment on this 2 directions.
As the beam is quite diverging after the collimator, this was done checking the beam positions around 20 cm after the collimator.
By unscrewing the 3 screws moving the Z position of the collimator (its distance to the laser),
we reach a power of 44.1 mW at the output of the collimator meaning 99.55% transmission of the collimator.
This seems good enough to go on the others steps (installation of the fiber, add another collimator at the output of this fiber and check the beam size at its output).
We also found another holder for the second collimator meaning we have all the needed components for the next steps.
Here I attach the rms integration of the four error signal curves.
Participants : Marc, Yuefan
Auxiliary Laser 1
Today we installed the auxiliary laser 1 on the squeezer bench.
Compared to the lens on the IR path far from the FC, its output port is at 14 holes (vertically toward FC) and 4 holes to the left (toward the main) laser.
This position seems convenient to avoid to twist its power supply while letting some space for the future optics to be installed.
Using 2 mm spacers, the beam height at the output of this laser is 7.5 mm.
Pre alignment of the collimator
Because of the divergency of the laser and the divergency of the not-aligned collimator, the beam at the output of the collimator is quite diverging.
In order to follow the alignment proposed by Thorlabs, we need to measure the output power of the collimator and incremantally find its maximum.
The powermeter had to be placed few cm after the collimator making this task quite painfull...
We could reach an output power of 10.35 mW (compared to the 11.3mW of the laser). However, a sad mis fixation of the collimator post made it moved quite a bit.
Tomorrow we will add a converging lens between the laser and the collimator in order to have a lower divergency of the beam and better work condition.
After the test of Mach-Zehnder, we need to install it in reality. But we change the original design. Thanks to the optocad code of tomuta-san, I revise it and do the simulation.
We have considerations:
1. The stable lock of filter cavity. We decided to use MZ only for the Mode cleaner and OPO.
2. The focal length of lense is limited. So I asked yuefan to give me the list of lense we have. I select focal length of 50mm and 75mm.
3. We need to adjust the position of these two lenses to have a good mode matching.
The change can be seen in the attached picture 1. We will start to install it tomorrow.
The result of previous comparison is wired. So we decide to change back the old servo and measured error signal again. At the same time, we also did calibration again.
Note here, the calibration method for green likes entry 750. The only difference is that we consider frequency much higher than unity gain frequency. Then we got result like attached picture 1 and 2.
The calibration for infrared got from attached picture 3 and 4. For the problem of saturation, we changed the demodulation phase.
Finally we got result as attached picture 5. We can make them overlap by multiplying a factor of 11 like picture 6.
As a result, we find the new servo make noise level lower than before. So now I put back the new servo, we can get the green transmission as 1.5V and very stable like picture 6.
Here I attach the rms integration of the four error signal curves.
The error signals for green and infrared account for the closed loop laser frequency noise filtered by the pole of the cavity (which is different for green and IR).
In the fist attached plot, the error spectra has been divided for the corrispondig pole in order to go back to the close loop laser frequency noise.
freq nois = err sig * ( sqrt( 1+(f/f0)^2) with f0 = 55 Hz for IR and 1.45 kHz for green
The two curves obtained shoud be coincindents. The discrepancy (about a factor 2.5) suggests that there is maybe an issue with the calibration.
Corresponding to the comment of Eleonora, the bandwidth of filter cavity for infrared is 114Hz but not 55Hz. Then I think we can explain the result (almost).
bandwidth=FSR/Finesse=500000/4355=114
I think there is a factor 2 missing in the formula: the pole of the cavity is FSR/(2*F) = 500000/(2*4355) = 57 Hz
- Can we identify the 4.9kHz resonance and maybe reduce it mechanically (e.g. by tightening screws)?
- Are the 600Hz noise features actual amplitude fluctuations coming from the SHG or are they added by the Mach-Zehnder? (This can be tested by putting the PD before Mach-Zehnder.)
- Is the low-frequency dark noise caused by ambient light or electronics? Can it be reduced?
It seems like the peak of infrared error saturates on oscilloscope from picture 3. Maybe we can put an attenuator for it?
Participaint: Emil and Yuhang
After change the new servo and new infrared demodulation board, we did a rough phase adjustment. We decided to make it as best as possible, so we changed the green and infrared demodulation phase.
For infrared phase, we change the demodulation phase and find a really small error signal. Then we add 90 degree to get a good phase. However, we cannot get a clear green error signal and we cannot see the difference when we change the phase. For the green locking, we did like this:
1. We lock green.
2. Change the phase until we can see the oscillation of error signal.
3. Decrease the gain till oscillation disappears
Because the gain of SR560 is 1 now, so we connect the demodulation signal directly to Rampotu servo. Now the gain is 4.5 on the Rampeauto board.
After change the demodulation phase, we measured open loop transfer function and error signal again. I put the result here.
(Note: this time I also attach the error signal before calibration)
It seems like the peak of infrared error saturates on oscilloscope from picture 3. Maybe we can put an attenuator for it?
Here is the 2nd version :
Changes :
OPO, homdyne and squeezed vacuum beam path have been added
For the article preparation as well as for future meetings, it could be useful to have a simplified optical scheme of the experiment.
Here we tried to do one following Oelker example ( Audio-Band Frequency-Dependent Squeezing for Gravitational-Wave Detectors ).
In the scheme attached to this entry is a preliminary scheme.
The "Not Yet installed" part should contain OPO PLL (?) and homodyne readout.
We haven't yet put the OPO, the PLL and the homodyne readout. The question being how to add them without making the scheme to difficult to read.
All the other main components are indicated.
Maybe it could be also useful to add the control loop?
Here is the 2nd version :
Changes :
OPO, homdyne and squeezed vacuum beam path have been added
Participants : Yuefan, Yuhang
When we tried to recover the FC lock, we had to act quite a lot on the BS control.
The pitch was saturating below -0.7 so we had to play with BS and IM yaw in order to reduce the saturation on the BS control while keeping a good beam position.
[ When the IM is misaligned it is really difficult to see the green transmitted beam because of another beam splitter has been installed on the green path in the squeezed bench]
We could finally performed a losses measurement still using the lock/unlock technique which gives us : 60.4 ppm +/- 7.3
With 2.54% misalignment and 0.25% mode-mismatching.
We also plotted the SHG stability over 1 000 s ("shgstability.png")
It seems to be quite stable around 1.5V even though some low frequency variations can be seen.
The last part from around 750s corresponds to the time we started to try to lock the FC.
It seems that some of the light came back towards the SHG.