Tentative schematic for FC 2 and 3

Now that I have a good idea of what fits after fiddling with the first filter cavity I have a rough sketch for OFCs 2 and 3 attached. Going to put optics on the sled to see if things fit tomorrow. Note that the four combos of camera/PD at the different exit ports are most likely temporary. Welcome any comments/concerns below.

LIGO C-HARD Pitch Optimal Linear Controller

[Ian, Lee]

Given a C-HARD Pitch plant as well as the related noise spectra with seismic (environmental) noise and sensing (measurement) noise, we calculated a number of optimal linear controllers for different weightings of two figures of merit. The code for this can be found in the github buzz dev branch in the AIC folder (with commit SHA). The outline of the system is given in the attachments as AIC_layout.pdf. This layout shows the the overall layout of the system we used to calculate our controller. The bandpass FOM, plotted in F1_plot.pdf, which we constructed using a 8th order bandpass and a 16th order high pass filter combined. The different roll offs on the different sides of the bandpass filter is better for our solver to process and is more like the real noise drop off on the SNR integral. In order to compare this controller with others made for this DOF, we are using a FOM that picks off right after the controller. This FOM is marked F3 on the layout attached. The Flat FOM (marked F2) remains right after the plant because it needs to hold a non zero D-matrix across the plant FOM combination. The original bandpass (F1) is not used here. A plot of the noises and plants is in noise_plant_bode.pdf. To meet solver requirements the measurement noise is modeled as flat. 

Results:

The RMS results are plotted in AIC_Hib_FOM_RMS_Scatter.pdf. In this plot, each point represents a controller, where its inner color is how heavily it is weighting the Bandpass FOM (F3) and the outside color is its phase margin in degrees. The red line is the H2 results which show the lowest possible performance RMS for a linear controller. The x-axis shows that controllers RMS after the bandpass FOM (F3) and the y-axis shows the RMS performance after the flat filter (F2). One set of controllers with a constat F3 weight of about 3e4 but increasingly more stable phase margins has been used in the HiB open loop bode plot (AIC_Hib_OL_BODE.pdf) and closed loop bode plot (AIC_Hib_CL_BODE.pdf).

I have included the H2 RMS plot in AIC_H2_FOM_RMS_Scatter.pdf, which shows just the H2 results as well as the associated open loop bode plot (AIC_H2_OL_BODE.pdf) and closed loop bode plot (AIC_H2_CL_BODE.pdf).

Vacuum Oven Software Setup

I was able to download the software for the JeioTech via this webpage (click OV4) and interface with the vacuum oven. It's a bit clunky, but it should work well enough when we have a computer permanently connected.

Filter Cavity Unboxing

Took out the coated filter cavity and there are noticeable dents/smudges on it. Pictures attached.

Leak Test for Vacuum Oven

[Lee, Daniel]

Following up on this post, we evacuated the oven for 5.5 hours. The final pressure was still around 15 Torr. I turned the oven off, but this time I closed the valve to the pump. After 18 hours, the oven pressure did not go up by more than 1 Torr. This means there is a leak upstream of the valve to the pump. I think is it the connection from the tubing to the oven itself. Other possibilities are the connection from the tubing to the pump (unlikely the dominant culprit as this connection is very tight), the KF O-ring seal (unlikely to leak this much), the pump itself, the tubing itself, or the oven between the valve when it is open and the tubing.

Vacuum Pump Test for Vacuum Oven

[Lee, Daniel]

We tested the IDP-3 that is connected to the Vacuum Oven. It turns on with a somewhat annoying sound, but this goes away quickly. The oven reaches its final pressure of 15 Torr in around 10 minutes.

We turned the pump off at 4 PM. At 11:15 AM the next day, the pressure was 30 Torr. This is a leak rate of 0.02 mbar L/s. 

With that leak rate and a pumping speed of 1 L/s, I would expect a final pressure of 0.02 mbar = 0.015 Torr.

The quoted base pressure of the IDP 3 is 0.25 Torr.

This could be explained by an improperly calibrated dial on the pump, and the vacuum is closer to 1 Torr.

Quantization of Bit Resolution on Filter

I have attached a presentation that walks through my whole process of finding the effects of bit resolution on a 2nd order low pass filter. In the first slide, I again highlight the -40dB/decade property of low pass filters with different critical frequencies, and in the following slide, I explain what I mean by adjusting the bit resolution of the filter. Afterwards, on the third slide, I have the normalized transfer function for a low pass filter with 10 Hz critical frequency and 10 MHz sampling frequency at float point, 36 bits, and 37 bits, showing that discrepancies are not noticeable. I follow with plots that show the difference between the float transfer function and the resolution adjusted transfer functions. These show the odd behavior of the difference as I increase the critical frequency, which we exploit further with the mean steady state difference plots (both in standard and log v. log format).

Particle Counter Measurements

I took some particle counter measurements in B102 and some LIGO optics labs in Lauritsen. x/x/x is the 0.3 um, 0.5 um, and 1.0 um cumulative count over 1 minute.

B102:

77973/4322/471 before fans were turned on. The fans were left off over the weekend.

9/0/0 after the fans were on for 90 seconds before starting the count

 

LIGO clean room to measure the scatter off of our optics, including the super-polished optics:

1/0/0 near curtains

0/0/0 2 feet inside curtains

 

Room with Fizeau Interferometer

84868/3495/397

GariLynn Billingsley says they clean the LIGO test masses a lot after they are in this room, but this is definitely not a clean room

Preparing Vacuum Oven for Initial Testing

I removed all miscellaneous items from the vacuum oven and set the maximum temperature from 320°C to 200°C on the red dial with a flathead screwdriver. I connected the included plastic tube to the vacuum port and tightened with the included hose clamp. I couldn't tighten the hose clamp as much as I would have liked because of the limited access behind the oven. I want to move the oven and fix this before we turn on the pump, although it might be fine as is.

I set up the Agilent IDP 3 vacuum pump. To its KN16 port, I attached a rubber o-ring and a KN 16 to 5/16 " hose adapter from Ideal Vac. I tried to use the Ideal vac KN 16 clamp, but it's unusable. I borrowed a KN 16 clamp made by LDS Vacuum from the Hutzler Group, and this works as intended. I attached the hose from the oven to the adapter and tightened the hose clamp. I was able to tighten this hose clamp much more than the clamp on the oven.

See attached photos.

780 Laser

[Briana, Ian]

We mounted the DBR780PN Distributed Bragg Reflector (DBR) butterfly laser. First, we placed the configuration card for the Type I laser into the mount (LM14S2). Then, we unclasped the mount's latches and placed in the laser on top of the pins. The laser output was put into the laser input from the vapor cell schematic. The temperature and current controller (TED200C and LDC205C respectively) are placed on top of the central table overhang (facing the entrance to the lab) and connected to the mount. The maximum current limit for the laser diode is 250 mA and the ideal operating power is 39.6 mW, so first the current limit was set to 250 mA, then the actual current increased until the reading of laser power from a powermeter placed directly in front of the laser output was around 39.6 mW (this current was ~230 mA). Then, the current limit was set to this value. The temperature controller was then powered on. It only has the thermistor sensor which reads out in kilo Ohms, so the conversion to temperature is done using the Steinhart-Hart equation (see spec sheet). The temperature controller's current and resistance seem to be coupled. The coefficients used for the equation are under the Absolute Maximum Ratings table on the spec sheet, relating resistance (kilo Ohms) and temperature (Kelvin). We confirm that the temperature controller works using Figure 2 on the spec sheet. For example, when we set the temperature to 29 Celsius (converting to resistance using the SH equation) and set the current applied by the temperature controller to 140, 180, and 220 mA, the output power (measured by the powermeter) was between that of the 25 and 35 Celsius curves. We initially had some issues because the temperature measurement tool did not show a significant increase in temperature when we decreased resistance. This was probably because the tool measured the surface which might not accurately represent the actual internal temperature of the diode. 

The spec sheet is found at https://wiki.mccullerlab.com/DCC/S2400003. Video tutorials lfor laser setup are found here: Setting Up a TO Can Laser Diode (Viewer Inspired) | Thorlabs Insights (youtube.com), https://www.youtube.com/watch?v=LAixCOso-FE

Shack-Hartman Wavefront Sensor Setup

[Erin, Daniel] Worked on developing a setup for using a Shack-Hartman Wavefront Sensor to image mirrors more easily. Light would pass through a polarizing beam splitter, then to a quarter wave plate, reflect off of the mirror we are sensing, pass through the QWP again for a total polarization shift of 90 degrees and then through the beam splitter again. It will then be deflected into a converging lens (f=50mm) as it is in the setup attached. Then the light will pass into the mounted lense array (MLA) and then to the camera. There will also be a lens tube to give about an inch of space between the array and the camera sensor. The wavefront sensor is an MLA with a square grid of smaller lenses that measure the intensity, which is how the wavefront is reconstructed. Mounted Lens Array for sensing the wavefront (Thorlabs MLA300-14AR-M - Ø1"): https://www.thorlabs.com/thorproduct.cfm?partnumber=MLA300-14AR-M (lens, tube, and camera we already have)

Logic-X digitizer up and running!

Today I was able to bring-up the Logic-X ADC/DAC boards and get the demo firmware and software running.

Said software was platform-specific for Windows, but I was able to install a Windows 11 VM using UTM for MacOS. I had a little trouble getting the JTAG drivers installed and then getting the USB and network interfaces to show up, but now it's all solved and it wasn't that difficult, relatively speaking.

I've used the pre-built artifacts provided in the BSP, so the next step is using a home-built bitstream and executable. I've made some progress with the bitstream but for some reason it's failing timing, so I'll continue to look into this next week. Then we'll be able to incorporate the filter and run some tests with real signals! In parallel I'll also try to run it on the RedPitayas via PYNQ.

By the way, the digitizer board runs HOT, so we're gonna need a cooling solution --- meanwhile, I'll try to get some desk fans to force some air through.

 

Attached are the program log as well as a plot of the ADC data read out by the demo routine.

Migrate the SOPs to the wiki

I was able to easily cut/paste the 1550 laser SOP into the wiki

https://dcc.ligo.org/M2300159 -> https://wiki.mccullerlab.com/DCC/T2400003

You can then use the "permissions" tab during "edit" mode to allow everyone to view (so it becomes public, even though DCC items default to non-public)

 

I would prefer we use the Wiki for these, so they can be kept up to date and can then more easily link to other SOPs or documentation.

We can/should make LIGO DCC entries, but they can just link into our system.

Installed Vapor Cell

[Briana, Ian]

Polarized beamsplitters and fiber couplings are installed and placed generally on the board. Optics needing disassembly (see previous post) have been placed in their correct locations. Still need more BD.2 780 mirrors, which are certainly somewhere. The exact location remains a mystery.

The vapor cell was installed. First, four screws on the bottom of the GCH25-75 heater assembly were unscrewed to separate the structure into a top and bottom. Because the vapor cell is too small to fit in the radius of the heater assembly, adapter rings are enclosed around the vapor cell. The adapter rings were placed on the inner section of the heater assembly due to the length of the vapor cell. The nub on the side of the vapor cell is oriented towards the side to protect the cell. To mount the adapters around the vapor cell, the small screws on the very sides of the adapters were removed to split the adapters into two, after which they were placed over the vapor cell and rescrewed in. For a single adapter, the screws are antiparallel. The set screws on top of the adapters were screwed in last to ensure the vapor cell was fixed in the adapters. The vapor cell was then rotated so that the green part of the screw faced upwards (from the bottom part of the mount). Then, the top part of the mount was placed over the vapor cell such that the upwards-facing green part aligned with the holes for set screws in the top mount. Finally, the four screws were rescrewed in. 

Unfortunately, one of the four screws on the bottom of the vapor cell is now lost in the optical table abyss. The screw type is Black-Oxide Alloy Steel Socket Head Screw, 2-56 Thread Size, 1/4" Long | McMaster-Carr. A replacement was found and used instead. 

Mr Brightside

[Ian, Torrey, Briana, Jeff]

We performed an experiment with the filter cavity. We split off the fast controller error signal going to the laser, and put it in the new Fosi speakers. Then used another speaker to play a song near the filter cavity. The speaker used to play the song was physically decoupled from the table and cavity. We then recorded the output signal using the moku data logger (original file found at "Nextcloud/GQuEST/B102/Bright.csv"). We then wrote a quick script that turns a CSV into a .wav file. Bright.wav is the result. 

Mr. Brightside can be clearly heard. Joking aside this is actually show casing the strong audio pick up of these very long bowtie cavities. The constant tone in the .wav file is a 840 Hz tone that is a result of poor grounding to the DC modulation input port used to control the frequency of the laser.

This is the way music like this was made to be heard.