I installed the Wilcoxon 731A seismic accelerometer and its associated P31 power supply. On the table near the filter cavities shown in the close up image of the seismic accelerometer. I hooked it up to the patch panel to port red 8. I then hooked it up to a moku pro to test it. It seems to be working and I tested it by jumping next to it. I have not calibrated the data coming out of it yet.

I moved the case to a shelf in the RbQ lab. See Seismic_Accelerometer_Case.jpeg for location.

[Ian, Torrey]

Continuation of 12256.

Testing the script in action. I rewired the set up to mimic how this data was taken and reconfigured moku 1. This includes VCO locking cavity 1 and having the FRA output going to the piezo. Take the updated data, run it through the code, upload to moku, and track performance with the uploaded filter on and off.

Between nofilter.png, the customfilter.png, and customfilterincreasedgain.png (increasing the single leading value in the text file to 3, from 1), reduces the overall RMS from 650 uV -> 550 uV -> 416 uV. Note that you cannot increase the gain arbitrarily high as it will cause piezo to ring up if too high (was too high at 5). This seems successful, but we still need better control overall.

OFC3 1550 alignment is complete. To recap, the 1550 OFC3 path has no test light. It is aligned directly to the output of OFC2. In order to achieve this we built a 1550 PDH scheme for OFC2 for more stable locking. I can then scan on the piezo in OFC3 to find cavity flashes. Additionally I had to implement some periscoping mirrors as outlined in previous posts.

One problem: OFC3.png is a cavity scan. In red is the PDH error signal for the 775 light and in blue is the 1550 TRANS signal. Fairly clean cavity scan. There seems to be a large unexplained frequency drift, as seen in ScreenRecording2025-04-22114505.mp4. This drift is significant portion of the cavity FSR (10's of MHz) and appears to just be the cavity/piezo. I don't believe it is the laser, as you can track the correction voltage being applied to the DC Modulation port of the teraxion via its interface teraxion.png. May be worth recording this data and making sure but quick look by eye it doesn't seem to be the case. Additional proof for this not being the laser is the fact the 775 and 1550 drift the same amount. If it was the laser the 775 light would be moving twice as fast due to the SHG frequency doubling.

[Jeff, Daniel]

We installed three 6" CF to 2.75" CF Zero Length Reducer Flanges on the 6" CF 6 way cross (horizontally except the south side). I tightened the west flange all the way to 34 Nm with full flange to flange contact. I tightened the north and east flange to 34 Nm but only a few turns at 34 Nm.

I borrowed the #8-32 Tangless Helicoil Insertion and Removal Tools from Stephen Appert. I also "borrowed" some Nitronic 60 #8-32 Length = Diameter Helicoils (it's going to be hard to return them once we use them). It's been logged here by Caltech LIGO Lab. I previously borrowed some #2-56 helicoils and tools and some #1/4-20 helicoils and tools. Everything is in B111A except the 1/4-20 helicoils, which are temporarily in B110 since I cleaned some of them.

[Jeff, Daniel]

We cleaned the 1.33" CF to 2.75" CF adaptor plate for the PICAS 2D Rb MOT and attached the Rb Vapor Cell by hand and lightly with a ball-end 9/64" screw driver. We don't have a stubby (low profile) 9/64" allen key which we need to finish tightening, so I bought some.

We also installed and tightened the screws on a 2.75" CF tee above the vacuum chamber. I haven't figured out what I want to put on it because we don't have a layout.

Since the pressure was 5E-7 Torr, I took some data with the RGA with the ion pump and turbo still on. See the attached graphs. Some takeaways: H₂, H₂O, and CO₂ are about 2-3x the LIGO beam tube requirement (note the old date; maybe they require better now). N₂ and CH₄ are below LIGO's requirements. The total calculated pressure is 1E-6 Torr, 2x what the agilent pressure gauge reads. This could be because the RGA is at the end of a 7" tube, an angle valve, and a tee before it reaches the center of the 6 way 6" CF cross. The Agilent ion pump has direct line of sight to the center of the 6 way cross. It does not appear like there are significant hydrocarbon contaminants (they should appear at 12-16 amu (C, CH₁, CH₂, CH₃, CH₄) and above 40 amu). The Holometer desired a partial pressure below 1E-10 Pa (7.5E-13 Torr) which is the same as LIGO's hydrocarbon goal. I think we are far from this target, but in my amateur opinion we have met our vacuum goal.

The Agilent Account Manager came by to inspect the pressure gauges because they read over pressure when the pressure is normal and they pressure seemed to go down even though everything was off. Nothing much came of the visit, but he will forward along the latter issue.

The pressure after the weekend was 1E-3 Torr. There at least isn't a major leak then. I vented the system (both the turbo and the main body of the chamber) and pumped down with the scroll. I turned off safe start as the turbo has been lubricated by being on recently. At around 1 Torr, I turned the turbo on and it increased up to its final speed in under a minute. After a few minutes, the pressure was in the 1E-6 range so I turned on the ion pump. The ion pump behaved as normal. The RGA showed a decent amount of H2 and was ~2-3x the reading of the Agilent pressure gauge. I closed off the gate valve to the turbo and turned off the turbo. The pressure rose by quite a bit to 1E-5 Torr and I decided to turn off the ion pump and RGA. I think we need to wait until the pressure is closer to 1E-6 Torr to close the turbo pump. After turning off the ion pump, the pressure rose to 1E-3 Torr in ~10 minutes and then stabilized. I think this explains why the pressure was so high last Thursday. I still don't know what caused the ion pump to go off. Hopefully the ion pump is ok.

Around 4:45 PM today, I vented everything and pumped out with the scroll and then turbo. I turned on the ion pump around 4:55 PM. By 7 PM, the pressure should drop to 2E-7 Torr. If not, then something is wrong with the ion pump, or something is wrong with the all metal angle valve, RGA, or the connections between them. Closing the valve to the RGA would be a good check.

[Ian,Torrey]

In order to more stably lock the filter cavities, we have been in talks of taking cavity transfer functions to isolate any resonances to upload an invert version of the TF as a filter in the DFB. We have written a script to make this work. Here are the following steps:

1) Take OFC TF. rawdata.png

2) Rescale the magnitude of the data so that it is 0 at low frequencies.

3) Invert it in complex form and choose the appropriate window for the filter. invertedandwindowed.png

4) Down sample the data to make the experimentally taken data smoother.

5) Fit the down sampled data to ZPK format. ZPKfit.png

6) Convert your ZPK curve into second order sections.

7) Write it to a text file in the format the moku wants. output.txt

success.png (I moved it down 20 dB so you can see the shape at high frequencies. This is done with the leading single digit on its own line in the text file.)

Some notes associated with this:

-The moku can only upload 4 SOS in this format as a custom filter. Limits the amount of shape we can add to our filter dramatically.

-We will upload this script to the log in a cleaner format but for now here it is piezoTF_calculation.ipynb. Its really messy though, we'll update with a cleaner version soon. 

I downloaded and briefly set up some software to talk to the turbo pump and pressure gauge on the LFC. I took off the interlock to prevent the turbo from accidentally turning on.

I cleaned 15 Nitronic 60 1/4-20 1/4 long helicoils and 4 of the 5 custom 1/4-100 bushings. I used the standard 1:30 simple Green:DI water, DI water, then isopropanol baths and sonication for 10 minutes each. I am currently baking them out at 200°C for 48 hours.

[Jeff, Daniel]

I connected the Stanford Research Systems Residual Gas Analyzer 100 (SRS 100) RS232 cable (9 pin D-Sub) to a RS232 (male) to USB-A (male) adapter and plugged it into the USB hub in the mobile clean room. I also connected the same adapter from the turbo controller to the hub. I turned on the RGA and took some spectra around 3:30 today. I've attached the graphs from the data. The y-axis is Torr. The total pressure is the sum of the partial pressures; these graphs imply a pressure of around 1E-5 Torr, even though the Agilent pressure gauge had a reading around 1E-7 Torr, two orders of magnitude different. Jeff and I came in around 11 PM tonight and took another RGA reading; H2 and N2 levels were higher than at 3 PM, up to 3E-5 Torr. We then found out the pressure is 1E-3 Torr on the Agilent pressure gauge and the ion pump was off. I don't know if the ion pump going off is a cause or an effect of the high pressure. If the 6E-10 Torr/s rate of rise measurement is accurate, then the pressure after ~12 hours would be ~3E-5 Torr. This seems to indicate the high pressure caused the ion pump to go off. If the helium from the helium leak check were in the system, it would be easily visibile with the RGA (and shouldn't cause such a huge pressure rise). The up-to-air valve and the gate valve to the turbo pump were closed and tight.

If the pressure is this high, this explains the RGA measurments being so high (it's not rated above 1E-4 Torr, so a little discrepancy there makes sense). There is a big mystery for why the pressure is 1.2E-3 Torr after being stable for days. To test the agilent pressure gauge, I rotated and moved the magnet's z-axis (it's a cylinder and on another cylinder). This lowered the pressure to ~7E-4 Torr, half as low. A factor of 2 in pressure isn't a big deal for these issues.

The H2 levels on the RGA also seem quite high.

[Jeff, Daniel]
We attached the Infleqtion PICAS 2D Rb MOT Adapter Plate to a 2.75" CF Bellows with 1.25" long 1/4-28 screws and no washers. 0.875" long screws weren't quite long enough. The gasket was initially incorrectly seated on the knife edges; putting the curved edge on the bellows seemed to help when I tried again with a new gasket. I tightened the bolts to 14 Nm, the standard for 2.75" CF thru flanges, and made steel to steel contact all around.

[Jeff, Ian, Daniel]

We vented the turbo/scroll system on the Laser Filter Cavity (LFC) by turning the knob on the turbo pump. There was a decent amount of air that came in, implying there is a good seal from the turbo to the outside world, even when the scroll and turbo are off. The gate valve was closed and we noticed no pressure increase in the main body of the LFC, implying the gate valve seal is good. We then vented the main body of the LFC with the up-to-air valve. A little turn increased the pressure from 1E-5 Torr to 1 Torr, so one needs to be careful there. I turned the valve a bit more and the pressure slowly increased. Notable, there was a lot of gas sucked into the LFC even when the pressure gauge read "over pressure" (over 760 Torr). I therefore tried to play with the magnet on the pressure gauge once the LFC was fully vented to get the controller to display 760 Torr. Rotating it and moving its z-position didn't accomplish it.

We took off a CF 2.75" to KF50 adapter + a KF50 blank flange and added a 7" long 2.75" CF tube to the all metal angle valve (the latter of which was quite dirty on the flange outside the vacuum ). We needed this 7" long part so that the cup of the Stanford Reserach Systems Residual Gas Analyzer 100 (SRS RGA 100) had room to be inserted into the vacuum system (it needs ~3"). We then connected the SRS RGA to the tube and we tightened the bolts.

I opened up the gate valves and had previously opened the all metal angle valve to the RGA and shut the up-to-air valve and turned on the scroll pump. After 20 minutes, the pressure was 0.1 Torr like before, so I don't think fiddling with the magnet changed anything. I then turned on the turbo (still in soft start mode) around 0.08 Torr. The pressure and power drawn behaved like last time. 20 minutes after turning on the turbo, the pressure is 4.9E-6 Torr.

To use the RGA (and have the turbo controller, which reads out the pressure) talk to the computer, we need some RS232 (male) to USB-A (male) adapers. Ian found this which looks good.