NAOJ GW Elog Logbook 3.2
1-Setting of switches on the front panel:
* The differentiator shall be disabled on the front panel in setting the switch on "OFF".
* The switch INV/NON INV on the front panel, shall be set on INV.
2-Setting of the 8 straps on the board:
Low-pass filter, Notch filter 1 and notch filter 2 are activated on the board in setting strap on connectors P7, P8 and P9 (3 pins) between pins 1 and 2
* The transmission signal is negative with a peak at about -1V.
It shall not be inverted: the strap on connector P4 (3 pins) is set between pin 1 and 2.
The threshold level must be tuned to -0.5V (THRESHOLD OUT).
* Strap is set on connector P11 (3 pins), between pins 2 and 3, in order to activate the sample-and-hold on the triangular signal, on the locking.
* Strap is set on connector P3 (2 pins) to connect the triangular signal to the output stage.
* Strap is set on connector P2 (3 pins), between pins 1 and 2, for test purpose.
To check low-pass filter, notch 1 and notch 2 filters (in scan mode) between TEST IN and TEST OUT. For this test the differentiator, shall be set on "ON" (not intuitive but important). After this test, the differentiator shall be disabled the front panel in setting the switch on "OFF".
* Strap is set on connector P1 (2 pins), in order to be able to tune the offset.
3-Modification of components:
* Integrator 1/f: corner frequency changed to 2.2 kHz
Capacitor CMS 1206: C38 = 33nF
* Integrator 1/f2: corner frequency changed to 220 Hz
Capacitor CMS 1206: C26 = C33 = 330nF
* Low-pass filter: cut-off frequency changed to 2.2 kHz
Capacitor CMS 0805 : C45 = 2.2nF
Resistor CMS 1206 : R59 = 33k
* Notch filter 1: notch frequency changed to 6.6 kHz / quality factor changed to 3.6 (measured)
Capacitor CMS 0805 1% : C49 ; C50 ; C51 ; C53 = 2.2nF
Resistor CMS 1206 : R65 ; R66 ; R67 ; R68 = 11k
Resistor CMS 1206 : R73 = 2.7k
* Notch filter 2: notch frequency changed to 19 kHz / quality factor changed to 0.9 (measured)
[Capacitor CMS 0805 1% : C60 ; C61 ; C62 ; C63 = unchanged (560 pF)]
Resistor CMS 1206 : R79 ; R80 ; R81 ; R82 = 15k
Resistor CMS 1206: R89 = 13k
* Gain adjustment (G): Gmin = 0.054 / Gmax = 2.15 / Gtyp = 0.3
Resistor CMS 1206: R33 = 4.3k
[Pierre, Yuhang, EleonoraP]
Since we noticed the not efficient gain margin, we decreased the gain of SHG servo to 2.25. Then the OLTF is measured and attached in figure 1.
After modification of poles and zeros of GRMC, we implemented it. The important parameter is the transmission threshold which is -1.3V. The measured OLTF is attached as Figure2.
Notice: Always remember to give the high voltage driver an offset of roughly 20V by turning the knob on the front panel of it. This is because there seems a lower limit of high voltage driver.
[Miyakawa-san, Eleonora]
We wanted to install the new hard disk brought by Miyakawa-san (version RTS- 3.1.1) on the new PC.
We found a problem with the USB driver that could not be solved, so we were not even able to use the keyboard and to log in. Miyakawa-san suspected that the problem comes from a driver incompatibilty.
We will investigate if it is possible to solve it.
Anyway we replaced the hard disk of the currently used standalone computer (which had the version RTS-2.8.3) with the new hard disk (version RTS- 3.1.1)
Everything seems to work fine but I still could not fix the problem I find using the "Fnc" block from the "LIGO part" library. Miyakawa suggested an alternative way to implement that function.
NOTE: we found out that we have to disconnect the timing signal in order to restart the computer.
[Pierre, Yuhang, Eleonora P.]
After the simulation of OLTF and the check of the SHG transmission signal level, Pierre realized the corresponding circuit. It includes the modify of 20 components(resistors and capacitors). These changes will be uploaded to our wiki.
Then we installed it on the NIM rack, we tested both manual and auto lock of it. It works very well and can lock SHG also pretty well. Then we measured the OLTF. The measured result is plotted and shown in the attached figure.
We can see from it that the phase margin is enough. However, the gain margin is not enough. Besides, there are several peaks around 20kHz have phase around -180. This may also bring some instabilities. So we decide to reduce the gain to reduce unity gain frequency from 3.55kHz to around 2kHz. So that we can have a better gain margin and also avoid the phase of peaks crossing -180deg.
Previously, I checked the behavior of PBS for splitting the laser beam into two beam paths.
At that time, the maximum transmitted laser power was 0.4mw, which was too low.
Eventually, it turned out that the PBS was Brewster PBS (see https://www.u-optic.com/Show/?id=177&siteid=2).
Though the incident angle should be 56deg, I put the PBS with 45deg incident angle which caused such low transmissivity.
I use BS temporary instead of the PBS, and transmitted and reflected power is 5.8mW and 12.8mW for each.
One thing I have to take care is current configuration cannot separate p- ans s- polarized light.
We received a new router that we will use to rapleced that used for DGS computers in TAMA.
The model is YAMAHA RTX830.
[Matteo, EleonoraC]
We installed the new computer that will replace the KAGRA standalone I/O chassis. We mounted it on the rack which is hosting all the DGS electronics (pic 1).
We inserted 3 DAC and 3 ADC board in the designated slots (pic 2). In the case of ADC we had to slighty modify the position of the board front pannel to reduce the distance from the board it self, otherwise it was not possiblle for them to fit the slots. (pic 3)
For the slots from left to right (looking from the front side of the PC) we have (See Pic 4 and 5):
1: DAC PCie 16A016-16-F0-DF S/N 180904 - 98
2: DAC PCie 16A016-16-F0-DF S/N 180904 - 133
3: DAC PCie 16A016-16-F0-DF S/N 180904 - 42
4: ADC PCie 16AI64SSC-64-50-M S/N 181013 - 10
5: ADC PCie 16AI64SSC-64-50-M S/N 181013 - 28
6: ADC PCie 16AI64SSC-64-50-M S/N 181013 - 14
Pic 6 shows the board connectors from the back of the PC
Pic 7 and 8 show PC front and PC datasheet.
[Aritomi, Eleonora P.]
We checked error signal of OPO reflection with new PD. Attached picture shows error signal of OPO reflection. The amplitude is 112 mVpp as expected.
As you can see from second attached picture, error signal of OPO reflection (and its control loop) is not stable mainly because CC PLL lock is not stable.
CC PLL lock should be checked.
Also we need to optimize the control filter.
1/ Every LEMO connector was tested successfully on the 7 SERVOFILTER modules
2/ The current consumption was tested successfully on the 7 SERVOFILTER modules:
* +12V: 130mA
* -12V: 85 mA
3/ The 5V supply was tested successfully on the 7 SERVOFILTER modules
4/ A closed loop test was performed successfully in injecting a perturbation on the RAMP IN input (square signal, 200 Hz, 1 Vpp, 5.1V offset).
The SERVO OUT output is connected to the input of a RC filter (R= 150 Ohm, C= 1 microFarad).
The output of this RC filter is connected to the ERROR IN input.
Thus, a pole is created (fc = 3.8 kHz - 42.7 Ohm/ 1 microFarad)
Tests performed successfully:
- SCAN MODE: the SERVO OUT was observed on the 7 modules to check that the triangular signal is correct
- Manual and Auto locking modes
- 1/f and 1/f3 filtering modes
- differentiator filtering (ON/OFF)
- ENABLE IN function
Remark: in Auto mode locking, the THRESHOLD is tuned in order to have -2.2V a voltage level on the THRESHOLD OUT output.
A voltage level Vth is injected on the TRANSMIS. IN input.
Vth < 2.2V => unlocked
Vth > 2.3V => locked
The ENABLE OU and LOCKED MON out outputs were checked:
LOCKED => 5V level
UNLOCKED => 0V level
Conclusion:
After transportation, the integrity of the 7 SERVOFILTER modules was checked successfully.
See attached pictures files of the 7 SERVOFILTER modules.
[Matteo, Eleonora C.]
In order to have a larger error signal for the CC in reflection from OPO we have modified the15 MHz resonant TAMA PD (S/N00Z405) to increase the gain of the RF channel.
We have modifiided the RF amplication stage as follow (pic 1 and 2):
R1 = 1K (before it was disconnected)
R2 = 9.1K (before it was shortcutted)
The gain has changed from 1 to 10.
The TF of the PD measured before and after modification are shown in pic 3.
The TF was measured by injecting the signal directly on PD pin (A), so they don't match those reported in (http://tamago.mtk.nao.ac.jp/tama/ifo/general_lib/circuits/011217_PD/011217_PD_fit.pdf) but they are just a relative measurements.
[Aritomi, Eleonora P.]
P pol PLL frequency where parametric amplification is miximized is 207 MHz with 49mW green.
We measured parametric gain.
Amplification: 2.92V, De-amplification: 120mV, without green: 312mV
The parametric gain is 2.92/0.312 = 9.4. It was 21.6 before.
We changed OPO temperature around 7.178kOhm, but it doesn't change so much.
Tomorrow we'll investigate this problem.
[Aritomi, Eleonora P.]
Today we recovered visibility and shot noise.
For the alignment of LO in AMC, peak is 8.08V and mismatch is 28mV+3.6mV, which means mode matching is 99.6%.
We maximized BAB transmission from OPO by changing p pol PLL frequency. It was 291MHz without green.
For alignment of BAB in AMC, peak is 568mV and mismatch is 18.4mV, which means mode matching is 96.9%.
Then we measured visibility. Max is 7.76V and min is 2.4V, which means visibility is 0.528. Power of LO is 1.2mW and BAB is 0.107 mW. So theoretical visibility is 0.548. Mode matching estimated from visibility is 0.528.0.548=96.4%.
After alignment of homodyne, we got shot noise above 1kHz.
Tomorrow we'll recover squeezing and check shot noise level with CC.
[Eleonora P., Eleonora C.]
We wanted to investigate the system performances after reducing CC power of a factor 10.
We locked the PLLs, the OPO and checked the CC error signal in reflection from it.
We could see a demodulated signal (at 14 Mhz) when the green phase was scanned (10 Hz 1.5 Vpp as last time) but the amplitude is quite small: ~10 mV. (see pic1)
The power of the CC beam (with OPO locked) just before the homodyne BS is 11.5 uW.
Then we wanted to proced with the shot nosie and squeezing measurement but we found out that the visibility (of LO and BAB) was very low. So we confirmed that the aligment of the two beams into the AMC was poor but we managed to recovered it with some difficulties. (We followed the usual procedure: first we align LO with the two steering mirrors before AMC then we align BAB with the steering mirrors on the squeezing path after PBS.)
Current status:
- BAB flipping mirorr is on (BAB injected)
- Homodyne flip mirror is OFF to check alignement to AMC
- BAB and LO are aligned into AMC but visibility has to be re-measured
NOTE: We discovered that the fibered PD used for CC PLL monitor has a broken part (see pic2). This part is used to connect the PD the power supply insead of using battery. Since we don't have any spare we replaced it with a battery. Before leaving, we took the battery off to avoid consuming it. Remember to put battery in the PD if you want to check the CC beat note.
I just measured there are 3uW of p-pol is going also into homodyne.
[Eleonora P. Eleonora C.]
=SHG=
IR input = 0.6 W GREEN output = 0.2 W Efficiency ~30%. (as last time, entry #1180)
SHG coupling 80%
It seems that the highest HOM is 2nd order and cannot be easily removed improving alignment (has matching changed?)
Note that sometimes HV driver for SHG gets stuck and need to be switched on and off.
=GRMC=
Max power transmitted when MZ is on the bright fringe = 80 mW
GRMC coupling 90%
A reference value for GRMC transmission: 49 mW corresponds to 1.78 V on the signal from PD in trasmission (TRA GRMC)
[Aritomi, 2*Eleonora]
We have realigned the CC and ML pick-off into the fibers:
ML: 1.4 mW → 0.19 mW, coupling: 0.19*2/1.4 = 27%
[Aritomi, Eleonora P., Eleonora C.]
Yesterday we have realigned both the BAB and the CC beam. After the difficuties found last week we decided to put a beamspliter in transmission from OPO and check the beam both with camera and PD.
The aligment was quite easy to achieve but even in a good condition we could see some residual scatteerd light coming from OPO in the camera.
We had a coupling of about 85% for BAB (pic 1) and 95% for CC (pic2).
The BAB is now injected using a magnetic flipping mirror (http://www.1md.co.jp/fbp1000s_E.php). (pic 3) We checked the reproducipility of the aligment after taking off and putting back the flip mirror. It is quite good (but not as good as reported in the datasheet (http://www.1md.co.jp/fbp1000s_E.php)): anyway a clear FSR with highTEM00 can be seen and the optimal condition is easy to recover.
Finally we had re-aligned the CC and Main laser pick-off into the fiber.
NOTE: Please when putting on and off the mirror follow the procedure in pic 4.
[Aritomi, 2*Eleonora, Matteo]
This is work on 8th last week.
First we replaced a flipping mirror for BAB. Then we changed the position of OD1 from f=125mm lens to before the flipping mirror as shown in attached pictures to reduce only CC power.
CC power before OPO with OD1 is 11.6mW and 121mW without OD1.
Then we tried to align BAB, but it was difficult. We also found that CC was misaligned a lot, too. P pol was fine. We checked the beam shape of transmission from OPO with a camera as shown in last attached picture. Apart from usual beam spot on the right of the camera, there was scattering light from OPO on the left.
We'll align BAB and CC this week with camera and PD.
