[Ian Torrey]

We turned the 780 laser on in B111A to make sure it was still functioning. The beam was very misaligned on the Rb vapor cell and will need to be realigned.

Aidan borrowed 7ish silver plated 12-point head, 5/16-24 x 1.25" screws. Part SBX-24-125 from Duniway. We have 25 left. We need 16 for interfacing with the gate valve and 8 more for interfacing with the ion pump. We'll need another 8 for the turbo pump which should arrive in a few weeks, so we should get some more.

I leak checked the RbQ Vacuum Chamber because once again the pressure is stuck around 1.2E-6 Torr. My conclusion is that the leak followed the western 6"x12" plate when I rotated it. I thought that the leak was still at the western mating point between the rectangular prism frame and the cube frame, but I couldn't replicate it. There was generally weird behavior with the leak checker, the with the helium partial pressure going up and down for a 4 minute stretch, peaking at 1E-8 Torr L/s, way above the leaks from the chamber. 

I vented the chamber, took off the o-ring (it was quite easy to just pluck it away on the long end of the rectangle; I can't find the o-ring removal tool which might be needed for the square plates) from the problematic plate and added a fresh o-ring. I put the plate back in its most recent orientation and tightened the screws to 70 in-lbs. I started the scroll pump at 4:15 PM and the turbo pump at 4:22 PM with a pressure of 3.3E-2 Torr. 27 minutes after turning on the scroll, the pressure is 5.5E-6 Torr, which is similar to previous values.

[Torrey, Sander]

Torrey and I assembled the base (legs and crossbars) for the 72 x 30 inch TMC optical table (see photo). This will support a 4 inch thick table top. We might be able to lift the table top onto the frame ourselves, otherwise we need to get Caltech riggers to do it for us. Location of the table in the lab is still to be decided.

Continuation of this post.

In my previous model, I was creating the PRC out of one curved input mirror (no folded cavity mirrors either, just one optic for simplicity). We want to avoid having the input mirror as a curved optic, so now PRM2 and PRM3 are now a convex and concave mirror respectively.

After getting the model working, I set the length of the cavity to a constant value (PRCM1->PRCM2: 1.2m,PRCM2->PRCM3: 1.2m,PRCM3->BS: 1.2m, ArmLength:.6m for a total of 4.2m) and iterate over mirror curvatures. I recorded the stability of the cavity and beam size on the end mirror (I think these are the main things we care about so far?) and here are the results. 

stabilityofpowerrecyclingcavity.png 

beamsizeonendmirror.png

So for example, something like a ROC2, ROC3 = 25,-25 would be the most stable, but ROC2, ROC3 > 40,-40 would yield the biggest beam size on the end mirror. This code is now quickly changeable if anyone has something they want looked at.

Also random aside: I was having trouble creating the cavity objects for these models. I found printing these plot_graph model objects were helpful. Here's an example of the functioning one with crudely drawn lines showing the cavity path: output.svg.

[Ian, Daniel]

In spite of my old proclamation, I decided to clean the 4.5 in CF Gate Valve from Alan Rice. All the materials are stainless steel, with two exceptions: some nylon is press fit into stainless steel parts to reduce friction, the gasket to seal the gate valve when closed is made of viton. I did the standard 10 minute sonication in Simple Green:DI Water 1:30, DI water, and isopropanol. I did the following batches: the main body with the CF knife edges, the top with the bellows and screw that allows for actuation, the inside parts (including the parts with nylon, which is fine according to LIGO's methods, section 12.2), and two outside parts which don't touch the vacuum. I am baking at 200°C all purely stainless steel parts that will go into vacuum. On Friday, I will clean the viton according to this procedure, replacing the pressure cook with a shorter sonication and baking out at 120°C for 48 hours. I will finally clean the nylon/steel parts at 80°C for 24 hours.

We need to get cleanish stainless steel screws to assemble the part: 1/4-28 socket head cap screw, 1 5/16” long (qty. 12); 1/4-28 nut (qty. 12); #6-32 socket head cap screw, 3/8” long (qty. 4); #10-32 socket head cap screw, 1/2” long (qty. 1). The 1/4-28 parts can be used from the CF Flange screws (1 1/4" long screws are too short and we have many extra nuts). I got the other screw types from the MCE shop.

A while ago, I contacted Thorlabs technical support on how to make their Polaris parts UHV rated. They generally rate their parts as 1E-5 rated without a bake out and 1E-9 with "proper bake out". They wrote the following and attached the document attached here

For the suggested Polaris bake out procedure, you can refer to the attached document. This will consist of heating the flight hardware in a clean, certified vacuum system (<1 x 10-5 Torr pressure) at the highest temperature permitted without endangering the hardware but at least 10 degree Celsius ( C ) above its in-flight operating extreme, assuming this temperature does not exceed the maximum exposure temperature. Our non-Piezo Polaris mounts have a maximum operating temp at 200° listed on our website, so we can suggest a bakeout at 210°C. The 8-32 and M4 cap screws included with the Polaris mounts are not rated for pressures below 10^-5 Torr. Prior to placing any components in a sensitive vacuum system, a thorough pre-baking in a bake-out oven should be performed to remove all moisture and surface volatiles.

We can't reach 1E-5 Torr in our vacuum oven, but I still think we should bake out the components at ~1 Torr and 200°C.

The pressure seemed stuck at 1.1E-6 Torr, so I vented, took off the 2D MOT, added a spare 2.75" CF blank flange from the Holometer (we have plenty of unopened 2.75" blank flanges) and tightened the screws to ~40 in-lbs. I took off the 12" x 6" plate that was attached to the 2D MOT. I noticed a small hair which I wiped away with an isopropanol wipe but otherwise the inspection looked good. I wiped the vacuum surfaces with an isopropanol wipe, rotated the plate 180°, and installed it, tightening the bolts to 70 in lbs. I turned on the scroll at 2:07 PM and turned on the turbo 6 minutes later when the pressure was 4.7E-2 Torr. The turbo reached full speed 70 seconds later with a power of 4-5 W a few seconds after that. 25 minutes later, the pressure is 8.7E-6 Torr, a bit worse than the past two times but we ultimately care about the final pressure. At 6 PM, the pressure is 2.2E-6 Torr.

This is a first iteration of a power recycled finesse model of the GQuEST demonstrator.

For this initial version, I have the curved mirror as the input mirror. The ROC is ~20m, which yields a beam size on the end mirror of 2mm. The total length of the power recycling cavity is 4.2m (PRCM1->PRCM2: 1.2m,PRCM2->PRCM3: 1.2m,PRCM3->BS: 1.2m, ArmLength:.6m). A propagate beam call of this system shows the input beam size for this configuration would be ~2.3mm. port_tree.svg shows a map of all the input and output ports of the mirrors and beam splitters (this is less useful but I had been unable to get one of these to work before and thought it was interesting). 

Next, after talking with Sander, I want to find a range of PRM lengths, IFO lengths, and PRM ROCs that yield a beam waist of 2mm on the end mirror. To simplify this idea and to reduce the calculation time, I have merged PRCM1->PRCM2: 1.2m,PRCM2->PRCM3: 1.2m,PRCM3->BS: 1.2m into one length. Then I calculate if my model is stable over a range of ROC values, IFO lengths, and PRM lengths, combining the two lengths into a single power recycling cavity length. PRCgeometrywithcontours.svg is the result. I've included contour lines to see lines of constant beam size on the end mirrors.

I plan on iterating on this but this was a first attempt. For now I have a working model of the single simplified system and a version of the code we can iterate over specific paramaters. Note that for my simplified version an ROC for the PRM of 1.6m (which is what we have in the lab already) is not stable (also 3m is not stable, the size of the existing spares set aside for the LFC, needs to be greater than about 4.5m).

Since the pressure only reached 1.0E-6 after 2 full days of pumping, I decided to leak check again. There was no leak 180° from the past leak on the plate I rotated, indicating the plate was fine (a visual inspection showed it to be fine as well). There was a leak in the same place as last time (bottom right (south) of the west 6" x 12" plate that goes to the 2D MOT), but maybe smaller, around 2.5E-10 Torr L/s. Interestingly, closing the valve at the end of the leak check spiked the measured leak rate to 4.8E-10 Torr L/s.

I vented the chamber and removed the plate to inspect it and the frame. There was very slight discoloration on the plate in a few places, but I couldn't feel anything. The chamber looked fine. I replaced the plate. Putting in the screws was harder with the 2D MOT in the way, but I don't see an easy way to redesign. We shouldn't have to take this plate to add or adjust components in the vacuum chamber. Due to the awkward angle, I very lightly stripped the top of the head of a screw. This is cosmetic as I could easily get a hex key in, but the screw should be thrown out if we take off the plate again.

I started the scroll at 10:55 PM. Since I suspect the 2D MOT is tugging on the plate through the bellows, I pushed the plates that hold the 2D MOT foward after the pressure was ~1 Torr. I turned off safe start and turned on the turbo at 11:01 PM, the turbo quickly ramped up to speed, maxing out at 99 W drawn. As of writing this at 11:39 PM, the pressure is 3.9E-6 Torr and slowly dropping. If the pressure doesn't get close to 1E-7, I want to take off the 2D MOT, rotate the plate, and add a 2.75" blank flange. We should also use the RGA to confirm the presssure is from a leak and not outgassing. If the leak stays in the same place, the next step would be taking apart the frames, which would be a total disassembly. Being able to valve off the turbo would also make leak checking a lot better, so hopefully the gate valve arrives soon.

Based on this leak checking, there seems to be a singular problematic area. The easiest idea is to rotate the square plate next to the area to see if that fixes anything or the leak moves. There could also be an installation error this would fix.

I turned off the turbo pump, waited for it to stop spinning, then turned off the scroll pump. I turned on the N₂ line and slowly opened the needle valve. I waited for the pressure gauge to read over pressure, removed the screws (a bit of air rushed in), rotated the plate 180°, put it back on, tightened at 10, 20, 30, 40, and 70 in-lb (twice). I then turned off needle valve and N₂ line. I turned on the scroll at 4:37 PM and the pressure was 6.7E-2 just 5 minutes after turning on scroll, faster than before. I turned on turbo at this point, keeping soft start on for a pure comparison, but we should turn it off soon.

After 16 total minutes, the pressure was 5.5E-6 Torr (slightly better than LFC, ~20x better than Dewer at this time), 5W drawn (LFC 10W, Dewer 16W at this time), 1167 Hz pump speed. I am hopeful this was the fix as we've reached a better pressure much sooner. There is a chance the pumps run better the second time, but it took 2 hours of pumping to reach this presssure last time on the RbQ vacuum chamber.

I made slots on both ends of three 1/4-20 to 3/8-18 thread adapters for use with a flat-head screw driver. I used the band saw meant for steel and held the adapters with a collet (a 15/64" collet and 11/32" collet I think) and a square collet holder. Ian checked that a flat-head screwdriver fits into the slot.

I leak tested the RbQ Vacuum Chamber with LIGO's leak checker. Since the 4.5" gate valve hasn't arrived yet, we had to keep the turbo on which makes leak checking more difficult. However, I did fine at least one leak: the bottom of the rectangular prism to cube interface on the south side (roughly 5 Torr*L/s, but this is pretty arbitrary since the turbo is on (and helium spray level ambiguous)). I also discovered all three screws that secure the windows were loose. I gently tightened them. I turned off the leak checker and the helium line. 

I plan on letting the turbo pump a bit longer and seeing if the pressure gets lower. I want to take off, inspect, and reattach these plates since it should be easy. i could rotate the 6" square plate to try to see if that changes the leak spot. The leak could also be caused from the rectangular prism to cube interface itself. We would have to take off 8 plates to fix that, so I want to try that last. If we do that, I think I want to take off the cube entirely and evaluate the pressure with just the rectangular prism. We can make a MOT without the cube, but not a conveyer belt. We could attach the tee and include the ion pump as well as the electronics flange.

I grabbed the LIGO Lab leak checker from Maty Lesovsky and put it in B111A to leak check the RbQ Vacuum System. We can also use it to leak check the SNSPD Dewer. Maty said she shouldn't need it for a few weeks.

I have noted the following voltages on the Laser Filter Cavity (LFC) Ion Pump Controller and the pressure according to the Agilent FRG702 pressure gauge. The voltage corresponds to the current draw by the ion pump. These are only from "steady" times, not when there is a sudden change in the system, for example turning the ion pump on or closing the gate valve to the turbo pump. I graphed this data. The fit appears linear, but the lowest voltage (not steady state) I've seen is 2.1 V when first turning on the controller, so the fit seems good up to ~1.4E-6 Torr. Maybe this is when the current maxes out, assuming a constant current to voltage conversion.

[Briana, Ian, Torrey]

Placed the SHG setup on a new 1 by 1.5 ft breadboard (see new schematic). We will separate the SHG and vapor cell setup to be on two different sleds since they are fiber coupled and can be easily put in place elsewhere. The beam dumps are not in place and the optics are not aligned yet. Replaced Torrey's five 780 nm mirrors with the ordered 780 Newport BD.2 mirrors. Assembled the second Faraday isolator for the 780 nm light. We will not need to use the 1550 nm mirror that has the piezo imprint since the new schematic reduces the number of 1550 nm mirrors needed from 4 to 3.