Wow, this is awesome. Can you plot the comparison between the measured and expected OLTFs for both (or either) Moku cases?
At low freq, the measurement of OLTFs gets difficult due to high loop suppression and the estimation will give us an idea of how the loop can be improved.

I almost finished constructing the mode-matching breadboard today.

What I did on July 26th:

• Constructed the mode-matching telescope on a breadboard (Attachment  1).
  • Height of optical axis: 3.5" (from the bottom of the breadboard)
  • Used two lenses of f = 200 mm @ ~6 cm and ~24 cm from the surface of the collimator mount
    (Used JamMt to obtain the solution)
  • Tuned the position of the two lenses monitoring the resulting beam waist so that the waist becomes ~ 500 um
    (OMC eigen mode: 489.6 um (horizontal), 490.5 um (veritical))

Next:

• Measure the dimensions of the resulting mode-matching telescope
• Measure the beam profile of the resulting beam
• Estimate the upper limit of the mode-matching efficiency to the OMC

Supplementary

Output beam profile from collimator:
I measured the beam profile of the output from the collimator (Attachment 2).

Beam waist X: 44.3 +/- 0.5 um
Beam waist Y: 46.1 +/- 0.6 um
Waist position X (from the front surface of the collimator mount): 1.64 +/- 0.04 cm
Waist position Y (from the front surface of the collimator mount): 1.70 +/- 0.05 cm

I found that the resulting beam is not collimated but focused.
The model number is not written on the collimator so I was not able to find its specification.

I borrowed a lens (f=200mm, LB1945-C, Thorlabs) from the air BHD setup [Attachment 1 (before removal), 2 (after removal)].

I am going to use this lens for the mode-matching breadboard for BHD OMCs.

The OLTFS of AUX locking system when locked with the PDH box, and Moku:Go and Moku:Pro running a similar filter as that of the PDH box. The Moku:Go for some reason seems to have a very low UGF (~10.8kHz) compared to the Moku:Pro  (~21.4kHz), which is a bit lower than than the original UGF of the PDH box (~24.2kHz). With a better, optimal digital controller, it could be possible to get higher low frequency gain and bandwidth. Unfortunately due to time constraints and the interferometer not being available to work on for enough time, I was not able to start testing optimal controllers. 

Today, I managed to completely substitute the analog PDH box with the Moku:Pro. Earlier, I was just taking the error signal demodulated by the PDH box only. Now i can show that the Moku can source the oscillator, and using the mixer and low pass filter retrieve the error signal using the lock in amplifier instrument, and oass it onto the digitial filter instrument to lock the box. This could mean the entire PDH box can be replaced with a Moku:Pro/Moku:Go, which allows for easy modification and upgrades, with optimal controller design. Still, the lock was brief, lasted 10-15 secs, and the transmission was noisier. I could not get a noise reading yet because getting it to maintain lock  out of the box was a bit finicky,  might have to tweak some paramters of the lock in amplifier and pdh box... 
(I also tried the same with a Moku:Go, but the transmission was even more noisier.)

(Moku Pro transmission and setup)

(Moku:Go transmission)

I removed all the DB37 and IDC50 front panels from the chassis and retained only the 2 BNC feed-throughs.

I also removed the IDC50 breakout boards from the DIN rail and put 4 optical isolators instead (attachment 1).

Since all the DC power supplies are in use, I took 2 Sorensens for +/-18V and put them on the bench (attachment 2). I connected them to the 20V input on the chassis. I also connected the 18V DC source to the excitation point of the Acromags in the same way that was done for the YEND chassis. I turned on the Sorensens and made sure that the Acromags are powered up and drawing current.

Next, I'll use the YEND wiring spreadsheet together with the old XEND wiring spreadsheet to wire the existing and new channels to the DB9 breakout boards.

I removed all the DB37 and IDC50 front panels from the chassis and retained only the 2 BNC feed-throughs.

I also removed the IDC50 breakout boards from the DIN rail and put 4 optical isolators instead (attachment 1).

Since all the DC power supplies are in use, I took 2 Sorensens for +/-18V and put them on the bench (attachment 2). I connected them to the 20V input on the chassis. I also connected the 18V DC source to the excitation point of the Acromags in the same way that was done for the YEND chassis. I turned on the Sorensens and made sure that the Acromags are powered up and drawing current.

Next, I'll use the YEND wiring spreadsheet together with the old XEND wiring spreadsheet to wire the existing and new channels to the DB9 breakout boards.

[advait, andrei]

We needed two additional ADC channels for the seismometer setup, to read out the temperature of the seismometer can/enclosure, and another for the ambient temperature near the can. Koji let me know that there are two EPICS ADC channels, C1:X01-MADC0_EPICS_CH29 and C1:X01-MADC0_EPICS_CH30 that are currently unused. These are also conveniently CH 6 and 7 on the ADC board referred to in the previous post. 

Andrei and I laid out a couple of new BNC cables from the 1x9 rack to the seismometer, along the older cables. The image below shows the current configuration of the ADC - 

The cable labelled EX-SEIS ... is for the temperature of the seismometer connected to CH 5, SEIS_AMB_TEMP is for the ambient temp connected to CH 6, and SEIS_CAN_TEMP is for the can temp connected to CH 7.

All three channels yield a count number on caget-ing, so I had to calibrate these too. The calibration is given by voltage = [ -3.0593-4 * (counts) - 0.00367 ] V and I obtained this by reading values for three positive voltages. 

[Yehonathan, JC]

I have made a panel design for the acromag chassis. These panels will hold DB9 connectors. The previous ones had DB37 cutouts (Attachment #1). A bunch of these were found with the acromag accesories, but none had DB9 cutouts. We are switching to DB 9 to connect to the coil drivers without an issue. We are essentially trying to copy what the Y End looks like as much as possible. After taking measurement of the current panel, I used Front Panel Designer to create a new panel with the DB9 cutouts. these spacing between the connectors are ~0.4 inches which will allow us to fit 4 DB9 connecters on each panels. (Attachment #2)

Currently, I'm constructing the mode-matching telescope on a breadboard for the alignment of two OMCs for BHD.

What I did on July 25th:

• Constructed the input optical system before the fiber collimator (Attachment  1).
• Measured the mode-matching efficiency to the fiber collimator: 0.878 +/- 0.006

Next:

• Fix the end of the fiber on the mode-matching breadboard.
• Arrange two lenses to match the output beam with the OMC eigen mode (Horizontal: 489.6 um, Vertical: 490.5 um)

I performed the ETMX matrix diagnolization based on the free swing test using Anchal's scripts.

First I calculated the resonant frequencies using

python /opt/rtcds/caltech/c1/Git/40m/scripts/SUS/InMatCalc/getResFreqs.py ETMX/SUSfreeswing_ETMX_1374048020_POS_PIT_YAW_SIDE.txt -d 600 --optics ETMX -s ETMX/1374048020_POS_PIT_YAW_SIDE_ResFreq.yml

Some changes had to be made to  getFreeSwingData.py to account for the fact that all the suspension models has been updated since the last time this test was done.

Then, I run the diagnolization script

python /opt/rtcds/caltech/c1/Git/40m/scripts/SUS/InMatCalc/sus_diagonalization.py ETMX/SUSfreeswing_ETMX_1374048020_POS_PIT_YAW_SIDE.txt -d 600 --optics ETMX -r ETMX/1374048020_POS_PIT_YAW_SIDE_ResFreq.yml -f ETMX/1374048020_POS_PIT_YAW_SIDE_diag_plot.pdf -b 0.01

Attachment 1 shows the spectrum of each DOF before and after diagnolization. There is some improvement in the seperation. I populated the ETMX input matrix with the new calculated values.

I now turn to tune the damping loops.

Aiming for Q=5, I find that POS, PIT and YAW are slightly overdamped and SIDE is underdamped (attachment 2).

These are the changes in the damping gains:

POS gain 75 -> 40

PIT gain 24 - > 2.5

YAW gain 17-> 5.0

SIDE gain 61 -> 250.

Attachment 3 shows the ringdowns after tuning the damping gains.

Yes, it seems like ITMX UL sensor issue has reared its ugly head again (see attachment).

Yesterday during the recovery work from the CDS restart, I found ASS was not working for both arms. Both arms were flashing and locking, but their alignment seemed suboptimal.

Also, something seemed wrong: For example, the ITMX was too excited when the OPLEV servos were on. The arm locking was achieved when the ITMX oplev was turned off.

Compared the network settings between some of the machines.

It seemed that we can write network settings into /etc/network/interfaces. (Comment lines omitted)

source /etc/network/interfaces.d/*

auto lo
iface lo inet loopback

pianosa has the same content, but I found chiara has much more elaborated lines. So I decided to put the details there.

source /etc/network/interfaces.d/*

auto lo
iface lo inet loopback

auto eno1
iface eno1 inet static
      address 192.168.113.215
      netmask 255.255.255.0
      gateway 192.168.113.2
      dns-nameservers 192.168.113.104
      dns-search martian
      dns-domain martian

And just in case ifup was ran

$ sudo /sbin/ifup eno1

This made rossa connected to the wired network as rebooted.
So reactivated the NFS line in /etc/fstab
Rebooting pianosa brought it back to the nominal operation. Victory.

{Koji, Yehonathan}

I had some issues with the Debian GUI which I thought would be solved upon reboot.

I restarted the machine but the boot sequence would get stuck at the cvs/cds mounting.

To get access to the command line prompt I edited the Grub script by adding init=/bin/bash to the end of the line that start with the word linux and booting.

This allowed me to access root with reading permission. To get writing permission I ran

mount -n -o remount,rw /

then, I commented out the cvs/cds line in /etc/fstab and booted. This brought back the normal debian GUI.

However, we were unable to manually mount cvs/cds. The problem seems to be that rossa got disconnected from the network. Switching port on the network switch and using a different network cable didn't seem to help.

Currently, rossa is bootable but there is no network access.

[Koji, Advait]

Summary

• Spare ADC/DAC channels were prepared for Advait's experiment.
  C1:PEM-SEIS_EX_TEMP2_MON gives us the raw count number for the temp sensor signal.
  C1:PEM-SEIS_EX_TEMP2_CTRL can provide the output for the heater control (calibration ~0.5/2^15 V/count)
• The remote switching of the X End shutter was restored.

End X configuration

• Advait already had the temp sensor connected to the power supply. And the heater cable was laid out between the seismometer and the 1X9 rack.
• We went to the X End:
  • Connected the temp sensor signal to the ADC adapter chassis at CH5 (fifth connector ... the numbering starts from CH1) - This corresponds to the ADC channel of CH28.
  • DAC connection
    • We were unsure if the DAC adapter was configured to be single-ended (SE) or differential. The unit is S2200003 . Fortunately, this DCC entry had a photo of the PCB board. We confirmed that the unit was configured to have SE outputs.
    • The DAC adapter unit has four general BNC output ports (CH13-CH16). CH13 was reserved for the ALS Laser Slow control.
    • (Unrelated topic) Connected green shutter TTL switching input of Unibritz shutter controller to CH14
    • The heater input was connected to CH15 (the channel last but one)

Software configuration

• There is no Acromag at the X End right now. Therefore, for now, we want to use the fast channels to implement the channels.
• Opened c1scx model:
  • Added ADC0 CH28 to the adc ch selector
  • Added a PEM box with "top_names" option. ADC0 CH28 signal goes into this box.
  • In the PEM box, the input goes into EPICS OUT supposed to be "C1:PEM-SEIS_EX_TEMP2_MON".
  • The EPICS IN, supposed to be "C1:PEM-SEIS_EX_TEMP2_CTRL", goes to the output of the box. This output is connected to DAC CH14 (i.e. 15th channel)
     
  • Added an AUX box. In the box, the EPICS channel "C1:AUX-GREEN_Shutter" is read. This goes to a filter module (C1:AUX-GREEN_X_SHUTTER_CAL).
  • The output of the AUX box goes into DAC CH13.
• Opened c1asx model:
  • Checked the DAC output. Some spare filter modules were connected to the DAC channels we wanted to use. They were cleaned up.
• Edited c1auxex EPICS database
  • Now the shutter channels are double booked. So, commented out C1:AUX-GREEN_Shutter entry in /cvs/cds/caltech/target/c1auxex/ETMXaux.db .
  • Restarted modbusIOC by "sudo systemctl restart modbusIOC.service" on c1auxex
• The model was built and installed.
  $ rtcds build-install c1scx
$ rtcds build-install c1asx
$ rtcds build-install c1x01
• Upon restarting the models, all the machines crashed with DK. As per Yehonathan's suggestion, /opt/rtcds/caltech/c1/Git/40m/scripts/cds/restartAllModels.sh was used to bring everything back online. It was painful, but it worked after a few times of repetitions.
  • By the way, I checked what's wrong with burtrestore at this opportunity.
  • First of all, when the restart script is used, most of the time something is already wrong. So in situ recording of the epics values does not provide a healthy snapshot of the system.
  • Secondly, applying auto snapshot at once may cause the incomplete restoration whatever reason (like the command line is too long etc).
  • It'd be good to write/use a script to apply a bunch of snapshots one by one.
  • Maybe cleaning of the request file is also necessary.

Hardware check again

• We noticed that the ADC channel for the temp sensor was saturated. It was confirmed that the temp sensor sig has -14V. This needs to be fixed before the cable is reconnected to the ADC adapter chassis.
• Checked the DAC adapter ports' calibration:  +30000, 0, -30000 count to C1:PEM-SEIS_EX_TEMP2_CTRL gave +4.563V, +0.0047V, -4.573V respectively.
  This corresponds to 1.523e-4 V/count, This number is close to 1/2^16 V/count.
• The output saturates around +/-5V because... the diff to se converter in the DAC adapter has the gain of 1/2.
• The X end green shutter:
  • The shutter output was amplified by 30000 at C1:AUX-GREEN_X_SHUTTER_CAL filter module.
  • The shutter mode was reverted to "REMOTE".
  • It was confirmed that the shutter can be switched by giving 0 (close) or 1 (open) to C1:AUX-GREEN_X_Shutter.
    I quickly made buttons to switch the shutter from the ALS and IFO_ALIGN screens. Also now the shutter toggle scripts controlls the shutter as before.

Also, if someone could share resources for adding channels it would be great. There were two old slow Acromag channels, C1:PEM-SEIS_EX_TEMP_MON and C1:PEM-SEIS_EX_TEMP_CTRL_WATTS before, but these were disconnected and are now in use for something else. Yehonathan said there are no Acromag channels available now, but I might be able to set up fast channels to replace these.

JC shared this wiki with me, and while the "To add or edit a channel to an existing subsystem" section seems relevant (to add new channels to the SEIS subsystem), the folder structure seems very different from what is described there and these instructions seem outdated. It also does not explain where the physical connections of the BNC cables would need to be made.

Any help with this would be appreciated.

On Friday afternoon, I added a couple of fuses to complete the circuit for a power supply for the heater that is inside the seismometer can. A cable had been set up previously and connected to the X end rack, but the connection was broken. I am adding before and after images to show the change.

As you can see, the fuses were already there on the rack but not connected to anything. The red and black wires to the left lead to the bottom most power supply, which is currently set at 25V. The black and white wires to the right lead to the seismometer. I was accompanied by Yehonathan while I did this. 

We will also set up some Acromag channels for the heater control and temperature readout soon.

I underestimated measurement time. When I came this morning it was still running and I had to stop it.

I'm setting it up again to run tonight 1 AM with 5 repetitions per DOF instead of 15, and 720 sec free swinging time instead of 1050 sec.

Here are the transfer functions of the individual components along with the OLTF.

RXA: please spend some time trying to make the plot more useful: grid lines, distinguishable colors, zoom into important range, some text descriptions of each trace, comments on differences between theory and reality, etc., etc., etc

While modelling the system, I had no way of knowing the gain of each individual component. The individual transfer function of the PDH box and overall open loop is known, so I combined the Cavity, PZT, Low Pass filter, and Photodiode into one block, and fitted its overall gain to match with the overall OLTF of the system. 

Using this setup, with the pole values taken from Gautam’s thesis, and the fitted PDH zpk values; I was able to get the OLTF of the model to agree with the OLTF experimentally taken. I now use the model as a base for the setup, to test controller designs, impulse response, and stability margins. There is some deviation from the experimental value (~10dB in magnitude in the low frequency range, ~20 deg phase). 

We can see that at lower frequencies, the PDH box influences the transfer function while at higher frequencies the other components contribute. This is a good sign as we primarily want to increase the stability and performance of the system in the 10Hz to 20kHz range (Which is where the AUX laser noise is suspected to dominate in the ALS beat note) and this can be easily adjusted by changing the PDH box parameters. 

 I tried to simulate this block using the SR560, I tested with both a single and double pole at 30kHz. The 2-pole setup seemed to be the closest to the combined cavity, pzt, photodiode, low pass filter and summer combo (plant).  

With two poles at 30kHz and the fitted PDH controller implemented on the Moku: Go, I took an OLTF, and it was deviating ~20 dB from the model (in hindsight, I could've increased the gain in sr560 (but I dismantled the setup to use the SR560 to take OLTF at with digital controller at XEND) will take another reading if I get time, but still the phase is way off).  

I have moved onto implementing the digital PDH box on the actual system, I have not been able to design an optimal controller, but using the fitted values, and increasing gain on the Moku, I was able to achieve similar (not necessarily better in terms of open loop characteristics as that of the original PDH box. I will get some more data and post the comparisons of OLTFs and noises when locked.
 

ETMX freeswing test will start at 1:00 AM.

To cancel it run:

I modified freeSwing.py so that it doesn't turn off the filter banks in the ETMX suspension screen as to not change the alignment before the kick.

Dean made the OMC cables for the BHD Platform. They are going through the C&B process.

From left to right: QPD cable, PZT cable with Picomotor etc, DCPD cables

Since we removed the Acromag chassie modbus service could not be started so the soft Epics were not availble and the watchdog was not working essentially.

I made a temporary fix by commenting out the Acromag related comands in /cvs/cds/caltech/target/c1auxex/ETMXaux.cmd (a backup was created named ETMXaux.bak.cmd). More specifically, this block was commented out:

This allowed the modbus to start. Once the chassie is back this bloc will need to be restored with possible changes if we decide to add/remove Acromag units.

Unfortunately, software watchdog is still not functional since the logic turns off the enable switch on the coil driver using an Acromag BO that is not there. I'm still trying to figure out how to turn off the MEDM screen coil driver output using the Epics logic.

The new laptop for the Beam Profiler is ready for usage.

From the last meeting, we came to the idea of having 2 laptops that are designated for the beam profiler specifically. We now one laptop over at WestBridge, and one here at 40m. This makes it easier to travel with just the Beam Profiler heads, rather than carrying the Profiler heads + laptop. This machine has been named Stella by ChatGPT. Feel free to use it! (Only runs Windows 10)

Too many cavity poles! It should be just a single pole. Please show us your OLG Bode plot on the same plot as each of the individual pieces of the loop, so we can see what contributes to the TF.

Wanted to see if digital filter on the Moku:Go can be changed without breaking lock. Loading different filter files broke lock. Tried to set two identical filters and have an output matrix to switch between them, but the Moku:Go lacks the capability for that, so I switched to a Moku:Pro, which lets me configure an output matrix in multi instrument mode. Changing the output matrix in multi instrument mode on the Moku:Pro does not seem to break lock, but the transmission kept jumping erratically, I did not know if the low transmission was causing the instability.Maximum transmission was around ~0.6 when locked.

Yuta let me align the arm cavity, was able to get transmission upto ~0.9 while locking with the Moku:Pro. Lock held for aorund 5-7 seconds before breaking, it was not enough time to test if changing the output matrix is breaking lock. I was able to keep it locked to a higer mode, with ~0.2 transmission (not ideal) and changing the output matrix does not cause any erratic changes in transmission (dont know if this will translate to TEM00 mode). Will try again tomorrow to change output matrix while stably locked to TEM00. I realigned the arm cavity to it's original state and got main laser transmission back to ~0.9. I also noticed that it's easier to lock when input gain is attenuated (~ -5dB), the plan is to lock with a low gain filter, and switch to a more agrressive filter to hold lock stably.

Locked with Moku:Go

Locked with Moku:Pro

I borrorwed a visual fault locator for fiber cables from QIL with permission from Yehonathan.

I am going to use this in the 40m to construct the in-air mode-matching telescope for the OMCs.

I found that the MC transmission was low (~13000. Usually it's ~13500) in the morning.
This was due to the low transmission of PMC (~0.675. Usually it's ~0.685), so I restored the PMC alignment using the two steering mirrors before the input mirror.
The transmission of MC and PMC was restored to ~13300 and ~0.687.

[Yuta, Hiroki]

We conducted the followings for the PRMI experiment on Jul. 18:

• Leaned alignment tips?
• Measured the OLTFs for PRCL (UGF~220Hz) and MICH (UGF~50Hz)
• Data for noise budget: gpstime 1373763480 - 1373763560
• Measured the sensing matrix 
• Succeeded in locking for ~20 min. in the end

Details

OLTFs for MICH and PRCL:
We measured the OLTFs for MICH and PRCL (Atatchment 1 and 2).
The data for MICH is saved as measurements/LSC/MICH/MICH_OLTF_20230718.txt
(Sorry for forgetting to save the data for PRCL...)

• PRCL UGF ~ 220 Hz
• MICH UGF ~ 50 Hz

Data for noise budget : 
 1373763480 - 1373763560 (controlled by the OLTFs above)

Tips for stable lock
We succeeded in locking the PRMI for ~20min with the following conditions:

• MICH: AS55_Q, Power normarization with POPDC=1; gain=800; Out MTRX BS=0.5, PRM=-0.33
• PRCL: REFL11_I, No normarization; gain=-0.01; Out MTRX PRM=1  

For the stabler lock, the alignment of MI seems essential.
Before locking the PRMI, you should misalign PRM, lock MI and carefully minimized the ASDC with monitoring the video.

Sensing Matrix 
We measured the sensing matrix using scripts/CAL/SensingMatrix/ReadSensMat.ipynb :

Sensing matrix with the following demodulation phases (counts/m)
{'AS55': 2.0999999046325684, 'REFL55': 76.0199966430664, 'REFL11': 32.638, 'REFL165': 92.23478110797188}
Sensors       MICH @211.1 Hz           PRCL @313.31 Hz           
AS55_I        (+4.44+/-2.93)e+08 [90]    (+1.69+/-0.37)e+10 [0]    
AS55_Q        (+9.89+/-1.00)e+09 [90]    (-3.77+/-0.50)e+09 [0]    
REFL55_I      (-6.12+/-2.24)e+11 [90]    (+3.70+/-0.37)e+13 [0]    
REFL55_Q      (+1.46+/-0.53)e+11 [90]    (-8.89+/-0.86)e+12 [0]    
REFL11_I      (-1.43+/-0.53)e+10 [90]    (+9.13+/-0.86)e+11 [0]    
REFL11_Q      (-7.73+/-3.71)e+08 [90]    (+5.13+/-0.50)e+10 [0]    
REFL165_I     (-3.73+/-0.49)e+09 [90]    (+6.20+/-0.60)e+10 [0]    
REFL165_Q     (-1.38+/-0.10)e+09 [90]    (-2.24+/-0.17)e+10 [0]    

We decided to verify the PWM accuracy by adjusting the oscilloscope to fit a hundred cycles of the signal and measuring the mean voltage - the variation is linear as expected, although it does diverge at the extremeties as noted earlier.

Additionally, I compared the experimental data in the previous plots with a simulation using the model that we constructed, and after accounting for time delays in heating the system, we see a very good agreement between the two - 

I was able to use the Moku:Go to implement a digital controller and substitute the analog pdh box to lock the green laser at x end. I took the RFPD IF OUT (the error signal) and Servo Input (Control signal) and passed them through the Moku:Go filter. The mixer of the analog box is still used. The controller implements a digital filter with the poles and zeroes fitted to that of the analog controller. Transmission was low (~ .2 - .4), but similar to that acheived with the analog controller at the time, probably due to misalignment of the cavity. Changing the filter on the fly using the Moku gui brings about a momentary dip in transmission, but immediatley rises back to normal most of the time. The smaller the change in filter parameter (eg. gain 1dB -> 1.1 dB) the shorter the dip period.

I will test out how feasible it is to switch filters on the fly and have some python scripts that will switch filters, and one to potentially take transmission/reflection data and figure out if lock is broken. Play around with gain and offsets and filter coefficients, see what happens..
 

What's the nominal AOI for the mirror and the T value at that angle in a number? We will need that later.

I took away the fiber collimator and the collimator mount from the air BHD setup [Attachment 1 (before removal), 2 (after removal)]

The removed collimator is shown in Attachment 3 and will be used for mode-matching telescope for the OMCs.
 

I removed the T&R measurement setup with permission from Radhika (Attachment 1).

I will put away the chopper used in the setup tomorrow (the location of the chopper will be written here: elog #17822).
I am going to construct the mode-matching telescope for the OMCs here.

I wrote a Python code to simulate the AUX laser stabilisation system. It takes the poles and zeroes of the individual components (cavity, pzt, summer, low pass filter, pdh controller) and combines them, and shows the stability margins and impulse responses. Now any PDH controller designed can be easily plugged into the code and the stability and perfomance of the loop can be tested.

Trying to make a tabletop simulation of the setup, using an SR560 preamplifier with a second order pole at 30kHz to model the arm cavity and other ocmponents. The Moku:Go will be the digital PDH controller, and transfer of the setup will be taken with another Moku:Pro (not enough Gos). The individial and combined OLTF of the setup is taken and compared to the similar data. 

I loaded the same digital filter into both the Moku:Pro and Moku:Go, to see if there was any difference. I used the multi instrument mode, with Digital Filter and Frequency Response analyser. To my surprise, every configuration else exactly same, the transfer function was different for both Go an Pro. I also ran the Filter on the Go, and took the frequency response using the Pro, the transfer function was identical to the Go taken internally.The Pro is giving the expected transfer function. Idk if i am missing something here.

I will be testing out the digital controller on the Xend AUX setup today, replacing the anlaogue PDH box with the Moku:Go fitted with, see if i can get a lock with the digital version, see any problems that might come up.

I've made a cleanroom assembly in solidworks to create a list of all the components we will need when making this. Sadly, the 80/20 that we have in the cage area was taken already. I can't point fingers because the lock was removed due to it being an optional emergency exit way pointed out by Caltech Safety. Here is the 3D model of what the assembly looks like. Please let me know if you have any questions. 

I have not calculated the cost yet, but I will tomorrow. 

Just realized I wrote out but failed to post this ELOG update before I left last week:

On Thursday 07/05, I simplified the transmission measurement setup since a gain reference signal was no longer needed [Attachment 1]. (This is because the reference beam previously had to hit the TRANS PD, and now this wasn’t needed).

For the s-polarization measurements, I placed a HWP after the PBS to rotate the polarization. I used another PBS to ensure that the light was being reflected and tuned the HWP angle to minimize transmission [Attachment 2].

I’ve included transmissivity plots as a function of angle of incidence, for s and p polarizations [Attachments 3, 4]. For each point, I include the result from the Moku:Go spectral peak finder and the result from performing a FFT on the time series data.

I unmounted the PR2 optic and placed it back in its original case. I’ve left it on my desk if anyone needs it [Attachment 5].

Update: we have a working "I" controller for the puck. Here is the data:

In achieving this, we had to make a few changes;

 1. In the heater circuit, the resistor is now connected to its own power supply set for 10V. This enables us to have around 50% duty cycle at equilibrium for T_ref=45C:

*This is the heater profile for the above "I" controller run
 2.  We changed the frequency of the Pulse Width Modulation (PWM) of the RaspberryPi that is used to switch the MOSFET. It is now 100Hz compared to the previous 1kHz. I observed using the oscilloscope that the PWM was delivering ~870Hz when set for 1kHz and that it also struggled to deliver accurate duty cycles, especially for low duty cycles (below 5%). The duty cycles should always be kept between 5%-95% for accurate modulation.
 3. In case we want to change the power supply voltage, the controller code can be easily adjusted by just changing the maximum voltage constant.

I borrowed two beam profilers (WindCamD and Beam'R2-DD) and a laptop from the cryo. lab in the WB.

I am going to use them for mode-matching to the OMCs in 40m for a while.

Borrowed an NPRO controller from Liyuan @ Downs B138

Model M125/6-OPN-PA
SN 359M
Manufactured March 1998

The laser illumination is back with this controller.

POWER ADJ 0: 1064nm OUTPUT 277mW / 532nm OUTPUT (right after the harmonic separator) 0.67mW
POWER ADJ +10: 1064nm OUTPUT 364mW / 532nm OUTPUT (right after the harmonic separator) 1.09mW
Inter lock connected

Y arm green was still aligned to the arm cavity. I could confirm the Y green was locked with GRTY ~0.2.
The crystal temperature was adjusted to be 43.6degC as before and this actually did reproduce the beat note between the Y end AUX laser and the PSL.

The broken controller was labeled and stored in Yarm cabinet E03.

Model 126/6 M
SN 239M
Manufactured July 1996

PRMI carrier on 1f now locks for ~1min but not more. Below is the configuration.
The configuration is the same as basically the same as 40m/17578, except for PRCL gain. Maybe the RF demodulation phases changed?

MICH
  - AS55_Q (24 dB whitening gain, C1:LSC-AS55_PHASE_R=2.1 deg)
  - C1:LSC-MICH_GAIN = 0.4 gives UGF of around 30 Hz
  - -0.33*PRM+0.5*BS
PRCL
  - REFL11_I (18 dB whitening gain,C1:LSC-REFL11_PHASE_R = 32.6 deg)
  - C1:LSC-PRCL_GAIN = -0.005 gives UGF of around 150 Hz
  - PRM

Lock stretches:
  - 1373325207-1373325231 (with MICH @ 211.1 Hz and PRCL 311.31 Hz lines on)
  - 1373328637-1373328718 (without lines)

Next:
 - Check PRM actuation
 - Measure beam spot positions on PRM, PR2, PR3 and BS (with PRX or PRY configurations to make it easier)
 - Investigate why PRG fluctuates so much
 - Sensing matrix measurement (need longer lock stretch)
 - Try PRMI 3f for PRFPMI

The LED blinking on the head is normal. We can ignore that.

It is likely that the controller is the issue. I brought the PSL AUX laser controller to the Y End, and the laser turned on / the green light was visible.

Resolution:

- Find a spare controller (West Bridge, Downs). I could not find a spare controller. There is a dead (~5mW) NPRO in a cabinet, so there should be an unused controller...where?

    - Asking Downs/Liyuan if he has a spare -> Yes he is
    - CTN has two NPRO sets. One is functional. The other is not. It was confirmed that the controller is broken and the head is alive. (cf [ELOG CTN 2619])
    - Use the PSL AUX NPRO controller.

- Bring the PSL AUX to Y End. However, some experiments with that laser are on going.

- Bring an NPRO combo back from UCB. UCB has borrowed an NPRO combo. We heard that it is coming back soon.

[Yuta, JC]

Yuta and I tested at the Y End for Green Beam

Yuta and I went over the the Yend to check on the green laser. Yuta and I tried a couple of test to see if there was a quick fix. 

· We used an ohm meter to check if the red interlock switch was working. (Beeps when pulled and silent when pushed.) The Interlock switch is working just fine. 
· We tried another laser that we found in the laser cabinet along the Y arm. This was a Lumentum 125 that said “5mW power (DYING!!!!)” on it. When we connected this to the controller, the laser did not turn on. I’m assuming that whoever labeled this laser “Dying” would have put “Dead” if the laser stopped working. So the laser should work, right ? 

Overall, we can rule out the interlock issue. It should be something with the controller, or the head. Though if we assume the laser works fine, the controller may be the issue. If we can make sure this spare Lumentum 125 works, this will mean that our laser at Y End is still alive. Is there anywhere I can test this ?

• ETMX damping loops are not good. ETMX is moving by ~10 urad (if oplev is correctly calibrated), and beam spot moves by ~0.5 beam spot on ETMX when Xarm is locked. TRX fluctuates by ~10%. Simply tuning gains did not solve.
• ETMX does not come back after putting some offset to misalign and remove the offset to align. Hysterisis makes me hysteric.
• Xend acromag is removed, and we cannot access Xend shutter and ETMX slow channels.
• XARM ASS is not working.
• Yend laser is not working.

[Koji, JC]

Koji and I had a walkthrough of the Y Arm and raised questions of what accommodations we can make for the lab. 

· The HP3563A Control System Analyzer - is this needed in the lab ? Has anyone used this ?
· We have the black instrument rack the is behind 1X2 Rack. We can move this by Section Y8 of the beam tube, where the step ladders are stored currently.
· We will remove the bulky cart from the lab space. This has been used consistently to store items on and forget about. It is better to not have it at all.
· The small optical breadboard which is along Y-Arm should be prepared for usage. Cardboard boxes under the table will be searched through and put into a google sheet. 
· Save the 1U cases from the modules that are by section Y10.
· Cleanroom garments and wardrobes should be moved by the dress area along Y arm. 
    ·Issue : We are currently using 3 cabinets along the Y arm for Cleanroom apparel. 
·There are HEPA Filter replacements under section Y11. We should swap these out with the mobile HEPA Booth because these filters have been in air and most likely aren’t as clean as they should be anymore. This will be done before our next vent. 
· Koji and I agreed that 6ft x 4ft is a good Length and Width for the cleanroom that is being assembled. This will give us ~1 ft of clearance on 3 sides of the table. 

[Yuta, Hiroki]

Calibrations of the actuators have been done

Summary of calibration:

AS55_Q in MICH : 1.16e9 counts/m 
BS        : 26.19e-9 /f^2 m/counts
ITMX    : 5.07e-9 /f^2 m/counts
ITMY    : 4.80e-9 /f^2 m/counts
ETMX  : 10.99e-9 /f^2 m/counts (matched to ETMY with the gain of "x0.414" put in coil outputs filter bank. Without "x0.414", 26.52e-9 /f^2 m/counts was measured.)
ETMY  : 10.99e-9 /f^2 m/counts
MC2    : -14.27e-9 /f^2 m/counts in arm length
MC2    : 5.10e-9 /f^2 m/counts in IMC length
MC2    : 1.06e+5 /f^2 Hz/counts in IR laser frequuency 

Methods

Calibration of MICH error signal:
To calibrate the MICH error signal (C1:LSC-AS55_Q_ERR), we followed the same free swinging method as elog #17285 and #16929 but used the update code: scripts/CAL/OpticalGain/getOpticalGain.py.
By fitting the X-Y plot of AS55_Q and  ASDC, we obtained 1.16e9 counts/m (Attachement 1).

Actuator calibration of BS, ITMX and ITMY:
We followed the same method as elog #17285 and #16929. 
We locked the MICH with UGF of ~10 Hz and measured the transfer function from C1:LSC-BS,ITMX,ITMY_EXC to C1:LSC-AS55_Q_ERR above the UGF.
By using the optical gain, measured transfer functions were calibrated to the actuator transfer functions of BS, ITMX and ITMY (Attachment 2).
Fitting code: scripts/CAL/MICH/MICHActuatorCalibration.ipynb

Actuator calibration of ETMX, ETMY and MC2:
We locked the X(Y)ARM with ITMX(Y) and measured the transfer function from C1:LSC-ITMX(Y)_IN2 to C1:LSC-X(Y)ARM_IN1. 
With this measurement, we can directly obtain the transfer function of (Optical gain)*(Actuator of ITMX(Y)), which is not affected by the 1/(1+OLTF) suppression (We should have done the same method for the calibration of BS, ITMX and ITMY).
Then we changed the actuator to ETM(X) and MC2 and measured the transfer functions from C1:LSC-ETMX(Y),MC2_IN2 to C1:LSC-X(Y)ARM_IN1. respectively (Attachment 3,4).
By fitting the ratios between the measured transfer functions, we obtained the actuator transfer functions of ETMX, ETMY and MC2 (Attachment 5,6,7).
Saved diaggui templates: measurements/LSC/(XARM;YARM)/(XARM;YARM)_(ITMX,ETMX,MC2; ITMY,ETMY,MC2)EXC_Template.xml

Balancing the actuation of ETMX and ETMY:
The obtained actuator transfer function of ETMX(Y) was 26.52e-9(10.99e-9) /f^2 m/counts.
The result for ETMY was consistent with elog #16977 but that for ETMX was ~10 times larger than elog #16977 because of the new ETMX coil driver.
To balance the ETMX and ETMY, we added the gain of “x0.414” to the coil output filter bank of ETMX.
Also, in order not to change the OLTFs, we added the gain of “x2.42” to the LOCK FILTERS of ETMX and changed the gain values of the damping filters.

Manually switched TP2 to "Low Speed (LS)" mode (aka "Stand by")

[Yuta, Yehonathan, JC]

We have moved a Smaller Table and moving forward with preparing a clean room.

What we did:
  - Moved the 3 ft x 4 ft Optical table which was by the red toolbox next to the 1X2 Power Supply Rack.
  - Moved the Portable HEPA Filter (which was used during the BHD Vent) from ITMX Chamber

Summary:

     · Moved the 3 ft x 4 ft Optical table which was by the red toolbox next to the 1X2 Power Supply Rack.

Yehonathan and I cleared off this table and disconnected any power connections that were running to it. There was a grounding cable that was bolted down to it and safely removed. Next, I used the manual forklift to lift one side of the table while Yehonathan rolled a piano dolley under the table support. Following, we did the similar procedure on the other side of the table and safely sat the table onto piano dolleys. We rolled the table over and placed it into it's new position by the 1X2 Power rack. To take off the piano dolleys, I used the forklift once again. Yehonathan held one side still to prevent the table from rolling away dangerously, and Yuta slid the Piano dolley out from under the table. Next, we lifted the other side and slipped the other dolley out of the table similarly. once the table was sat down, we re-leveled the table.

     ·Moved the Portable HEPA Filter (which was used during the BHD Vent) from ITMX Chamber

I rolled over the the protable HEPA filter from ITMX Chamber. The height of this was too tall to clear the short ceiling by the PSL table. Yuta and I have to lean the HEPA sideways to clear this and we did so smoothly. The HEPA fit perfectly around the optical table. I will wipe down the table contact Maty to see if we have a proper cleaning procedure for this. I would prefer to add more of the plastic drapes to enclose the table, but we can use plastic Mylar sheets in the mean time. 

I modified the Binary Input DB9 connector. The coil drivers are now operating with the Run/Acq LED off. As per 17656 they should have a flat TF now.

"gain_offset" for ETMY coil outputs has been turned on

As mentioned in elog #17671, the "gain_offset" of gain(0.48) for ETMY coil outputs had been turned off for some reason.
I have turned on all of the "gain_offset" for ETMY coils and have changed the alignment offsets for ETMY to compensate the effect of "gain_offset": 

P: 2703 -> 5631
Y: -2296 -> -4783

After the operation above, I confirmed that the Y arm is flashing and the OPLEV laser is hitting on the QPD.

[andrei, advait]

We wrapped the puck in foam. We also performed a calibration on the AD590 sensors @ 25.3 Celsius and then proceeded with an additional 6-point calibration using a digital thermometer (DT310LAB).

This is the current setup There is also a lid similar in size with the bottom lid, and the entire foam is currently held together with a heavy wire role that was put on top of the lid.

We couldn't replace the red tape with the yellow-transparent tape because the yellow tape is not strong enough and an air gap forms between the resistor and the puck. As a result, we will avoid getting above ~45 Celsius in to prevent the red tape from becoming too hot. We are now proceeding with a step response measurement and will follow up on that soon.