[Jeff, Daniel]

We assembled a 6" to 2.75" zero length reducer flange on a non-rotatable flange of the 6" 6 way cross for the Laser Filter Cavity (LFC). The 6 way cross has 3 rotatable flanges and 3 non-rotatable flanges. Opposite sides have one of each.

After tightening the bolts to 34 Nm and seeing no copper gasket, we added the Agilent FRG702 Pressure Gauge. The magnet was 0.6" below the main body of the gauge before moving it up to fit the screws. We added the screws and tightened as hard as I could with the ~8" long wrench. The copper gasket wasn't visible. We tried to move the magnet down to its original location but could only get it 0.45" in down and unsecured.

We tried to add the 6 way cross to the 8" to 6" reducer tee but it was too heavy for 2 people if one person help up the rotatable flange on the tee. We'll need a 3rd person for this job.

Now that we have an operational timing circuit, we can assemble the main chassis of the Babbage cymac and hopefully not neet to poke around inside anymore. The lid does not fit on top due to the bulk of the pcie extension cables, so Daniel is going to cut a hole in the top and the cables will poke out.

Notes from assembly:

- The standoff holes on the General Standards ADC and DAC are a bit small for the 6-32 standoffs. However the standoffs are nylon and can be forced through.

- The angle brackets for holding the Adnacom backplane in place are not the correct size. Daniel made something work but it required stacking 7 washers to make a spacer.

- The main headache of this assembly are the pcie extension cables. Most of the cables Todd and I can find on Amazon are quite bulky, and there is very little room to fit them between the PCBs and the walls of the chassis. I ordered a few types and tried a few of Todd's, and they are all bulky. Todd has some old ones that work well but he can't find them sold online anymore. Instead of continuing to deal with this issue we are cutting a hole in the chassis.

Big todos on this project are to order the input/output PCBs (for whitening, dewhitening, and possibly other tasks?), and to push the software along so that we can begin communicating with the ADCs and DACs.

[Sander, Rodica Martin, Daniel]

Summary

We imaged the surfaces of our crystalline silicon optical substrates from Knight Optical using an optical profiler from 4D technology. The substrates are dusty, which significantly increases the surface RMS roughness. Removing the dust, either by cleaning with First Contact or by digitally masking the surface image to exlude dust, improves the RMS significantly. From these observations we conclude RMS roughness of all substrates is certainly < 0.3 nm and likely < 0.2 nm. Given these numbers and estimated scatter losses due to this roughness (see post 12119), we believe no further polishing is neccesary.

Measurement procedure

We unpackaged the substrates, handling them either by suspending them on some lens tissue and holding the lens tissue, or by holding the barrel with gloved fingers. We then aligned the optic visually with the center of the imaging aperture of the profiler. We then focussed the image using the profiler's software and manual focussing controls. The images captured were 3.547 mm by 4.255 mm, or 2056 px by 2464 px. An average of 32 images was taken. We saved .4D data files for all measurements. We marked the packaging of each subtrate that we measured with a serial number. According to Rodica the profiler software has some functionality to remove artefacts such as those from dust and fringes, but this removal is incomplete as these artefacts still show up in the image and contribute to the RMS.

Results

Spoked Circular and Octagonal (SC-2in/SO-2in)

The spoked optics are the cleanest substrates out of all measured. We managed to image unmasked sections of some of these without any dust present, giving RMS < 0.18 nm. These images still have fringe artefacts so this is to be taken as an upper bound on the RMS.

Unspoked Circular and Octagonal (UC-1in/UO-1in)

The uspoked optics are very dusty. We cleaned one side of an UO-1in substrate with First Contact (peeling off 35 min after application, Rodica says this reduces residue left after peeling off compared to waiting longer) and compared to the other uncleaned side of the optic. This uncleaned side has RMS = 0.675 nm, the cleaned side had RMS = 0.194 nm. Applying a digital mask, excluding parts of the image with dust, reduced the RMS of the dirty side to 0.165 nm.

Unspoked Wedged Circular (UCW-1.5in)

These substrates are quite dusty (like the other unspoked optics, they were all part of the same order, so this makes sense). Notable arc-shaped scratches were seen on both sides of the optics, presumably from the procedure that produces the wedge. The wedge was observable by focussing the profiler on one side and observing the changed fringe pattern once the substrate was flipped over. Images of a substrate with lots of scratches and some dust and fringe artefacts had RMS = 0.298 nm and RMS = 0.274 nm. Given the RMS improvement from cleaning dust and masking seen from other substrates, I would estimate the RMS is actually < 0.2 nm.

Data for all measurement is in the attached spreadsheet (will upload to the wiki, we might want to add more data into it), as well as some screenshots of the measurements.

We have one working clock circuit for a cymac main chassis. We have two more PCBs which are both slightly incomplete.

Today I went to Downs and Todd flashed our 3 arduino nanos with the software outlined in LIGO-E2300449. The steps look simple enough, but since Todd had already done this for his Arduinos it was as simple as connecting to his computer and hitting upload.

I then returned to bridge where I tested one of our clock circuits. I learned that the ATX power supply only outputs a voltage if the large ATX connector is plugged in, hence the PCIe backplane resting on the anti-static bag in the photo.

The one complete clock PCB that we have works. We have two more PCBs from Todd, but they are each missing a component or two. One is missing two buffers (center and marked by orange) and the other is missing an RF relay (look top, K1)  Perhaps I will work with Todd to complete them.

It is also possible that we don't need the relay, which appears from the schematic to be related to the external input.

In order to facilitate the development of a lab power supply standard, I have performed a census of all the devices currently consuming DC power. I have divided the space into three main locations, presumably each location will be served by its own supply.

Electronics rack

Optical table

Under table

Here is a pinout for the AOM Power connector for properly creating the cables using a 5-pin connector. 

The pinout is intended for the plug/male type connector, and the cable colors are to match the diagram. The 5-pin connection has two GND and +24V receptacles. 

Always double-check the output on each pin before plugging it into an AOM chassis. 

NOTE: The D-shaped symbol at the bottom of the pin diagram refers to the connector slot and should be used to orient the connector and how you wire it!

I set up a remote USB hub on the shelf over the optics table. Currently the TeraXion laser is connected, and I successfully sent commands via the control GUI on Brewster. The space under the table is getting a bit cluttered, perhaps we should think of a way to keep these cables off the floor.

I used the 'Mr grip' stripped screw hole thread repair kit and some longer screws to replace a door stop on the door connecting B111B and B111D. I used two strips per hole.

[Jeff, Torrey]

We had previously reported on some high frequency modulations in the error signal while locking the cavities with the piezo and EVAL-ADHV4702-1CPZ amplifier. 

This is an opamp with the standard differential input, and I (Jeff) previously created this this splitter doohicky to connect the input ground to the negative port of the amplifier and the input signal to the positive port. 

We found that upon removing this from the setup we successfully removed the spurrious pulses from the error signal. Currently the negative input of the amplifier is SMA terminated to the ground of the amplifier, and the input is connected to the positive input. 

[Jeff, Sander, Alex, Daniel]

I cleaned the inside of the 5" long, 10" to 8" nipple. We then slid the input vacuum cube on Laser Filter Cavity (LFC) toward the 5 in Long 8 in CF flange bellows. We stuck screws through the nipple and eventually through the bellows as well, getting them somewhat tight with plate nuts. We then further slid the input vacuum cube into place and bolted it down. Jeff and I then removed the screws on the top half of the connection and loosened the others so that we can drop in the copper gasket. The gasket was not properly seated intially, so Jeff pushed the gasket up and away and the gasket dropped into place. We reattached the screws and tightened all the screws. I still need to tighten the screws to ~23 Nm. Waiting a bit for the bellows to plastically deform back might make this easier.

I thought we would want to use the engine hoist, but we decided to just push the cube. We did make some metal to metal contact, but I think the knife edge is ok. I briefly inspected the window attached to the output cube since it is fragile, and it looked fine. I have no real reason to expect damage. I'll install the input window last.

I measured the length from the center to center of the vacuum cubes (around where the mirrors will go), and it is 44 in (1.12 in), almost exactly what Ian designed.

See attached photo.

I loaned ~15 1.1 Ω resistors (they need a 0.1 Ω resistor) to the Hutzler Lab from the rotating electronics rack from the Kimble Lab in B102B. Point of contact is Yuiki Takahashi.

[Jeff, Daniel]

I cleaned one of the 8" CF Flanges on a 5" long bellows from the Holometer from tape and tape residue. I then cleaned the other end and Jeff and I installed it on the open end of the reducing tee we just installed. The bolts have to be installed from the tee because of the ripple pattern on the bellows. The 2" long bolts can be inserted along side the 6" flange on the tee. I tightened the bolts to 23 Nm to get a steel to steel contact, 2 Nm more than the last 8" CF Thru flange connection.

I thought we could use a 6 way reducer cross with a 6" CF Flange (qty. 2) to 2.75 in CF Flange (qty. 4) from Alan Rice, but one of the knife edges is damaged. It was on a blank CF Flange, so maybe vacuum was still able to be pulled. I would rather not have to figure this out after cleaning and installing the part, so I will not use it for now. See attached photos of the damage.

[Jeff, Daniel]

We removed the tape and cleaned the residue from one of the 8" CF Flanges on a 8 in to 6 in Reducer Tee from the Holometer. We then attached it to the 10 in CF to 8 in CF 12 in Long Custom Nipple on the Laser Filter Cavity (LFC) output vacuum cube. We tried to make the 6" CF Flange stick out vertically as best as possible since the 8" flange on the custom nipple is rotatable. The 8" flange on the nipple on the input vacuum cube is also rotatable, so we should be able to align the bolt holes without being constrained by this. I started to implement the 80-20 structure to support the reducer tee. The bolts are quite bad the hex input is easily stripped. We should get better screws (they are 1/4-20 button head socket cap screws, around 1/2" length but I need to check). We have not yet tightened the bolts on the CF Connection.

I added 1.75" long 5/16-24 Screws, Washers, and Nuts to the Agilent TwisTorr 74 Turbo Vacuum Pump to a CF 4.5" Flange to KF 50 adaptor. Neither flange nuts nor washers fit next to the turbo pump, so I just used regular nuts. We still need to tighten the screws. 1.75" long screws were the perfect length for this application (matching the length of the nuts), but 2" screws, which are standard for connecting 4.5" CF Flanges with through holes, would also work.