NAOJ GW Elog Logbook 3.2
[Aritomi, Eleonora P&C, Yuhang]
Today we tried to improve IR alignment into filter cavity.
First we increased green power to 68mW. OPO temperature is 7.2 kOhm and p pol PLL frequency is 150 MHz. IR injected into filter cavity is around 13mW. Then we locked IR on TEM00 and maximized TEM00 by aligning two steering mirrors for IR. During IR alignment, green transmission was above 5000.
Since lock of green phase is not stable, we scanned green phase with 10Hz and 1.3Vpp. Then we scanned AOM frequency at 2 mHz. After the scan, we locked IR on TEM00 and HG10,01 and measured each power with dataviewer. We repeated alignment of IR and AOM scan twice. Attached pictures show cavity scan. IR transmission when IR is locked on TEM00 and HG10,01 is as follows.
After scan 1
| AOM frequency (MHz) | IR transmission | mode |
| 109.036 | 100 - 2000 | TEM00 |
| 109.43 | 100 - 400 | HG10 |
After scan 2
| AOM frequency (MHz) | IR transmission | mode |
| 109.03646 | 100 - 3800 | TEM00 |
| 109.43133 | 100 - 270 | HG10 |
| 109.432 | 100 - 400 | HG01 |
This scan seems strange and is not compliant with measurement when IR is locked due to green phase scan. Pic 3 shows TEM00 in scan 2. Given that linewidth of filter cavity is 100Hz and AOM scan speed is 8kHz/s, frequency of green phase scan should be higher than 80Hz. Actually we tried 1kHz once, but TEM00 peak hight became smaller.
Anyway now TEM00 is 10 times larger than highest HOM. It seems that mode mismatch is small. We'll continue the alignment tomorrow.
This afternoon, Matteo checked my experimental setup, and gave me some advices.
Following his advice, I measured beam profile.
I will upload the results tomorrow...
Today, Aso-san pointed out that the silicon mirrors may have higher finesse than we expected.
Therefore, I checked the spec sheet of mirrors, and the designed reflectance is more than 99.997%, which corresponds to more than 100,000 finesse, though we ordered as 50,000 finesse.
The manufacturer may assume that the loss (probably ~30ppm) reduces the finesse from ~100,000 to 50,000.
However, it may be almost impossible to see transmitted beam with such high finesse and loss...
Today, I re-started the installation of optics for double-pass AOM which is used in HOMs paths.
The beam is aligned roughly.
In addition, I cleaned silicon mirrors using First Contact.
Then I installed them inside the chamber, but still cannot see any flash.
Also I re-designing the cavity in order to make another set of mirrors which have lower finesse and high coating quality.
Firstly after the new measurement of the mechanical transfer funtion of GRMC, the notch filter 2 was modified as below:
The center frequency of the notch filter was tune to 9,47kHz without changing the Q factor (1 as previously).
C60; C63; C61; C62; : 560 pF (unchanged)
R79; R80; R81; R82: 30 kOhm (instead of 24 kOhm previously)
We didn't succeed to lock the GRMC, so I checked completely the filter section of the servo: it appears after measurement of the transfer function of the 1/f filter, that the corner frequency was displaced from 2,2 kHz to 9 kHz.
It appears that the C38 capacitor was damaged. The C38 capacitor was replaced by a 33nF capacitor and the transfers function was restablished.
We always didn't succeed to lock the GRMC. It appears that the error signal in scan mode has an amplitude of -20mv / + 30mV which is weak for the current design of the servo. So the hysteresis of the detection comparator (U23 - LM311DR) was reduced drastically from 10mV to 0.5mV (largely less than the amplitude of the error signal in scan mode).
The R108 resistor was changed to a 10 MOhm resistor (instead of 470 Kohm previously).
After this modification, the GRMC locking was succesfull and robust.
We tuned the gain (position 3 of the gain potentiometer) in order to have a gain margin of 10dB (at phase 0 deg) and a unity gain frequency of 1.6kHz (more than 50 degree of phase margin).
So the GRMC servo is now functional and we can go on the modification of GRMC and MZ servos for synchronized locking.
Error signal is 30 mVpp and conversion factor from V to rad is pi/0.03 = 105 rad/V.
[Aritomi, Eleonora P&C, Yuhang]
This is work on July 2nd.
First we increased green power to 60mW. OPO temperature and p pol PLL are as follows
| green power (mW) | OPO temperature (kOhm) | P pol PLL (MHz) |
| 60 | 7.19 | 150 |
Then we scanned AOM frequency and measured green and IR transmission with diaggui. Note that this data is slow channel and sampled at 16Hz. Setting of AOM is shown in Pic 1. FM is 2 mHz and deviation is 1 MHz (peak to peak: 2MHz). During this measurement, we scanned green phase with 2kHz, 700mVpp. We also took a movie of green and IR camera to characterize higher order modes. Starting time is 19-07-02-07-08-57 UTC. Pic 2 shows transmission of green and IR. Higher order mode is as follows.
| time (s) | mode |
| 22.75 | TEM00 |
| 46.81 | pitch HOM |
| 72.5 | HG10,01 |
| 97.94 | HOM |
| 147.4 | TEM00 |
| 197.1 | HG10,01 |
| 248.6 | HG11,02 |
| 347.8 | HG10,01 |
| 423.2 | TEM00 |
| 472.8 | HG10,01 |
Note that speed of scan is 2 MHz / 250 s = 8 kHz/s and measured FSR is 147.4 - 22.75 = 124.65 s which means FSR is 0.9972 MHz in AOM.
As you can see, green transmission is also scanned by AOM since AOM scan changes alignment of green injected into filter cavity. But this shouldn't be a problem as long as alignment of filter cavity is fixed. When you look at TEM00 closely (Pic 3), sampling rate seems not enough. We'll try with oscilloscope today.
Yuhang, Pierre, Eleonora P, and Aritomi
In the beginning, Matteo and Eleonora modified the TAMA PD by apply two resistors and this amplifies the AC signal. This activity increased the SNR and make it possible to lock the coherent control loop.
Then I and Chien-Ming make this amplification even larger again, but the amplification increased both signal and noise. After this, we found there is a bandwidth limitation of AC channel op-amp when we increase the gain.
So we decide to use a higher speed op-amp and try to increase a bit the gain and to see if we can increase SNR furthermore. This replacement was did based on the modified PD improved by Chien-Ming. There were problems while Pierre was soldering the new op-amp. The copper board pin was fallen! But, fortunately, the lost pins are the pins which are not connected to anything else. Also, the day when we just finished the modification, we tested it on the bench but we didn't see any coherent control error signal. Actually, we found the very similar thing(cannot see BAB be amplified and de-amplified after OPO transmission) on the next workday. I and Aritomi-san we did the measurement of parametric (de)amplification measurement again while scanning the green phase with frequency slowly. After that, we could see everything well. We also realized at that time, the characterization of best OPO temperature/p-pol PLL locking frequency should be done maybe once a month.
After the replacement, we characterized the dark noise and compare it with the old dark noise before replacement. We found an increase of dark noise, we think this may come from the rise of a large current bias on the op-amp.
Then we measured the free-running noise and error signal noise spectrum (residual locking noise). The results are shown in the attached figures.
In the future,
we will also try to play a bit with the amplification of the AC signal. And also try to reduce the current bias of op-amp.
Error signal is 30 mVpp and conversion factor from V to rad is pi/0.03 = 105 rad/V.
[Takahashi-san, Yuhang, Eleonora C]
Last sunday afternoon (30/06) we recovered TAMA vacuum, air conditioning, dehumidifier system and DGS after the planned shutdown on 29/06.
Some details on the procedure can be found in the attached PDF.
Note that while I was In the end room I heard a short but very strong noise (as a small explosion). Takahashi-san told me that it is likely to be due to the presence of air bubble in the dehumidification water pump during the restarting process.
green setting:
| green power (mW) | OPO temperature (kOhm) | p pol PLL (MHz) |
| 40 | 7.175 | 174 |
IR into filter cavity is around 1 mW. TEM00 is as follows.
| AOM frequency (MHz) | IR transmission (cnt) |
| 109.03683 | 100-130 |
With 40mW green it was difficult to see higher order mode with IR camera. So we decided to increase green power.
On 27/06 Miyakawa-san came to Mitaka to try to fix the keyboard problem with the new DGS computer. (Since the gentoo linux cannot reconize the USB driver the keybord doesn't work). He tried to access the PC through the network but network port, but the kernel was too old to recognize it. It is not easy to update kernel because the kernel was patched to be a real time operation system. We also tried several USB PCIexpress board whitout success.
Anyway the actual standalone PC can host 4 ADC/DAC cards. Only two of them are currently used. We tested one of the new PCIexpress ADC board on the current stendalon PC and it seems to work fine. We decided to leave it installed.
The old ADC card is currently stored on the shelf close to the DGS rack.
In the near future we might need to install it too to increase the number of ADC channels. We agreed with Miyakawa-san and Akutsu-san that in this case we can use some of the hardwere from the ATC DGS which is currently unused.
In particular, in order to be able to use a second ADC we need 1 timiing adaptor, 1 AA module, 2 BNC-> Dsub coverter, cables. This will allow us to have 64 input channel.
Miyakawa-san showed me how to include additional ADC block in the simuling Real time model (NB: remember to copy the ADC block from library and not from those already in the model.)
He also helped to fix some minor issues:
1) remote connection to the DGS (Finaly we can access the system from the office and even from home!)
2) take/restore snapshot button on MEDM
3) rotation angle block
Yuhang and pierre
This work is motivated by the difficulty of locking the GRMC loop.
We measured again the loop, we found the resonant peaks have changed relative to the original measurement.
The first peak was changed from 7.23kHz to 6.9kHz and the second peak was changed from 11.85kHz to 9.9kHz.
We have already investigated by changing the position of clamps. But the peak position was not changed. However, I didn't try to move the end mirror of the mode cleaner. The change of these peaks position is still not clear.
Aritomi, EleonoraP, Yuhang
As described in the title, we characterized more about coherent control loop 1.
The dark noise is measured when there is not light. And other situation is the same(for example, PD is on, demodulation signal is on). The measurement point is error signal.
The free running error signal noise spectrum is measured after we checked that green phased shifter scanning and coherent error signal was fine.
EleonoraC, Takahashi, Yuhang
Today we have power outrage in TAMA, we switched off the vacuum system and air compressor yesterday. The procedure is similar to entry 1138. The list of things we did is summarized in the attached pdf file.
Here I also attached some numbers we need to recover later.
| power supply for AOM RF amplifier | 24V |
| RF driving frequency for AOM | 109.03535MHz 7dBm |
| power supply for homodyne | 19V |
| power supply for TAMA PD | 12V |
After the implementation of CC1 mirror mount, we used the 1/f integrator and locked this loop(with unity gain frequency of ~2.2kHz). We measured open loop transfer function and the noise spectrum of the error signal. The result is as follows.
But the measurement of the noise spectrum when the turbopump was on/off was not so reasonable.
We decide to characterize this loop again. Including the measurement of free-running noise, noise while locking, PD dark noise and also the optomechanical transfer function.
We always need many people to check several checking points to align the filter cavity. Also, the channels for monitoring video signals from the end room were not enough(two channels were working at that time).
This means: (1) camera is not enough (2) channels are not enough
- For the camera, we set up a new camera to take real-time video inside PR chamber. By using this camera, we could check the overlap of injection/reflection beam.
- For the not enough channels, we found two cables(actually is fiber) were broken from west-south corner to control area(we can also call it 'short part' of fiber). We replaced them(Broken fibers are 1-13 and 1-14. 1-13 was replaced with a working fiber and the label was changed as well. Now the 3-11 short part is broken). Then we found one of the channels of board1 was broken after one-day work. We changed it to a good channel later.
In the end, we have four real-time videos on the screen and an additional one (first target) which can be monitored after we switch from channel A to channel B. The organization of four videos are as follows
| Second target | infrared camera(FC Tra) |
| green camera(FC Tra) | monitor inside PR chamber |
The connection situation at the end room is shown in the attached figure 1 and 2.
[Eleonora, Matteo]
Up to now we have been using only 16 of the 32 channels available in the ADC of the standalone DGS recently installed in TAMA.
Today we received from Kamioka the BNC->Dsub converter that we missed in order to use them all. We have installed it and connect 4 more cables from the AA to the converter (each of them as 4 channels) (pic1).
We borrow these 4 cables from ATC after asking Akutsu-san (pic 2). The cables are much longer than what we need (pic3) so in the future it would be good to change them with shorter ones.
I have modified the realtime model to include these additional channels and I verified that it is working fine.
In conclusion, we have now 32 ADC channels in our DGS.
According to Emil's thesis P. 42 figure 2.10 or P.48 figure 2.13 (b), 200 mrad of phase noise seems to degrade 15 dB of squeezing to almost 0 dB of squeezing. Our situation seems around 100 mrad of phase noise. Did you consider the effect of control bandwidth when you calculated rms phase noise? As you noted, it's better to measure the spectrum and integrate it within control bandwidth.
[Aritomi, Eleonora P]
You are right. We'll scan AOM frequency in the range of 1MHz.