Yuhang, Eleonora

We installed 12 dB attenuator on the LO of the QPD2 demodulation board, as we did for QPD1 demodulation board 1 (see entry #2112)

We performed the same test and confirmed all the channels are working fine.

The LO was -6dBm, we notice that if we put additional attuantion the demodulated signal is slightly increased. For a LO of -21 dBm the demodulated signal reaches the maximum amplitude and it is a factor 1.2 higher that in  -6dBm case.

Anyway we will split -6dBm LO between the 2 demodulation boxes. The actual LO for each board will be -12dBm which is fine.

Yuhang, Eleonora

We report three additional failures of Anti Aliasing (AA) channels.

Symptom: ADC channel reads large, almost constant offset (order of ~1000 counts). The offeset is still present if we disconnect cables towards AA board -> it is not coming from the BNC2Dsub converter. The offset is reduced when we switched off AA board and disappear when we disconnect AA from ADC. 

This time the broken channels are:

- Ch.10 of AA0 used for INPUT YAW OPLEV (We realized it while charachterizing input oplev on 13/07)

- Ch.16 of AA0 used for END L SUM  OPLEV (not currently used, we know it is broken since several months)

- Ch.14 of AA1 board, used for  WFS2 Q3 (We realized it while charachterizing demodulators on 10/07)

For the moment I rearranged the model to used spare Anti Aliasing channels  for INPUT YAW OPLEV (chan 9-12 of AA0 moved to chan 25-28 of AA1).

We shoud understand the origin of such problem and solve it. According to Miyakawa-san It might be due to a ground-connection issue in the design of KAGRA Bnc2Dsub converter JGW-D1604781 https://gwdoc.icrr.u-tokyo.ac.jp/cgi-bin/private/DocDB/ShowDocument?docid=47

Eleonora and Yuhang

To have an idea about how much amplification is required for demodulated signals from QPD, we compared the QPD demodulated signal with DGS and network analyzer.

The test is done by taking signal from QPD2 segment1 with green beam reflected form the locked filter cavity. The beam was centered with galvo on the QPD and power was ~1.5mW

In the first attached figure, the in-phase demodulated signal (measured in DGS/network analyzer) is shown together with DGS ADC noise. We  see that the measurement in DGS and network analyzer agree with each other. However, in the measurement of DGS, the peaks are barely higher than the ADC noise.

In the second plot, there is a comparison of demodulated in-phase/in-quadrature signals.

The large peak at 88 Hz is due to galvo loop, we adjusted the gain of galvo  servo and it is reduced.

Yuhang, Eleonora

Some measurement on the local control noise was done, in order to compare their performamces to those of AA. 

This is for the input mirrors. The calibration in rad/count is (reported in entry #1874) used for this measuremnt is:

We recall that that Input and end mirrors oplev use commercial PSD (thorlabs PDP90A) and the signal is amplified by a factor 100 with a SR560 and filtered with a 2nd order lowpass with cut-off frequency 100 Hz.

PIC 1:  Piitch signal compared with ADC noise and dark noise (the noise measured when the laser is switched. Factor 100 amplification from SR560 still present)

PIC 2:  Yaw signal compared with ADC noise and dark noise (the noise measured when the laser is switched off. Factor 100 amplification from SR560 still present)

PIC 3:  Piitch signal compared with ADC noise and signal withoud the factor 100 amplification (red curve should be multipplied by a factor 100 to have the good calibration)

PIC 4:  Yaw signal compared with ADC noise and signal withoud the factor 100 amplification (red curve should be multipplied by a factor 100 to have the good calibration)

Matteo, Eleonora

We test QPD1 demodulator after adding the 12 dB attenuator on the LO path (see elog #2106) and confirmed that all the 4 channels were working fine.

We sent a LO and RF signal from DDS (-6 dB both). LO was at 78 MHz and RF was at 78MHz + 100 Hz. We could correctly seeing a signal at 100 Hz after the demodulation, shifted by 90 deg in the I and Q output. (see pic).

We tested other offsets (from 1 Hz to 10 kHz). For all these values demodulated signal has a constant amplitude fase offset between I and Q is always ~90 deg. We verified this for all the 4 channels.

Matteo, Eleonora

We put a reflective mirror after green FI so that the light is reflected back to the QPD2 using the standard path. The power was changed acting on the AOM modulation amplitude. 78MHz and 87.6 MHz modulation are switched off.

Pic 1: RF output of QPD2 Seg 1 in with different power level. 

Pic 2: Same of Pic 1 but zoomed around 78 MHz.

Pic 3: Peak height at 15.2MHz and 30.4 MHz as a function of the green power. (This is a real signals come from PDH of SHG)

Pic 4: Noise at 20 MHz as function of the green power. 

Measured data are reported below:

We used the following setting :

RBW: 300kHz.  VBW: 10kHz

Average: 10

Internal amplifier: OFF

Spectrum analyzer attenuator: 10 dB

When measuring the peak hight we used different attenuation levels (20/30  dB) to avoid saturation

Yuhang, Eleonora

We changed the BS used to split the reflection of green FI FC lock PD and QPDs.

Before we were using a BS T 10 (70:30 for s-pol) and we replaced it with a BS X 10 (90:10 for s-pol).

Note that we are using p-pol so the actual ratios are different.

In the previous configuration we had 2 OD (0.5 and 0.6). Now we use only one OD (0.6) and the power is ~0.35 mW and the power reacing the pd is ~150 uW

We need to check FC PDH signal.

Yuhang, Eleonora

Pic 1: RF output of QPD2 Seg 1 in with different power level. 

The 15.2 MHz modualtion was switched off. The SHG was set on resonance manually. For each measurment we double-check that the power was stable.

We put a reflective mirror after Green FI so that the light is reflected back to the QPD using the standar path. The power was changed acting on the AOM modulation amplitude.

For 10 mW of  green power we measured 1.8mA

Pic 2: Same of Pic 1 but zoomed around 78 MHz.

Pic 3: RF output of QPD2 Seg 2 in with different power level. This time the 15.2 MHz modulation was switched on. The EOM RF driving signal was - 9dbm in this measurement. (Note that DDS output (-6dBm) is usualy amplified by 14dB and attanuated by -6dB, in this case we divided the DDS output by 8 in the software (corresponding to -9dB)). We want to confirm how much modulation depth are we using for 15.2 MHz EOM.  I put a curve with modulation switched on (P = 10 mW) for comparison. The stuctures are at harmonic frequencies of 15.2 MHz. It seems there is a quite strong saturation effect. This is strange becasue this effect was not observed in e-log #2067 when the 15.2MHz modulation was even higher (not dived by 8 inside DDS software). [UPDATE: we found this is due to saturation of the spectrum anlayizer internal amplifier which was off in the previous measurement]

###################################################

We used the following setting (see Pic4):

RBW: 300kHz

Average: 10

Internal amplifier: On

We found that electronicn noise on seg 2-3-4 is ~ -67 dBm while for seg 1 is ~ -69 dBm.

Matteo, Yuhang, Eleonora

An attenuator of 12 dB was added in QPD1 demodulator between the LO amplifier and the splitter.

See attached scheme and pictures.

Now a LO of -6dBm (DDS output) shoud not exceed the maximum power allowed by the mixer.

What I Did

• Lock the laser to the folded cavity
• Did ringdown measurement for rough estimation of the finesse

First, I tried to lock the laser with the servo by adjusting the offset.
Then I could lock and transmitted power was 60 uW.

After that, I turne off the AOM in order to test the ringdown measurement.
The decay time was about 2.5 usec.
It should be noted that the decay time of diffracted beam power by AOM was about 100 nsec.
Therefore, the impact of AOM can be negligible.

From this result, the finesse can be estimated about 1.4-1.5*10⁴.
Next time, I will record the result and calculate the decay time by fitting.

Next Step

• Measure the OLTF
• Ringdown measurement test
• Thermometer installation

Yuhang, Eleonora

On 24/06 we quicky checked the output of demodulated signal QPD2 acquiring it with DGS.

The measurement was done by putting a reflective mirror after the last green faraday before viewport. The beam was centered into the quadrant. Power on QPD was 3.3 mW.

There was no additional amplification.

As shown in Pic. 1 the signal is dominated by ADC noise.

Pic.2 shows a calibrated spectrum of the ADC noise for the 4 channels used to acquire the demodulated RF signals.

I replaced the screws which were used to fix the input and output mirrors.
Then I tried to improve the alignment.
Finally, the transmitted power became about 180 uW.

Tomorrow I will try to measure the finesse and lock the laser.
Then try to measure the OLTF.

What I Did

Today I locked the laser to the fixed folded cavity.
After the alignment work, I measured the output power by the power meter.
The output power was 4.7 uW.
This may be due to the alignment problem.

Using this value, one can estimate the intra-cavity power.
Since the transmittance of the mirror is 0.017 %, the intra-cavity power can be estimated about 28 mW.
There is room for improvement.

Next Step

I am planing to improve the alignment of the mirrors and try to increase the output power to 100 uW.
Then I will try to lock with pumps on.
Also, some pedestals are to be replaced to vacuum copatible ones.

Log on June 23, 2020

What I did

First, I tried to lock the laser with turbo pump on.
Actually, the lock was not stable and it may need suspension to reduce the vibration.

Then, I inserted the spacer and installed mirrors to compose the folded cavity.
The procedure was as usual.
This time I removed the block between the spacer and mirror holder to reduce the cavity length as shown in the picture.
Thanks to this, I could see the transmitted flash as shown in the picture (red line).
Still transmitted power is low and need to be improved.

Next step

I will improve the alignment to increase the transmitted power.
Then I will try to lock the laser to the cavity.

Log on June 19, 2020

What I did

I installed the servo which was from TAMA and modified.
Then I tried to lock by adjusting the gain and the offset.

Result

I could lock the laser to the cavity though the lock lasted only 10 sec or so.
The transmitted beam power was about 160 uW.
As the transmittance of the mirror is 0.017%, the intra-cavity power is estimated as about 1 W which is a reasonable value.

Next step

I am planning to lock the laser under the vacuum pressure to check whether I can or not lock when the turbo pump is on.
Also I will measure the open loop TF to see the phase margin and so on.

Takahashi, Yuhang, Eleonora

We replaced the TMP near NM2 from TMC1000MC (broken) to STP-1003 (new).

1. Removed the screws for the ICF253 flange. Dismounted the old TMP.
2. Mounted the new TMP. Set the rotational direction of exhaust to 225deg. Fastened the screws for the ICF253 flange. Final torque was 250kgf/cm.
3. Connected the dry pump (DSP250). Started the dry pump. Opened the angle valve.
4. Turned on the controller. It showed a fail due to different combination between the TMP and the controller. Reset the controller, then the controller went to auto tuning.
5. Started the TMP. The TMP went to normal operation.
6. After monitoring for two hours, opened gate valves to the 300-m pipe. The pressure became 4x10^-7mbar.

Yuhang, Eleonora

During the "ADC trouble" we found out that PR oplev PSD was not working fine: 

- SUM channel was giving a negative output

- Y channel doesn't respond to vertical movement of the beam position on the sensor

We replaced the PSD with a spare one.

[Note that this spare was previously used for BS oplev (see elog #1947) and we changed it as we suspestected it was noisy but after the replacement we kept oberving occasional noise floor excess in BS so we confirmed that it was not PSD fault.]

After the replacement I mesaured the TF of PR (pic1) and found that it perfeclty overlaps to the previouse one, so I didn't need to modify the gain of the damping loops.

We suspect that PSD falilure can be connected to the AA burnt card. Investingations are on-going.

Yuhang, Eleonora

We removed the broken card from the AI module (see elog #2093) and we found that at least 3 opamp AD8622 were burnt. (see pic 1,2)

The evidently broken opamps are: (see pic 3)

CH2 -> Oplev PR X

CH3 -> Oplev PR SUM

CH6 -> Oplev BS SUM

CH3 and CH6 are the most damaged. They corresponds to the SUM of oplev PSD for PR and BS. We remark that this two PSD uses the same power supply. 

We also remark that PR PSD is not working fine anymore (in particular channel Y and SUM).

We asked Miyakawa-san who said they have never had a similar problem in KAGRA. 

Yuhang, Eleonora

I changed the cable connection in order to move the channels that were connected to the broken AA card (AA 0 1-4 and 5-8) to a spare one in another AA module ( AA1 17-20 and 21-24). See Pic1.

I modified real time model accordingly. See pic2.

We tested the ADC input 1-8 and they are working fine now.  

Power supply is also ok since we disconnected the broken card inside the module, two days ago. Pic 3.

While doing ADC test we realize PSD of PR Oplev has some trouble. We will investiange it later.

Yuhang, Eleonora

On Tue evening 9/06 we found out that some of the ADC channel were not working fine and one of the two power supplies was delivering the maximum current (3 A) and the saturation led was on.

I switched it off and on but after ~30 min it stated to saturated again. A strong smell of overheated plastic was present around the ADC/DAC modules. I switched it off for the night.

Yesterday after confirming that DACs were working fine we check the ADC channels one by one. Here what we found:

We suspected the first card of the AI module (channel 1-8). We extracted the module form the rack and we confirmed that the first card had a very strong smell of burnt but we couldn’t individuate the broken component (we coudn't remove the card from the module to check the bottom part). We disconnected the card from the alimentation and reconnect again the AI module to the power supply. No saturation occurred since then. We will consider how to replace it. 

Note that ADC channel 11 even if is not connect to the broken card shows a strange feature (see pic 3). This channel is now used to acquire OPLEV INPUT SUM and it is not used in any control loop but the problem should be investigated.