Meghan Lab Notebook

Long term tasks

• Manufacture more beamsplitter cube holders (in progress)
• Figure out windows equivalent to a mac mini where we can ssh in and boot remotely etc. (Eske had an idea for this)
• Create a full assembly in SolidWorks or Fusion of the system (importing parts)

1.26

Tasks with timeline O(weeks):

• find the windows equivalent to a mac mini where we can ssh in and boot remotely etc. (Eske had an idea for this) – moved above to long term tasklist!

order smaller piezos (have a conversation with Thorlabs about best product for this) - we ordered APY001 new stages and should be able to use the PIA13s

Tasks (before Feb. 2)

• conceptual design review slides (link to them here in wiki!)
• read some of
• the papers from the slack
• make wiki organization clearer - separate page for meeting notes/misc, timeline, working updates, BOM
• read up on / play wit
• h kinesis software 
  • Download parallels for mac, download kinesis

Windows on mac, kinesis download, GPT testing

Was able to use the parallels app to run windows on my mac (M1 chip). Successfully dowloaded and opened kinesis, but after looking through app and manual, seems like it will be more interesting to play with once I have a device plugged into my computer. Is there anything lying around in the lab that would use this type of control? If not, I will parse the manual without trying to do on a physical object for now, and see when the stuff comes in.

On a less serious note I asked GPT to help me come up with an acronym for this project. It came up with some interesting ones, nothing I love yet though. Will do some more thinking.

One question has come up thinking about the system tonight; with the imaging mirror system resembling something like the picture below, will our line of beams be oriented radially (like shown in the photo) going into the telescope or on a secant line of the primary mirror? I remember a mention of radial symmetry assumption today so I assume it is like this photo but want to be sure.

1.29

Conceptual Design Review Slides

PowerPoint and my notes are attached here. I have some confusion still about the alignment procedure, especially the second part. Also am curious which direction (multiple?) of angular tilt the beams will have relative to one another and how we will decide that.

1.30

Discussion with Prof. Stubbs and Elana; key takeaways:

• agreement that number of beams will have to be 2^n; decision between using 4 and 8 beams. Beam spacing TBD, tentatively use 4 beams
  • this would mean we need four adjustable optical mounts (for each optic ABCD in drawing in DOME-SUM page)
• there will be some shuffling on the mounting side of things - might need multiple beams etc
• light any direction into a beamsplitter will go the other two directions!
• will need some way to mount a corner cube to the APY001 tip/tilt stages from Thor

Updated tasklist

• CAD something (or think about how glue would work) to mount the corner cube to the tip/tilt stage
• draw out beamsplitter diagram with light from multiple directions

2.1

Updated Conceptual Design Review Slides

Updated PowerPoint slides with new part numbers and answered questions from previous conversation.

2.2

Notes from conceptual design review meeting

Conceptual design review! Overall went well. Key takeaways:

→ We still aren't quite measuring what we want to (the during the night seeing on the optical path) because we don't want to mount on the telescope, and observers don't want incoming light at nighttime (understandably). However, it's a step in the right direction.

→ Beams that we send in will span the entire diameter of the primary mirror.

→ No reference beam is needed for beamsplitter cube alignment - we can just align the beams with respect to one another with our corner cube! This allows us to avoid the issue of the laser beam pointing back at itself when aligned (and right back into the source).

→ Continue updating risk page on project, and allocate costs to potential risks

→ We will leave space for a second beam expander to solve the problem if the spots on the focal plane are too large

→ Elana and Prof. Stubbs will work on confirming mounting specs for the dome, where electronics will go, if we can get ethernet from electronics box, etc.

Actionable items:

• Make a list of risks, potential solutions, and associated costs. 
• (moved to longer tasks) Create a full assembly in SolidWorks or Fusion of the system (importing parts) - can do this from home so will work tomorrow afternoon on this depending on other things

Tasks

• Open stuff!
• start getting piezo actuators to move around
  • update new laptop
  • install kinesis
  • connect piezo 
  • connect to stage
• control a stage in two directions
  • motor controller with two piezos

Computer needs many updates - taking a while before I can troubleshoot Kinesis installation being blocked. 

• After downloading all updates, it seems I have to switch out of S mode to install any non-verified applications. I will not be able to return to S mode but that's ok. 
• Have Kinesis downloaded! With one piezo, the GUI looks like this:
• I am able to move a single piezo actuator back and forth. It makes a terribly annoying noise, but the position shows on the motor controller which is great!
• The connection architecture:

• • The piezo plugs into the back of the motor controller, where there are four channels.
  • The USB port on the front connects to the laptop. 
  • Note that there are only 2 USB ports on the computer, meaning that only 2 KIM101s can connect (and that this computer can only drive 8 piezos = 4 stages!) There will also be USB needed for flip mounts...
• Each piezo has a maximum step size of 20nm, and maximum step frequency of 2000Hz, which is a speed of 40 micrometers per second. → is this linear speed?

It does not seem like the PY001 I have opened has the barrel mount required for the PIA13. I am going to initially mount to the PY004 pitch/yaw platform. 

Issues:

• the micrometer screw is incredibly hard to loosen
• micrometer must be tightened entirely to access this screw which could cause damage to the platform when it comes out
• there aren't really instructions anywhere on how to remove this thing
• intended hex wrench size is somewhat unclear

I sent an email to Thorlabs tech support, and will proceed based on their response.

In the meantime:

• CAD a mount to the APY001 stage for the beamsplitter cubes
• 3D print one such mount, and machine another out of aluminum in the machine shop
• Lay out individual components (at least one of everything in the entire setup) with mounting hardware to make sure we have everything else we need

List of mounting solutions needed:

• Beamsplitter cube to stage (CAD)
• Amazon-bought beam blocker to LMR2 flip mounts - screw in optic retainer that comes with the flip mount!
• Piezos to either set of stages (emailed Thorlabs, we will see what they respond)
• Figure out wider optical table for beam? - if doing multiple rows on the same optical table we may mount multiple optical tables, or just mount stuff directly to beam, but this has not been fleshed out yet
• LIGHT SOURCE! We only have mounting for the source from Edmund which has wider divergence than we want - Prof. Stubbs will order a holder
• Camera (LP126MU) - can mount with a bracket and optical screws; see photo below under 2.4
• Should I have a tube for the beam between source and beam expander? - not necessarily to start with, although we might enclose the beam later

EOD Recap

1. We had a good design review meeting! There are minimal changes in my understanding and came up with a name for the project.
2. There is some confusion with how the PIA13 piezos mount to the stages; I have reached out to Thorlabs and will see what they reply.
3. I was able to mount a bunch of components (shown above)! There is a list of mounting solutions that needs to happen going forward.

2.4

Talked about mounting problems with Prof. Stubbs: added solutions in green above. 

Mounted camera:

Note that we have to download a separate software package for this camera - it does not run on Kinesis. Have updated a page of computer requirements on the Organizational Notes page. This camera also needs a USB 3.0 port to work at the correct speed - plugged it into a regular USB and the on-screen view updated extremely slowly.

Camera update: we can use a C-mount lens to get actual images (yay!)! I am thinking that the lens resolves all the incoming photons so that the CMOS detection can see an image and not just lots of light.

End of Day Updates

• Finished CAD of mount for beamsplitter cube; emailed Alejandro to see if I can machine one tomorrow morning.
• Might need a brief electronics refresher (or a nicer oscilloscope?)
• Next steps: wire up beams and test, machine a proof of concept mount, begin aligning stuff (CALL THORLABS ABOUT PIEZOS TOMORROW MORNING!!)

2.5

Talked with Alejandro this morning - the part will likely take me a significant amount of time to machine. Still planning to make one as proof of concept, but will probably not get to it before Friday! Did not have time to call Thorlabs during the day.

2.6

Messaging with Thorlabs online now. They said that the PIA13 stages can be used with APY002 platforms and PY004s, but not with the APY001's (the micrometers and thumbdrives have different mounting). We ordered some APY002s. Next steps:

• machine beamsplitter cube holder
• attach beamsplitter cube to PY001 with manual thumbscrew adjustment - just need screws to come in, but then should be set!
• get everything mounted on the right size posts to be aligned for light - moved to 2.16
• electronics review
• get laser working and on and responding to signal - moved to 2.16
• figure out laser beam mounting (when parts come in) - moved to 2.16

2.9

Spent morning in machine shop - should be able to get parts done by eod. Also borrowing oscilloscope from Jim MacArthur - the one I found in the lab did not seem to work very quickly.

End of day update: "finished" parts, but the beamsplitter cube does not fit. I will enlarge them, make sure they work and remachine/3D print during next week.

2.15

3d Printing Beamsplitter Cube Holder

Working on correcting the cad for the part right now. Planning to 3D print one so I have something to use tomorrow at least.

First PLA print failed – increased bed temperature to 215, and added a brim underneath the part. Was told for future prints that the Prusa Slicer is better to use than the Ultimaker Cura slicer the REEF defaults to using. Print should take 1 hour 39 mins.

Second iteration print also failed - the outside of the part would not have been solid. Onto third iteration - part will take 2 hrs 36 mins.

Figuring out required length of screws:

The 91251A876 2 1/4" screw is too short, according to my CAD. Going to try importing the 2 1/2" (91251A871).

This looks good! They are somewhat more expensive, but we should order a few of them, as the machine shop stock is all too short.

End of day update

The parts are printed! Want to do a bit of filing/deburring in the machine shop to make them slightly nicer, but other than that these should work as a proof of concept at least and allow me to align things. The beamsplitter cube does rest in the cavity, which is great.

2.16

Deburred the 3D printed mount; the beamsplitter cube fits great! Will just need the screws to secure it, which were ordered by Helen this morning.

Tasks

• get everything mounted on the right size posts to be aligned for light (best possible w/o other input)
• get laser working and on and responding to signal 
• figure out laser beam mounting 

I am going to start with mounting the laser diode.

Was able to mount with the AD10F, the threads of which went into an SM1 lens tube that connected on the other side with the 5x beam expander (BE05-532). The beam expander connects to the SM2A30 adapter, which has external SM2 that connect through the SM2RC ring to the external SM2 threads of the other beam expander (BE20-532). That is held up by an SM2RC ring on a post.

I found a SM1TC ring that can support the first beam expander (the current setup is not stable enough yet), but I could not find a post without a threaded end. Update: I have discovered that the top screws are removable on all of the posts! Super cool.

I also turned Jim's scope on for the first time since last Friday today, and the screen woke up with a bunch of fuzziness. I'm not sure why it's not working, but I have a multimeter on hand that should be able to tell me enough to keep working on trying to get the laser on.

I am using the HP E3631a power supply, but there doesn't seem to be any power coming out. Trying to troubleshoot with the manual. There is no error code on. Update: had the ground output instead of the -6V output. It now provides exactly 5V!

Laser is working! Shown is just the laser, modulated with a 10kHz, 5Vpp sine wave. The light also makes it through each beam expander individually:

However, I am experiencing an issue where there is no spot on the wall for light that travels through both beam expanders. You can see there is some light that makes it into the second:

I can't tell if the intensity is so reduced that the light is no longer visible. Messaged with Elana and Chris - Elana suggested trying to see it with lights off. I'm not sure if that will involve going to the other room; it is relatively bright in the room I am currently in. I think first I am going to try to align the system without the beam expanders in place, leaving room for them and leaving in their mounts.

Have been exploring mounting for a little while. My thoughts:

• Although posts can be screwed directly into the optical table, it is better to use post holders and mounts because then position and rotation are more easily adjusted.
• For height only alignment, changing out the alignment key on every single post is frustrating. So long as each part of the system goes on the same heigh movable mount, this works great.
• The laser level is helpful for seeing where things fall, but it is difficult to find a spot for it at the correct height.
• System-specific things:
  • the piezo stages and the camera would have to be propped up more off the optical table based on the height things are now
  • the difference between the height of the source mount and the beam expander mount does not correspond to the difference in thorlabs' beam posts; are there more spacers somewhere?
  • Also, where are the corner cubes in the lab?

For the end of today, I think I will make sure that the PIA13s work in a APY002 platform.

I did get a PIA13 in! Not sure if the reverberation/clicking from insertion damaged the stage. I think it's okay though. Note that I had to use pliers to get out the micrometer.

For Monday:

• prop up camera/APY002 to correct height - use right angle bracket and right angle mounting thing, order longer 1.5'' post for platform
• find more spacers/fix gaps in height alignment
• figure out how post spacers work with posts - are there longer screws somewhere without socket head? – buying longer screws with "low-profile" socket head
• check that light is emitted through both beam expanders in dark room (using photodiode and lockin instead, but same idea) - moved to friday
• find corner cubes - ordered and Chris will bring in from home

2.19

We are planning to align initially without the beam expander in place. Note the camera is large enough to measure the PSF center were we to align with the beam expander.

Mounting math

The minimum height for everything is determined by the bottom of the smallest post to the center of the MFF102. However, this is much shorter if we use the flip mount in its low position. The bottom of the mount to the center of the low position is 1.97 inches (50mm), and the distance laterally from the center of the device to the center of the lens is 2.15in (54.5mm). We can flip this down into the beam.

The other big constraint is the difference in height between the laser beam diode holder and the holder for the beam expander. These must be held separately even though we can connect them continuously, as we wanted to align with the laser dot being visible. Backup plan can be to align with the lockin and a photodiode.

The difference between the bottom of where the diode (in the SM1M20 tube) or the first beam expander (BE05-532) is supported by a mount and where the bottom of the large beam expander (BE20-532) is supported is (2.20" - 1.20")/2 = 0.5". So assuming the mounts are the same thickness (I will check that this is standard but from eyeballing looks like a reasonable assumption) we would ideally have a .5" difference between the two posts.

After checking a few part numbers, the mounts are not the same thickness. So I will do my height categorizing by mount.

This can be achieved using spacers - the TR_M series from Thor has spacers for _ mm between 1 and 10. However, the problem then becomes that I don't have long enough screws for the alignment keys to work and either a) have less control over the rotational alignment of optics or b) I try to find these screws (which seems difficult from a preliminary search but I'm sure they are on McMaster somewhere). Update: have found them https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=4101&pn=SH25LP38. So I should probably get an assortment of lengths when I finish calculating spacer heights, from McMaster. 

Thorlabs' 1/2" diameter optical posts come in the following lengths: 3/4", 1", 1.5", 2", 3", ", 6", 8", 10", 12". There is a line of 12.7 mm diameter posts that come in metric heights, but they seem to have metric setscrews, while the other posts have 8-32 setscrews that work with all of the parts we have bought. So, I'm sticking with the 1/2" diameter posts. Say we put the flip mount on the shortest possible post (0.75"). If we want 50mm vertically to the center of our beam, the post for the SM1M20 tube (held by the SM1TC clamp which has a radius of 1") should be 1" above the 0.75" post, putting it on the 1.5" post with one 5mm spacer correction. The rest of the calculations are in the images below:

Still need to figure out corner cube mounting and the FMP2 mounting to the APY002. For the mounting to the APY002, we are going to buy a plate that will allow us to just screw in the lens mount. Alternatively, we could mount it directly to the stage with an 8-32 hex screw through the center hole and the FMP2 on the other side (doesn't work with original 1" diameter post adjustment idea but does work with new one).

I've also just figured out I forgot to account for the height of the alignment key or the post holders in the lenses that go directly on post mounts. So I will have to do a bit of reshuffling.

Method:

1. calculate how much raw height everything has (raw height meaning all parts without added posts/spacers)
2. get everything about equal to the largest of the raw heights (add height to all if that reduces the number of parts

My drawings and math in PDF form:

Sent a Thorlabs order and a McMaster order; will add them to my inventory. Sadly the photodiode light test and the beginning of alignment will have to wait till Friday. Will also slack Elana about translation stage in the darkroom, not sure whether she needs it but it could be useful for alignment. Might mess up heights but hopefully I have ordered enough spacers to spare. 

Just thought - I might have messed up some math because I was not factoring in the base plate on the lenses/mirrors/beamsplitter but was on the tip/tilt stages. Adjusting for that gives us less spacers necessary for one of the large tubes, but more for the other - so luckily it evens out. I will leave the cart as is, but the adjusted PDF is below.

Actually I also calculated MFF102 height incorrectly (I tried to put things together and they are not quite as expected so this is good that we have not ordered anything yet). → new adjusted pdf is below.

PDF with Mounting Specs

Have to re-do cart and tabulate order.

Tasks for Friday:

• check with photodiode and lockin - not necessary yet as not using beam expanders yet (moved)
• align beam (moved)

Before Friday:

• slack elana about translation stage
• make Thorlabs cart

2.23

Beginning by trying to get everything on right sized posts based on what I have. Have also confirmed that the beam expander will go on a translation stage, and we will get rid of the tube connection between the laser diode and the first beam expander. I realize this could change my mounting somewhat but will work it out when the stage arrives.

Plan for today

I have enough hardware up and running to align a beamsplitter cube. That is the goal for the day. What needs to happen:

• mount things in a line and screw down
• mount camera at correct height and remove lens
• connect camera to computer, turn on beam
• align the dot on the camera!

Below are photos of the function generator and power supply, aligned components, and camera feed on the computer.

I have successfully "aligned" a single beam (quotation marks because now that we are doing relative alignment, this just means I have a single beamsplitter cube sending light back to the camera)!

• We can change the duty cycle on the signal generator to change the brightness of the diode, which makes a smaller spot on the camera feed and allows us to have better visibility over how aligned the spot is.
• We could also block some of the beam to make the spot dimmer and take up a smaller number of camera pixels.
• Lengthening the distance between the beamsplitter and the beamsplitter cube allows for more minute control over how well the angle is aligned.
• Chris and I had a short discussion about whether to have the beam expanders on a translation stage or a system of mirrors/beamsplitters that moves in and out of the beam path; we did not come to a conclusion.
• We ordered a translation stage from Thorlabs
• We also discussed how to tell if the light we have is collimated, and we bought a shear sensor from thorlabs that will do so. We can now use the collimation adjustment on the beam expanders with that to properly collimate our beams
• Chris confirmed that May 12-17 would work for a trip to Chile.

Next steps

• Get in touch with Alejandro on the machining side - we want 12 sets of the pair of parts that hold each beamsplitter cube
• Motorize an APY002 mount and put a beamsplitter on that
• Set up beam expander on motorized translation stage
• (moved later) Set up single beam system with correct alignment and space for the beam expander.
• (moved later) Figure out angular acceptance of beam expander to evaluate the translation stage setup
• (moved later) Photodiode test with beam expander system!

2.26

Spoke with Alejandro this afternoon; going to make some adjustments to design (round out corners, and add corner clearance holes that we machine with a flat end mill so that corner cube can sit more flush in the recess) and then we can send the files to Steve who can waterjet the parts. He did mention that it's fine to have one as proof of concept but for a lot of parts being manufactured we will need a billing code (I assume this is for Steve's time/waterjet time). I asked Chris/Elana about the code; Chris said that Helen has it. I am going to make 1 or 2 before going ahead with all 12. Emailed the file to Alejandro this afternoon, and will see what he says.

2.28

Alejandro said the parts are in the machine shop. Planning to get to the shop when it opens on Friday to run them! He also asked me to CAD a workholding piece for the parts, so I am going to do that now as well.

3.1

Today is my deadline for phase 1 of the project! Going to try to get things done. I am already working on getting ready for phase 2 with machining.

Holdups:

• alignment keys- sent email to Chris to order
• would like Chris' help removing the second micrometer from the tip/tilt stage - done
• double check lockin method (connect the trig out form the signal generator to the modulation signal for the lockin, connect the photodiode to the in on the amplifier)
• get translation stage up and running

Alejandro very nicely made me the plate on which I am machining my parts yesterday!

Machining procedure

Bottom part:

1. Waterjet holes, contour
2. Center find and counterbore (6-32 counterbore) the two holes on the diagonal (that hold this piece to the top piece)
3. Flip part, re-tap (6-32 tap) holes on diagonal of work holding piece so it can be held in opposite direction, and use same 6-32 screws to hold edges of part
4. Program rectangular pocket (program with below info worked well, ran in 2 steps 0.0250 depth of cut each) with 1/2" end mill
   1. 
5. Cut rectangular profile on same part with 1/4" end mill
6. Cut corner holes with 1/8" flat end mill, programmed as a drill bit (take picture of coordinates that Alejandro gave)
7. Counterbores for center 4 holes

Top part:

1. Waterjet holes, contour
2. Steps 4-6 from bottom part

My one machining mistake today was that I did not change the tool number between operations (as Alejandro said it did not matter), but then when I changed the dimensions of the tool on a different operation, it remained different on tool 1, meaning when I tried to cut the pocket, it ended up cutting I wider pocket than I wanted. Fortunately, it was on the top, so I could just flip the part over and machine the other side (which I did!). 

Yay! It works. I am leaning towards having the rest of them machined if that works with budgeting.

Now trying to set up the beamsplitter on the translation stage. It seems like the holes on the stage are not quite the right sizes for my 1/4-20 mounting scheme for optical posts. I remember talking with Chris about buying a mounting plate – not sure if that is something we did. For now, I have come up with the idea that I can take two mounting posts, mount their 1/4-20 holes to one another with a set screw, and then mount the 8-32 to both the SM2RC holder and the holes in the plate on the translation stage. Surprisingly, I seem to have guessed heights (0.75" post + 1" post) that work close to the right height.

Have settled on using a longer set screw with a 5mm spacer in between two 0.75" posts. Not perfect still - will do out the math later.

I also don't think the translation stage itself mounts to an optical bench, which is a much bigger problem. For now I can try to find cap screws, but there is a mounting plate on thorlabs that we should purchase (and that I am surprised did not come with the part). If we are purchasing that, we might as well purchase the mounting plate for the top of the stage, except for the fact that then everything gets much taller...

Let's try to do the photodiode setup, then will try to ask Chris at the same time about all the holdups. 

So it turns out I didn't make the part entirely correct. My CAD file is right, but it makes a difference which side the holes are on compared to the holes to hold the cube together. I accidentally aligned this upside down. However, we can still send the parts to be machined - the file Alejandro has is correct (I'm sure because the 3D printed version works). 

End of day updates

It may not have felt like a very productive day, but it was! I understand my beamsplitter cube holder now better than I did before and feel good about having 12 sets of them machined. I will reach out to Alejandro/Steve about machining out of black anodized aluminum.

I understand the photodiode testing after talking to Chris, and will have to make sure that I low pass filter (extremely low, ~10kHz) the combination of signals that it returns using the time constant (this is because there will be a beat frequency between two - we want the difference to be zero as they are the same frequency! there will also be a 2*10kHz beat that we filter out). I will also need to insulate the photodiode with Kapton tape so it does not ground through the optical mounting setup.

Chris helped me remove the micrometers from 4 sets of stages, and we put in the piezos. So I should be set to scale up once we have fabricated the rest of the beamsplitter cube mounts.

We also found the mounting plate for the translation stage to the optical bench and the plate to mount optics on top of the stage in a box that I did not realize was mine! So, a few less things to order. I have to update my parts list with this order, and will try to get that done tonight. 

Next steps

• email alejandro and steve about machining
• figure out how to and set limits on the ranges of the PIA13's for the APY002 stages
• insulate photodiode
• test lockin setup by sending sine wave from function generator directly to the lockin A/I input, and make sure the phase difference is zero
• mounting heights for beam expander on translation stage
• mount apy002s with spacers that we now have
• update bill of materials with new thorlabs order
• (from today) Set up single beam system with correct alignment and space for the beam expander.
• (from today) Figure out angular acceptance of beam expander to evaluate the translation stage setup
• (from today) Photodiode test with beam expander system!
• re-align the beam with the more realistic set up!

Even though I didn't entirely hit the deadline today, I think I have made good progress and am moving towards scaling up simultaneously, so I should still be on track for this coming month. Chris said he has some time this weekend to come in and work; I am planning to come in for a few hours tomorrow and/or a few hours on Sunday (either morning or around 4pm). After my midterms this past week, I should have a less intense week for school, and be able to make some serious forward progress before we leave on spring break.

3.3

Got a single stage to move with the piezos. It makes a really loud noise, but this is mentioned in the product manual. The step size should be increased when moving, at least for coarse adjustment - the motion is extremely slow. On the flip side this means our alignment should be very precise! 

There is no obvious way to set limits in the kinesis software. I am going to do more digging on this, because it seems intuitive that there should be limit capabilities when the software can read out my position with such accuracy. While looking online I did see the forward limit and reverse limit like in the image below, but I can't get that to show up for these actuators.

The more annoying way to do it is to zero the stages properly and remember what the limits should be while aligning. However, this will be frustrating to do while trying to look at the camera to get the beam onto that software. 

I also just remembered that Chris and I talked Friday about getting a USB hub to make sure we have room for everything to plug into the computer. My one concern with this is that it seems like the camera speed is dependent on being plugged into the USB 3.0 port, so I would hope extending the other USB port would be enough to control all other devices.

I also heard back from Helen that the billing code is 370.31710.xxxx.016110.600200.44505, which I will give to Steve when I go in to see him this week. I also should return Jim MacArthur his oscilloscope because I am not currently using it.

3.5

I emailed Steve yesterday at 8am asking if I could come in to talk with him about machining. He has not replied to me yet, so I decided I would just go in to 38 Oxford St. and find him. I wandered until I got onto the floor of the machining facility, where I found two people who looked like grad students/postdocs. I asked if Steve was in today; they responded that they have been looking for Steve for two days. I asked if they knew at all how I could get in contact, and they then said that the only consistent way to find Steve is that he covers the Physics instructional machine shop for Alejandro on Thursday mornings 8am-12pm. I can get there between 8 and 8:30 to talk with him if he has not replied to my email at that point!

I also had a conversation with Eske about the computer hardware we will need for the windows equivalent of a Mac Mini. The main idea of scaling the computer down is we want a small motherboard; the form factor of a motherboard will give us the shape/size. I have inserted an image from Wikipedia below, but he said to look at micro/mini ATX sizing.

Once we have chosen a motherboard size, there are computer cases designed for that size, and the other things should scale. We need a CPU (which should have built-in graphics), a power supply (PSU), memory, and a hard drive for a functional computer, and potentially some cooling. I asked if Eske knew about any commercial off the shelf option, and he suggested that I check out Dell, and then that I talk to the people at MicroCenter who should give me a better idea (and even if they don't have a COTS part, will be able to give me the parts needed to build the computer).

He also mentioned that the ssh-ing will just require us to have ethernet in the dome, and that we will have to set a static ip (which might be annoying to find in windows settings). Then from the software side, there is standard procedure on how this will work.

From a quick search of the dell website it does seem like they have a small form-factor option. This could be an off the shelf solution. If I have extra time at the end of the semester I could see building a computer might be fun (and maybe? would save money), but that is a big if.  

Talked to my father as well who mentioned an Intel Nuc might be a good option for a small off the shelf computer. He also told me that we would not need to ssh into a computer, as we are not running any command line programs (we just need the Thorlabs Kinesis GUI), so our best option would be to use Microsoft remote desktop and remote in.  I can install and test this even on the windows laptop I have in the lab. Also, I should make sure I have windows 11 for professional/enterprise, as the home version may not have all the features I need.

Concern with USB switching which device I am talking to:

• unplug hub and plug it back in
• reboot computer and test if the devices are the same
• can i find a serial number for each device on the computer?

3.6

It looks like from the photo I have before where I have connected to the K-cube that there is a serial number displayed at the top right corner of the UI in Kinesis (labeled S/N - see photo under 2.2), and because we will keep the piezos plugged into the same ports, we will be able to identify which actuator is which if we categorize them somewhere initially.

For the problem with homing and setting limits on the piezos that I was thinking about this weekend, I don't think there are absolute encoders on the piezos, meaning that each time I turn on the software for Kinesis, it will set my position to zero no matter what point I am at in the actuator range. This is slightly annoying because the range of the piezos exceeds the range of the stage, so being able to set software limits would be nice.  Relative encoders should still work if my understanding of what we need is correct, and we have a few options for how to solve the issue.

→ solution options:

• We can "calibrate" to absolute zero the first time we use them and record what position we set to zero (we would run it through its entire range, and then have some measure). Then, on subsequent usages, record the position at which we leave the actuator.
• We have a mechanical limit switch that physically prevents the actuator from pushing the stage beyond its range
• We have an electrical limit switch that completes the feedback loop and stops the piezo from moving past a certain points

Other thoughts/concerns:

• Will this system be set it and forget it, such that I can make sure the beams are all aligned in the middle of their ranges for the most part, and then I don't really have to think about the limits as we will only offset the stages a little bit in each direction?
• The piezo manual states that the actuator steps are not repeatable because of the way the actuator works. If we are setting an absolute zero, we should really have another encoder somewhere that records position (Thorlabs recommends this as well)
• Will we need to realign the system multiple times, or will it just be small adjustments?
• If we set a mechanical limit switch (or something goes wrong with the electrical one), we need to know what happens when the system fails. Does the motor stall? Does it smoke? Does a spring in the stage snap?

Chris also sent me a note this morning about triggering the laser diode with an external strobe bit, and mentioned the I/O bit on the K-cube piezo controller. He suggested that I figure out how to toggle a logic bit on/off for a millisecond. I've seen the logic I/O configuration tab under settings in Kinesis, and can configure one of those ports from the K-Cube as an output. From reading the KIM101 Kinesis Manual, the logic level of the output is based on motor specifications (if the motor is at a limit, moving forward, moving at max speed, etc.). There is also a User IO connector that is labeled as "for future use" but seems to have encoder capabilities.

The last thing I was thinking; because of the way we are aligning the beams relative to one another now, I am not sure the flip mounts are necessary, as their primary purpose was to get a reference beam onto the camera surface and then block that reference beam during our alignment. This would reduce the number of USB ports we need by 2!

3.7

Spoke with Steve this morning in the machine shop. He can make the parts with a turnaround of 2-3 days, but then needs to send them out to get anodized, which is a notoriously (apparently) long process to even get a quote. The game plan is that he machines 12, 8 get sent out to be anodized while we work with 4 that are not anodized, then once the 8 come back we send the last 4. This will be the way that least hinders progress of scaling up.

Chris also sent me this link that seems to allow multiple K cube controllers to connect out of the same port and mount side by side; there are already mounting solutions included in what we bought for each k cube individually, but this could be a more centralized way to do it. I'm not sure it's necessary, or that it would necessarily make mounting cleaner, but it is a good idea.

3.8

Will want to have a conversation at some point today about the absolute vs relative encoders and how to solve it. 

→ talked, we are going to try to just set up the system physically so that the "aligned" position is in the center of the range of the tip/tilt stages. In this case we should never move the piezos beyond the range of the stage. 

The shearing interferometer has arrived, so we can now test our beam collimation! Also, the rest of the alignment keys have arrived. Will update the bill of materials later tonight.

Chris and I also talked about controlling the laser diode with a Lab Jack, as a function generator is not a feasible long term solution. Will work on figuring this out as well.

• Update BOM
• Connect up Lab Jack such that I can turn the laser diode on/off
• ask about flip mounts

I am trying to work with the lockin right now. When I do a frequency sweep of the signal generator and have the signal generator plugged into the lockin, the needles on the lockin move each time the sweep ends and begins again. For the remainder of the duration of the sweep both needles are around zero, which is great. When I end the frequency sweep and just do a regular sine wave, my output on the lockin is both needles at zero. I had to change the range on the lockin, but now I get a difference when I have one channel at 4 kHz and one at 8kHz the needles on the lockin show that this is true. I am now going to try with the photodiode/laser setup.

Tried and talked to Chris about lockin. We came to the conclusion that the Y output has some wacky offset, and we can look at the X output for information. To my understanding, the X output corresponds to the in phase component of the signal, and the y corresponds to the out of phase component. This means that if I am sending an 8kHz signal to the photodiode and have an 8kHz reference frequency, there should be some nonzero amplitude on my lock in channels for the product in Fourier space. 

I am now going to use this to test the beam expander. I set up the translation stage in Kinesis, and homed it. I am easily able to move it back and forth on the stage!

The process of testing the angular acceptance of the beam expander is rather frustrating. I had to figure out the mounting of the beam expander on the translation stage at a correct height for my setup, but also such that I can rotate it to make sure things are correctly aligned. 

I'm also getting some kind of reflection that is spatially different than what I would expect – I would expect it to be sent back along the beam, but I think the photodiode is not planar with the laser diode; is there a good way to test this? I asked Chris who said the best (read: only) way to do this is to mess with it until things are right. A nice alignment trick is matching the location of the reflected beam to the laser diode. We decided to permanently attach the 5x beam expander to the laser diode so the alignment through there works.

With some small collimation adjustment we could see the beam outputted from the beam expander system and read it with the photodiode! We will be able to move forward with alignment using the beam expander on the translation stage. The strategy will be to lock the position of the beam expander, then align the laser diode with respect to it.

Also brought home a LabJack! Have gone through the quickstart manual, and the software/hardware seems to work great! Will add the software to the list of things to install on the nuc. With the LabJack, I can configure one of the IO ports to be an output, and change the logic level from high to low and back to turn the diode on and off (having another port at +5V and utilizing the ground). I will test this tomorrow!

3.9

• LabJack into diode
•  Test collimation with shearing interferometer
• Align new beam using translation stage setup!!!
• Continue scaling up (everything at right height, etc.)

Going to begin with LabJack stuff. I plugged the ground into GND, the +5V into VS (which according to the datasheet is 4.75-5.25V regulated), and I plugged the modulation wire into DAC0. From the Kipling app, I am able to change the voltage of DAC0 between 0 and 5V as the modulation signal takes TTL, which changes the brightness of the laser and gives me the capability to turn the beam on/off with software. Works great, and now I don't need power supply/function generator/breadboard/lockin anymore!

I will still need to mount this to the optical bench, but this should not be a big issue.

I am now going to try to align a beam from the start, using the translation stage, motorized APY002, full setup version of the system.

Weirdly a faucet in the back room just turned on, then off with no one else in the lab. Not sure if this is a normal water discharge thing, but it was a little strange.

I am actually getting a visible spot on the wall that is just less than 2" in diameter! I think this is because I am feeding in a constant voltage with the lab jack instead of a sine/square wave like I had with the function generator. Very cool and makes alignment much easier.

Something about the laser diode mount is not as secure as I would like it to be; I want to look into other mounting options for the SM1 tubes.

Changed my mounting to having the AD10F and the BE05-532 both mounted to an LMR1 mount as in the photo below. This will also make adjustments much easier than using the SM1TC mount. 

I put 15mm of spacers under the APY002 (note that the screw in the bottom looks like 1/4-20 but is M6). I am now working on aligning the beam with the beamsplitter and beamsplitter cube. One note with the beamsplitter cube – it seems like either the shadow of the piezo, the screw, or both are in the image (although sometimes on the very edge); I could update CAD to move the screw further out, but I have already sent the parts to Steve for him to machine. For now, I should be able to align with the smaller sized beam without any problems.

I'm amending my previous flip mount statement; I think we still do want a flip mount for the first beamsplitter (not the cube, but the plate). This will increase the intensity of our system overall by a factor of 2. Because the beamsplitter is on an angle, I have to be careful to keep the flip mount itself out of the beam.

I am able to align this way! I increased the step size of the piezos on the software, and it is not hard to tell which way the beam is moving. This seems like it will work remotely going forward, if everything is mounted initially in reasonable positions. Pictures of the windows I will have open while aligning (note that the Kipling app will be running on the same computer as the others, I just have not yet installed it on the Lenovo), and the setup, are below.

When I put the beam expander back in the beam, it was 1mm off from before - I will test again to make sure this was just me bumping something in the setup while working and not that this is not repeatable. My next deadline is the end of March, at which time I want to have aligned a multibeam system. I also need to check that we have enough spacers to mount three additional APY002 stages (although these are something that could be machined quite easily on a lathe in an emergency).

I'm noticing that there is no difference in what the camera reads of the laser diode output whether I have 5V or 1V on the LabJack. So, it seems like brightness can only be controlled by modulation; there is probably a way to program this onto the lab jack.

→ Sure enough after a quick search, we can do that quite easily according to this page. I will work on coding that next!

I also wanted to check out the shearing interferometer today. This video from thorlabs was very informative about how shearing interferometry works.

Measurement of the beam with the shearing interferometer was interesting. I decided to use the small interferometer and the beam that only goes through the 5x beam expander (and is between 3-5mm in size). I used the viewing plate as the beam is very small. I still could not entirely make out fringes, and the collimation adjustment from the beam expander was too small for the fringes to appear at any point in the range. Next week I will try using the collimation adjustment on both beam expanders and the larger shearing interferometer, which hopefully gives us room to fully collimate the beam.

The beam collimation is important because we are relying on the fact that parallel light rays will end up at the same point on the focal plane to do this testing. If the rays within our beam are all reaching the telescope at different angles, we will get a messy blob on the focal plane and we won't even be able to distinguish between points, let alone the differential motion between points.

I was also thinking - do we have a backup laser diode? I was looking at the specs online today for the same model and it said lead time was 8 weeks. Nothing is currently wrong with the one I am using, but it would be really tough if it broke and I had to wait that long to keep working. I was thinking the edmund optics one that Chris showed me a month or so ago could work with minor adjustments to the system.

Overall today was very successful - I am glad I came in! Planning to come in next Saturday (Mar. 16) by which time I hope Steve will have machined the four beamsplitter cube holders I need. I am also running out of length of optical bench, so will either need to compress the setup in the x direction a bit to fit all beamsplitters and mirrors (which can be done) or I can open one of the new optical benches we have and make sure they work. 

Another thought - the one reason the vertical mounting for Kcubes would be helpful is if we run out of space on the optical bench, and it reduces the number of wires everywhere. Separate thought about K cubes, but there is a power switch on each controller itself; how often do those need to be power cycled? Can they be power cycled by turning off the power to the port or is the flipping of the switch important for some reason?

Questions (compiled from above):

• do we have enough post spacers for all the beamsplitter cubes?
  • for some reason here my bill of materials does not match with what we actually ordered and what was in the cart I sent to Chris. We only have 5 spacers that are 5mm each for the beamsplitter cube posts. I will need at least 4 10mm spacers in addition to these. This was an oversight on my part with all of the mounting math that happened at the time, and now that I am readjusting heights based on the translation stage, more spacers are necessary. I will send Helen a note to order. (resolved)
• should I consider vertical mounting for the k cubes (like Chris sent me here, used with an angle bracket)?
• will the k cubes need to be power cycled at any point or can we just leave them on? (hearing back from Thorlabs by Monday)
• do we have/need a backup laser diode?
• is it normal for that one sink in the back of the lab to randomly turn on and off? (yes it is)

Next steps:

• large shearing interferometer (to collimate)
• (moved to Friday) modulate laser diode signal with labjack (to change brightness and get more precise positioning on ThorCam software) 
• (moved to Friday) get laser diode to have a TTL driven pulse of ~1ms through LabJack 
• mount more beamsplitter cubes
• open longer optical bench and set up on it

3.16

Reached out to Thorlabs tech support about power cycling the k cubes. They said it's not good to leave them on indefinitely (which I would tend to agree is not great), but that he was not sure if there is a difference between cutting the power from the cord and flipping the switch. He is reaching out to colleagues who work more closely with K cubes, and will get back to me.

I picked up the four machined beamsplitter cube holders earlier this week, and the new spacers that Helen ordered have arrived, so I am going to set up the four beamsplitter cube stages.

Beamsplitter cubes are set up. I accidentally did one too many, and put them all on piezo stages; I really want two mirrors on the same stages instead, and the third beamsplitter cube does not have to have remote tip/tilt capabilities. Remounting those now.

This is the updated mounting! I do need Chris' help - I tried to flip a piezo on the final mirror while mounting, and I can't seem to push it back in, so I have left it somewhat up. That should not impede my progress before next Friday.

As far as optical benches go, the MB648 that we are planing to use is relatively narrow, which means I have less room to mount the camera, the motor controllers, the third beamsplitter/second mirror, and the translation stage. However, the majority of the setup is in a straight narrow line. I will play with having the benches side by side vs. straight in a line.

Next, I want to do the shearing interferometer test with the full beam (beam that goes through both beam expanders). To my understanding, I should be able to change the collimation of the beam by turning the red knobs on both of the beam expanders, and by adjusting the distance between the beam expanders. I will try this now before switching to using the MB648s.

Note that I always will want to home the translation stage before moving it to a position - the position resets if the device is not homed.

Weirdly even with the larger interferometer (using the larger beam expander to have more control over collimation), I can't get an interference pattern to show up. I will ask about this when I see Chris next.

Moving on to the LabJack stuff!

To get pulse width modulation, I will have to use the FIO0 or FIO2/FIO3 ports rather than just setting a voltage on the DAC0 port. 

The software that shows you what your code is doing is windows only, so this seems like a good time to install labjack stuff on the lenovo. 

The labjack is giving Error #1230 in LJStreamM software, indicating that the device is occupied by another process than the one I am trying to measure. I am going to do some more online searching, but I am not quite sure what this message means. I will return to this on Friday.

3.22

Talked with Elana this morning - she mentioned that another way besides the duty cycle to get a smaller spot on the camera would be to use an ND filter. This is a good idea especially at this phase of testing to try out; I do think that the duty cycle method would allow us a higher degree of adjustability, and it does seem like something the LabJack is designed to be capable of even if I am having some trouble implementing it.

We also talked about collimation! She said that sometimes, she has found shearing interferometers finicky, and that the best way to do a rough collimation is to send the beam across the room and adjust its size with the dial so that the size where it exits the beam expander is the same as its size at a distance. I am going to try this now.

So I found that with sending the beam 25 feet, there is a maximum of collimation adjustment allowed by the first beam expander. At the site of the beam expander the beam has a 4mm diameter, and at the wall the (highest flux area of the) beam has a diameter of 8mm.

Now, testing with the larger beam! The tricky part of this is that the distance between the two when we mount them will also make a difference, so I will measure what distance gives the optimal collimation and replicate this on the mounting side. 

Note: I remounted the beam with a 9mm spacer, and then moved it up a bit with the set screw to get the full diameter into the beam expander.

With the same setup but now both beam expanders, I was able to get the same size beam at 0' and 25'. I included pictures of the spacing between the beam expanders while mounted so that I can mount them again. I have locked the red knobs on both beam expanders, so hopefully the beam remains collimated. But it does seem like I can collimate with only these degrees of freedom!

Now moving on to mounting things in the equal flux configuration.

Right now, mounting of all components works with the two MB648 boards side by side:

The one problem with this setup is that the camera is not mounted onto the optical bench. I may be able to shift everything down and get it on (moving the KIM101s and LabJack). Alternatively, I may be able to flip the beamsplitter+flip mount and put the camera more towards the bottom of the photo, between the bottom beamsplitter cube and the mirror. I don't like this as much because it is closer to the true optical path where the light will be traveling during alignment.

The Intel nuc would go on this setup as well; it seems like there would be enough room for it by shifting the things in the above image to the right, but the spacing between our beams also matters. I would like to discuss this with Chris. The other item that is not shown is our corner cube, which we said we would have on the ground while aligning. If my interpretation of this is correct, we could put it on a tripod-like structure so that it reflects back the beam and I have my hands free to work with the computer.

Also, I talked briefly with Eske about LabJack, and he mentioned they have a Python interface! I will look into this as it would make the software side much easier for me.

• Modulated signal on LabJack
• 1ms pulse (LabJack)
• Play with ND filters

The black anodized beamsplitter cube holders came in and they look so much cooler so I have taken another picture of the more finalized setup from a different angle:

This example seems relevant to signal modulation: Examples/More/DIO_EF/dio_ef_config_1_pwm_and_1_counter.py

I was able to run this script in python using VSCode (on my personal laptop). I initially got an error saying my device was being claimed by another process (error code 1230), but this was resolved when I quit the Kipling application. I am going to plug things into the correct ports (need the modulation to be in FIO0 to receive the signal) and test this out with the camera!

So the camera only registers that there is some difference in modulation when I put an ND filter over the lens (even with tiny duty cycle and frequency):

I'm not sure if this is because there is just so much flux initially that the cells are overloaded. This should be less of an issue when actually aligning, as we have beams that are coming back through a series of beamsplitters and will lose a LOT of flux before reaching the camera. However, it will be good to set the modulation for each beam here proportional to the number of beamsplitters it travels through so that we have an alignment procedure.

With the correct math, this program can also be what we run to turn the beam on for a finite time (1 ms pulse). I will need to figure these specific numbers out, but the architecture is correct! I will reach out to Chris to maybe set up a meeting sometime over the weekend, but I do think I am on track for hands off alignment next week which means we should begin thinking about next steps and final testing of the system. Hooray!!

Sometimes I have noticed that the beams seem to shift between when I use them (the beam expanders are no longer aligned with one another). I can't tell if this is because I accidentally hit stuff in between setup and use, or if the mounting itself is loose, but I will test this next week with the other alignment things.

The beams:

End of day questions:

• What are the complete specifications of the test our system must undergo?
  • Which telescope will we be using?  We will use Auxtel!
  • When can this happen? Potentially in May (week of 12th-17th)
  • What will the mounting in this telescope look like? Mounting logistics are shown on the Aux tel installation notes page that Chris made;
  • Using all four of the beams? Yes!
• Is there an Intel nuc in the lab somewhere that I can work on setting up? We ordered one on Amazon that should arrive at the lab tomorrow (Monday 3/25)
• When I plug in the camera for a long period of time, it does heat up quite a bit, and it seems like I will also have to power cycle the Kcubes somehow. Is there a power strip that can turn on/off with software? What about a USB hub? So there are power strips that turn on/off with a user interface that allows us to control when/how frequently. For USB ports, I will have to look into this for Windows.
• Will we be transporting this exact system to Chile, disassembling and reassembling, or sending new parts down there? Should we start looking into ordering things to Chile in case they are backordered? What threshold should the system pass before the travel is confirmed? We will send the big parts (basically just the optical bench) to Tuscon to be sent to Chile. Everything else, we should be able to pack up all together and take on the plane.
• Wire routing (zipties? wire loom/braided sheath?) Zip ties should work well for this! We will also plan to mount the K-cubes and the nuc on the vertical piece that allows us to mount to the dome, and will route the wires over/down from there.
• LabJack mounting?
  • was thinking of just drilling out the small holes that are already in the plastic so they are the correct size to accommodate 1/4-20 bolts, if this is ok. Will need some sort of fastener for the other end still. Drilling out these holes should be fine.
  • There are also slots on the back for hanging but these do not seem secure enough to me
• Attaching the two boards together? Alternatively, we can either: a) are just going to 
• Camera mounting alternatives: mount it above the configuration and reflect the beam up, mount it below the optical components on the piece parallel to the dome wall, or shift stuff so there is mounting room on the board itself
• Spacing between beams in this configuration? The spacing will be determined by the diameter of the primary mirror, the amount of the mirror that the secondary mirror blocks, and the beam size. Adding to next steps that I will draw a diagram with these things.

Up next:

• bring Steve other beamsplitter cube mounts so they can be anodized
• finalize modulation code and pulse code
• rotate piezos so that none impede the beam (Chris)
• Test vance wedges size
• Figure out beam shifting things (lock down mounting)
• FULL ALIGNMENT TEST! Set the system up, then align without touching anything (MOVED)
• Turn USB ports on/off with software (no longer necessary, as we can just turn the computer on/off with software)
• Switch over to intel nuc, download software
  • remote in
• Draw out entire configuration with primary/secondary mirror, beam spacing and size
• From the full system drawing and the length of the breadboard constraint, figure out distances between different components for mounting
• Come up with answers/have discussions about the above questions

3.24

Met with Chris this morning to discuss the last few weeks of progress and next steps. We were able to talk about the above questions, and I have put the answers above in green. I also added to my list of things to do!

For USB ports, I found https://superuser.com/questions/502046/how-to-disable-and-enable-a-usb-port-via-command-prompt and https://stackoverflow.com/questions/404097/power-off-an-usb-device-in-software-on-windows that I will look into more closely later in the week.

Some initial spacing thoughts:

Spacing on breadboard (we have 48" in total to work with, and we are somewhat limited by the length of the beam expander setup):

Mirror spacing on Auxtel (using the data from here):

(outdated) Math on where we should try to put the beams:

3.27

Setting up new computer! It seems like there is a way to mount it vertically with a mounting plate that came with the device. Attached a keyboard and mouse to get set up. 

For some reason the device does not like my school email microsoft account; I will set up with my personal email. I remember having a similar problem with the Lenovo. 

I have downloaded the Thorlabs and ThorCam software, and I am working on setting up Microsoft Remote Desktop. I was able to enable it on the computer, but for some reason when I open it on my Mac, I get a blank window. Looking into this.

Talked to my dad about remote desktop's lack of functionality, and he confirmed that it is likely blocked on the Harvard networks. I will talk to HUIT about this on Friday (made an 8am appointment).

Talking to IT points:

• static ip address (subnet mask + default gateway)
• remote in
• ethernet

I will need to briefly borrow a monitor when we get to Chile to figure out what IP address is assigned to it, so that I can set up remote desktop again.

He also mentioned that it might be easier with a USB hub to turn ports on/off remotely. Another thing to check out on Friday. Did a brief order of magnitude sizing test for the wedges as well:

It looks like if we have an extra 2" or so in the dome, they should work as they are (and will not need to get cut down to size). Chris and I talked about having Johnny measure the distance while he is down there, which I think would be very helpful!

3.29

Talked to HUIT this morning, and they got my computer set up with ethernet access. I think the problem when I would remote in on Wednesday is that the computer disconnected from the wifi before (which is strange because I was downloading apps), and so I would get a black screen. This should not be a problem when using ethernet, which they were able to set up for me. 

Having trouble seeing the LabJack in Kipling on the new computer. Going to try to reboot and see what happens. 

Rebooted and having same issue. The LabJack also will not connect to my computer. However, when I run code from VScode the laser diode turns on? Not quite sure what is going on here. I am going to try to factory reset the device using SPC and a jumper cable.

→ Factory resetting worked! I don't love that I am unsure why this happens, and in the case it happens in Chile we would need to go up there to put a jumper cable in the device.

Had to restart VSCode so that it recognized the LabJack module that I installed.

Came back to computer after group meeting and the laser diode had the same connection error 1298. Going to unplug and replug using the jumper cable again.

Now it works! Wondering if I should just keep that jumper cable in, so that the device resets each time it is powered up.

Code notes:

The waittime parameter controls for how long the signal is on. 

The pctDutyCycle parameter controls effective brightness/time on in the pulse case.

The outputFreq parameter controls the frequency of the signal.

Beginning alignment now.

Pre-alignment:

• make sure light path hits all critical points
• take off all lens caps

Also wanted to talk to Chris about sending in the beams to the telescope.

Meeting w/ Chris

• Retroreflector will give us the freedom to reflect things back regardless of what angle they come in, but is sensitive to lateral displacement... 
• We do NOT need translation stage anymore!
• We will flip the beamsplitter in and send the beam to an off-axis parabolic mirror that focuses the beam into the camera
• This will somewhat simplify our system!
  • I do need to recalculate some mounting heights. we are back to the flip mount being the constraint.
• We can also send all the beams in on a chord of the primary mirror, so we will not need to account for the central region being out of bounds, because we will just compress all the spacing.

4.1

Met with Johnny this morning to confirm which dome measurements are helpful to know for the project. Since to my knowledge we haven't decided on an exact height for dome mounting, he will take a measurement on a low rib, a rib halfway up, and a high rib. He also mentioned that he is going to chat with the engineers about my mounting scheme and the best way to install it safely. He also recommended that for testing initially in Chile, we could install it on the low rib so that we could still tinker with stuff if we absolutely needed to.

I also heard back from ThorLabs about the power supplies; they recommend not turning the K-cubes off using the switches (which they would). Their message is below:

I think we will have to take our chances.

Heard back from Johnny - he said there is more than 30" of clearance for the setup, which is amazing! He also now has the SolidWorks files for AuxTel, and will send them to me tomorrow.

Updated task list:

• remount things at correct heights
• mount camera in correct position
• mount lens and test it out
• alignment procedure (moved)

making system better after alignment procedure complete:

• improve collimation (now have to re-do because beam expanders are connected) (moved)
• beam stop for beamsplitter's incorrect direction
• power outlet/ethernet on inside of the dome
• figure out power outlet setup - automate this?
• cut down wedges (moved)
• learn to do auxtel data analysis (moved)
• set computer to never go to sleep

4.4

Have received an off axis parabolic mirror and mounting supplies! Remounting the beam expanders threaded together.

Running into an issue where with the beam expanders threaded together, somehow the beam is coming in on an angle so only part of it is showing up. When I change the angle of just the diode and its mount, the full beam comes through, so I believe the problem is at the beginning (either with the mounting adapter or the threading). Going to try a new mount now. Have tried connecting the two with a lens tube as well with a very similar result. I'm going to try a new AD10F adapter next.

Tried a new AD10F with the same result: 

I think maybe the laser diode is not sending light out straight? Not entirely sure what the problem is. I am going to try to set everything else up and figure this specific problem out tomorrow.

In the meantime I will set up the picture that Chris drew on the whiteboard:

Drew a new whiteboard picture:

Current setup:

Things to do pre alignment:

• figure out mounting error on beam

4.5

Talked to Chris this morning, and he said that the most likely the beam is coming at an angle out of the laser diode itself. There should be a mount with angular adjustment capabilities lying around the lab, something like these (SM1 threaded) or these (smooth bore). 

I have found a KM100 mount in the back. However, it is made for 1" diameter parts, not SM1 threaded parts. I machined down the AD10F to have a 1" diameter (it is no longer threaded), and attached it to KM100. Will now make sure that it gives me enough freedom to send the beam through both beam expanders.

The beam adjustments are perfect for getting the beam sent through the beam expanders at the correct angle! However, the beam expanders must be at the correct angle in the first place.

It is quite tricky to get this mounting correct. Things I have thought of trying:

1. Re-set the heights of things based on the camera or beamsplitter flip mount
   1. not sure if this would actually simplify anything
   2. Just worked on this for 10 minutes before realizing that the only way to attach the spacers and still maintain correct heights is through the alignment key at the bottom. Going to rework the heights we have (with alignment keys) to make sure everything is mounted properly.
2. Change the mounting height for beamsplitter cube (a little bit of the top of the beam is cut off)
   1. this is not my main concern right now. I want to do a proof of concept alignment first.

I think the procedure to set this up will be:

1. Put the beamsplitters on the optical bench, as straight as possible
2. Align the beam with the beam expanders.
3. Put each beamsplitter on individually, and for each make sure light directed back with a corner cube can hit the camera.

I am also running into an issue where moving the corner cube does move the position of the beam, so I have to position it just right for the beam to come back and hit the camera. We added the off-axis parabolic mirror to solve the alignment problem once the light gets to the camera, but getting the light to the camera in the first place might prove to be tricky unless we have almost impeccable alignment to begin with. 

I am going to go back to the one beam alignment, and work with the parabolic mirror to get a single beam aligned. First, I am going to double check that all my mounting heights are correct, as the beamsplitter cube cuts off a little bit of the beam right now.

Final Mounting Heights

It was a good thing I redid these, because I had some things that were a few millimeters off in height. I now have this as reference for when I rebuild the setup.

Now setting up the single beam. I can see that the OAP does reflect all the beam to a certain point, but then it re-expands. I have to make sure the camera is at the mirror's focal length. The mirror I have set up has a 6" focal length, and even from eyeball alignment this creates a little tiny dot on the camera: 

The current setup looks like this:

Some notes on single beam test with this setup:

• It does seem like there is still sensitivity to lateral motion of the corner cube (see video below)
• There is an adjustment cycle with setting up so that the beam travels down the optical bench like it's supposed to.
• Tilting the mirror affects the angle the camera should be at (which makes sense)
• I'm not sure adjusting with the piezos is actually changing the beam position on the camera. Going to try adjusting the micrometer beamsplitter cube to confirm.

Video taken sliding the corner cube laterally (along the bench):

Going to look into off axis parabolic mirrors more over the weekend to be sure I understand what they are actually supposed to do.

4.6

Was thinking about alignment this morning; the lateral motion of the corner cube causing movement on the face of the camera doesn't matter as much if we can differentiate between parts of the beam with different brightnesses. Said differently, if we can see the point of maximum intensity on the beam and align the corner cube such that the individual point is at the center of what is reflected back, then that alignment should be consistent across beams.

As for off-axis parabolic mirrors, this diagram from Edmund was helpful in understanding:

The first few minutes of this video were also helpful in understanding the requirements of alignment.

The video also showed the use of post clamps, which allow more precise height alignment and rotational degrees of freedom. I know we have a few in the lab; they are definitely something I should consider for improvements to the system.

4.7

I am still confused on how the changes in angular position of the beamsplitter cube. Testing things now. We tested that the reflection off of a mirror works to send the beam back, and that rotation of that mirror corresponds to translation on the plane of the camera. The case where it does not work is when we use a corner cube.

Talked to Chris about alignment issues. We figured out that indeed, the angular position deviations are "canceled out" by the insensitivity of the corner cube to angular variation. So our alignment scheme as-is, even with the off-axis parabolic mirror, will not work to adjust the beams. 

Options moving forward are:

1. using the accelerometer on a phone to align the tip and tilt on the beamsplitter cubes.
   1. we would have to machine a hole in the top piece of the beamsplitter cube holders, and insert a cylinder with precision flatness on both sides to then place the phone on.
   2. the system would have to be tilted up at 45 degrees, because we cannot measure rotations around the gravitational force with an accelerometer, and one of the directions we care about is directly around g
2. Use the ghosts off the faces of the beamsplitters, aligning them with one another
   1. they will all show up at the same time depending on which beams we align
   2. we have AR coated beamsplitter cubes, so not very much light will make it back in the form of ghosts
   3. there will also be three visible ghosts per beamsplitter cube (3 reflection surfaces)
3. Strap a 1" mirror onto each beamsplitter such that the face is flush with the cube
   1. this would ensure that the angle we are returning at is only affected by the angle at which the cube is pointing
   2. we could just cut the rubber band post alignment
   3. requires total hands-on system access during alignment
4. Using a mirror on an optical rail to align 
   1. relatively rigid system, and allows us to translate one mirror between beams
   2. we would be off by the same angle in the mirror for all beams
   3. not as rigid as last method below
5. Using a mirror on a ball-bearing translation stage
   1. most rigid feasible system (considering 3 to be infeasible)
   2. we would need to find a long ball bearing translation stage - from preliminary thorlabs/edmund searches the longest I can find translates only 6"
   3. off by the same mirror angle in all beams, as in 4
6. Use a laser level to get coplanar beams, then do a 1D search pattern with the telescope
   1. we would need to make sure that the beams are parallel to the optical bench in this case and not just on a random diagonal line 
   2. questionable that this will give us a high enough degree of precision

For methods 4 and 5, we will need to bolt the optical benches together, put the translation stage on the second optical bench for alignment, and then unbolt to mount the first optical bench on the dome. 

Remounting notes:

I have found 14 post collars in the lab that are not currently being used. I am going to put these into the system as an alternative to the alignment keys and post spacers, which will allow us the rotational degree of freedom while keeping the heights constant. I am also going to remount the flip mount on a 90 degree post system so that it comes into the beam but the bulky part does not block the beam no matter how we want to rotate the lens. The 90 degree mounting scheme also allows us to come from above or below the beam. This will fix the issue I had with wanting to minimize the vertical movement of the beamsplitter cube tip/tilt mounts.

During the alignment process today, it was helpful to reconfigure the exposure time in ThorCam app (under settings), and rescale the pixels (the black/grey/white box icon). 

Todos from breakfast:

• 24" translation stage (on ball bearings, as the rail might give us more degrees of freedom than we would like)
• make an LSST summit account to get access
• try to get a longer focal length lens/mirror
• make sure new mirror system beam is not blocked when it arrives (set up and test)
• software architecture
• code to find centroid in python (moved)
• recalculate mounting heights (control for height of beamsplitter cube)
• remount all components at correct heights
• full alignment (moved)
• image quality meeting slides (presenting Friday May 19) (moved)
• CAD mount for mirror on translation stage
• manufacture mount for mirror on translation stage

Notes on content of paper

Instrument description paper

Introduction: why are we doing this, what is dome seeing about, lightning review of dome seeing in the past (reference Stubbs paper and other paper Chris mentioned today)

Citations to AuxTel/Rubin design

Say we are making a prototype for the system that is to be installed in the main dome

Talk about the importance of short duration imaging capabilities, which allow us to freeze atmospheric effects

Compare to DIMMS - DIMMS do on the sky what MOSS does on the ground  (or we are building the transmitter to which regular DIMMs are receivers)

Figure 1 - conceptual diagram; try to finish the CAD once design is truly finalized

Describe and rationalize design choices (out of all the ways to do this, why did we choose what we did)

Preliminary results, including measured differential motion data from Chile

I have started a brief outline of the paper, which is accessible here. 

Notes on Chile planning

• I will now spend two weeks down there, arriving May 5 and departing May 18. 
• Elana and I will install the setup initially, and Chris will come down the following week for calibration and data analysis
• Chris will talk us through the bolting into the dome mechanics pre-departure
• Meghan needs to find steel toed shoes
• We have less than a month before departure to create and finalize an alignment procedure.
  • Meghan will be taking exams and writing papers for at least the final week of this time, so realistically this is less than it sounds like...

Updated mounting heights

4.8

Chris emailed that we will need to make an adapter plate to sit the mirror on the translation stage. I will CAD that today and try to machine it tomorrow. Added to the to-do list above.

The dimensions (according to Chris' email): a 2" side length square, with 25 mm separation between M4 through holes and a 1/4-20 tapped hole in the center.

Quickly CADed the above piece, which should be very simple to machine. However, I am completely going based off of Chris' numbers - want to confirm that I have things correct according to the Amazon vendor before I proceed with manufacturing tomorrow and Wednesday.

Have to do mounting heights for two new mirrors (see Thorlabs order) and the translation stage with the mirror.

4.9

The stage is this one. 

List of Components we plan to leave on the summit is here. Note that this can also be used as our packing list for Chile.

4.10

I will have to re-adjust the mounting heights based on the stage. Assuming we will use the stage linked above, we have 

3.1496 (stage height)+ .25 (mounting plate) + .61 ( post holder and BA1S) + .75 in (minimum post height) + 1.38" (height to center of FMP2 for mirror optic) = 6.13"

which is much higher than I expected to go. This just means the tip/tilt stages will also need to be propped up; I can machine spacers for them on the lathe. 

What height will they need to be?

For FMP2:

2.45" (height up to platform of APY002) + 1.38" (height to center of FMP2) = 3.83"; 6.13" - 3.83 = 2.3"

→ need 2.3" of extra space. I can machine a 2.3" post spacer, but the longest M6 screw we have is 2" long so will not reach from the bottom through the spacer to the post to lock everything together.

I could alternatively put a short optical post on the APY002 stages that have mirrors, but this would not work for the beamsplitter cube ones. I could machine a counterbore for a 1/4-20 in the mounting plate for the mirror, so I only need one screw and can lose the 0.61 inches from the post holder and stage. A combination of this and a 1" mirror (which would reduce the height of the mount to 0.87") would reduce the height to 5.01", meaning the FMP2 spacer could be thinner.

Emailed Chris asking if the above changes would work. In the meantime, am going to try to get the new Thorlabs things at least on mounts, even if they are incorrect heights.

Chris suggested mounting the thing sideways, which puts the center at the top of the .75" post. With this and the counterbore, we have 3.1496+.25+.75 = 4.1496, which should work with minimal additional machining (as the 3.83 height will need to be brought up .3196" = 8.117 mm. I may need to move the post over to the side, or use a longer post to mount.

The LMR75 is too thin for the mirror; the retaining ring does not fit inside with the mirror. I could use the FMP3 instead.

Heading to the machine shop to begin work on the plates. Planning to make two (one with threads, one with counterbores) so that I can swap them out if need be.

The hole locations (to be programmed):

Soon after beginning to counterbore the center hole, I realized that you can't do a 1/4" deep counterbore in a 1/4" thick part (duh). I will do the two side holes tapped, and can quickly remake another mounting plate that has all three tapped (if we need it made). It only took about 2 hours start to finish doing it the first time.

This changes the heights somewhat, but not by too much - it adds 0.61" (BA1S and post holder). I could also do another right angle adapter and mount it coming down if this height change is too big of an issue.

Chris said to glue the mirror to the mount with 5 minute epoxy. Surprisingly, the cafe had very nice tools for the job - a coffee lid worked as a mixing tray, and the knife was able to spread the epoxy around the lens mount.

After 5 minutes I did a shake test, and the mirror did not come out. It is still slightly sticky though.

Finished machining the adapter:

I am not sure there will be room for a BA1S, so I still might have to work with an alternative mounting scheme, but at least I should have options when the stage arrives. I need to mount another mirror as well to put on the translation stage (FMP2 mount would be ideal).

Also, I'm not sure what comes with the translation stage in terms of motor controllers, but it does look like we need some wiring; depending on what kind of signals need to be sent, I may want a LabJack extension board.

The translation stage arrived; it is extremely long:

I do think I found a mounting solution for the stage; I can screw in a post holder from above if I get the correct sized 1/4-20 bolt, and then use a post collar to hold the post above the screw. I might want a slightly longer post holder to do this. This takes the additional .61" down to .31", and my machined spacers can be smaller. 

They would look something like this:

Also, I'm not sure what comes with the translation stage in terms of motor controllers, but it does look like we need some wiring; depending on what kind of signals need to be sent, I may want a LabJack extension board.

4.11

The smaller, manual rail arrived today. I am somewhat tempted to try using that first, so that there is minimal increase in complexity in the system at a time, and if that works it would be simple. However, I think Chris' argument that it is inherently less stable makes sense, and there will be a high degree of precision required for alignment.

If I were to use this rail, 6-32 screws fit counterbored through the holes, so I would have to have some sort of 3/8" aluminum mounting plate with some 6-32 tapped holes and some counterbores for 1/4-20 bolts that could then attach it along an optical bench line of holes. I would also need an adapter plate to mount to the top, where there are M3x3.5 holes, spaced 20 mm apart in both directions; I could essentially make the same exact plate I made for the other stage, with 20mm spacing between M3 holes instead. I could counterbore the M3 holes for ease of mounting.

Motor controllers for the other stage have also arrived.

List:

• mount additional mirror on translation stage
• mount translation stage to optical bench
• figure out computer wall mounting, KIM100 wall mounting, LabJack mounting
• machine adapter to allow mounting between translation stage and optical bench
• preliminary test with mirror (instead of corner cube) and new elliptical mirror setup
• remount everything with post collars instead of alignment keys
• IPA account

4.12

I am first going to make sure everything works in this new alignment scheme, not using any translation stage for the mirror. I am out of FMP2 mounts, so I will epoxy the smaller mirror to an LMR2.

As far as mirror alignment goes, I found it helpful to put a post collar on, set the height, and then set the rotation based on the angle of the beamsplitter (hold an index card in front of the elliptical mirror). Due to spacing reasons I might need to mount the elliptical mirror above the setup as well, but this should work with the camera. 

Image Chris drew on board of idea of new setup.

I am somehow still not getting any angular beam displacement with the elliptical mirror and spherical mirror setup. Going to try to use some optics software to figure this out. If it doesn't work I will return to the off-axis parabolic mirror to retrace steps.

This issue was resolved - I figured out that I had the initial beamsplitter rotated 90 degrees from what it's supposed to be. Obviously with this setup, turning the beamsplitter would not change the path of the light coming directly out of the beam expander.

Aligning with a single beam works! I can set the piezo in the lateral direction (even channel, side to side motion) and it moves very nicely:

Some notes about alignment:

• Make sure all components are flipped in the right direction!
• The beam does not hit the camera in this configuration. However, the ideal setup has the distance minimized between the initial beamsplitter and beam expander, so the camera can be at the focus of the mirror (and not hit the kinematic mount for the mirror.
• The flip mount does not entirely like being in its up position when gravity is pulling it down; will have to test if this works when motorized. I could potentially make it work coming up from below.
  • If it will be coming up from below, I can remove the post holder and use the 90 degree angle bracket to adjust the height. I will need to rotate the beamsplitter 180 degrees if this is the case.
• I can draw a circles (and add timestamps) on the ThorCam live window that stays up as long as the window is there; this may be helpful for coarse alignment. I can also see a histogram of the plot while it's up, giving me a more quantitative measurement of where the beam is on the graph.
• It is a little harder to see the movement in the up/down direction, which can show up more as a defocus. I'm wondering whether this will be more pronounced motion with a rigorous alignment procedure.

Setup photos:

Time to scale up! Last time I was scaling up, I remember having trouble in that my beam was not traveling entirely straight along the table. Procedurally now, this is somewhat interesting to handle because light is not coming straight out of the laser diode itself, so I have a relative alignment between the beam expander system and the diode in addition to the alignment of the entire beam itself. There are two ways of going about this:

1. Align the beam so that it travels straight down the row of the optical bench, then insert the beam expander into its path
2. Align the beam expander so that it looks like it's going straight down the row of the optical bench, then align the laser diode with this

I think I'm going to do a mix of both strategies: I'm first going to align the laser diode, then going to see how far off I am with the beam expander, and will adjust accordingly. Sadly, my alignment rulers do not arrive for another 10 days, but I will hopefully have them for Chile, which will make things much easier!

MFF102 mounting options:

I personally prefer the second option, and will just have to route wires out of the path of the beam with some zip ties. I am going to set the heights based on the standard height of the APY002; I took out the spacers and they should work as is. There is a ~5mm difference in the heights of the FMP2 APY002 stages and the BS031 ones; I am going to see how much light I lose by taking out all spacers, but I can always put the 5mm ones back in if it works better. Now that I am working with post collars instead of spacers, alignment happens much much more quickly!! I feel like I can change the correct degrees of freedom at any given time.

I thought of a mounting trick for the beam expander assembly; I don't feel great about having it only supported in one location, but I also don't want to over-constrain mounting such that the light path ends up on an angle. I can first mount with one optical post to the SM2RC holder, and can finalize that height. Then, I mount the second optical post, and slide down the post holder until the base touches the ground.

Photo of everything that will be left on the summit (sans some wires):

Photo of optical bench setup only:

Photo of optical bench setup with idea of alignment procedure shown:

Alignment Procedure

Note: assembly starting with mounted components, not necessarily at requisite heights.

1. Clamp second optical bench to main optical bench; mount translation stage and mirror on edge of second optical bench.
2. Screw laser diode into top left hand corner of optical bench. Align such that the beam is the same height traveling along a row of optical bench mounting holes.
3. Move beam expander into place. Screw mounting down. Minimal adjustments to laser mount repointing as necessary.
4. Put first beamsplitter cube on bench (in neutral position?). Turn on vertical laser level line. Align the center of the mirror with the center of the cube with the reflection on the wall from the mirror.
5. Translate the mirror, using the laser level reflection trick to mount the other beamsplitters and mirrors.
6. Use laser level to mount beamsplitter plate so that sends beam and reflection on same plane.
7. Insert the elliptical mirror and the concave mirror such that the light from one beamsplitter makes it to the concave mirror. Translate the mirror to check that this is true for all beams. Make adjustments as needed to get this to be true.
8. Fine adjustment stage: align in the order shown in the image below. Align on the focal plane of the camera, placing a small circle at the centroid of the beam you want to align to (and time-stamping it). See if you can get different colors for different timestamps
9. Detach the second optical bench

The details of the beginning and end of the setup procedure will depend on mounting height on the dome.

Was also thinking about stability and translation stages today. If I use the larger translation stage, I will need to use a few right angle connectors to mount the mirror from above, which might have the same amount of instability as the sliding stage. I think I might make adapting plates for both translation stages, and then I can pivot quickly if either doesn't work. Unfortunately the machine shop doesn't open to me again until Monday, but I should be able to get a good amount of machining done then. In the meantime I can CAD the adapters.

I've made a separate page with Chile reassembly notes and packing list (tools and components).

4.14

• CAD of adapter plate for sliding stage
• CAD of mounting plate (long stage)
• CAD of mounting plate (sliding stage)

It looks like there are a few required components for the linear translation stage to work:

Tried connecting up the SMC02 motor controller to the translation stage, but initially this did not work. I brought in a power supply to test it, and it worked when I plugged the motor controller into a negative voltage line, which after some searching it seems like the LabJack does not have the capability to do. Oddly, the motor only seems to move when I give the motor controller 10V or less, even though the motor controller says it can be used from 3-25V DC. Will now try plugging in the translation stage motor (before was using the stepperonline motor). This video was somewhat helpful.

The stage also does not seem to want to move with this motor controller. Will come back to this issue, but overall, we need a power supply solution if we are going to use this translation stage rather than the other one. Or, online it says that you can flip the leads to get a "negative" voltage; not sure I entirely trust that the LabJack has chassis ground but this is worth a try.

4.17

The screws on the top of the manual carriage are M3, and the counterbored holes are for a 6-32 screw and are spaced 25 mm apart. To mount this, I was thinking of using 3 2" blocks each tapped for 6-32 and counterbored for 1/4-20".

The mounting plate for the top of the carriage should have 4 M3 through holes, in a 20mm by 20 mm square, centered on a 32.4x27mm box. They will be counterbored.

The larger stage has M4 mounting holes for which I will need M4 SHCS on the longer side. Also, the holes are threaded, which is not ideal for threading into another machined piece, but I will see how it works.

Alejandro changed his schedule at the last minute this morning, so I didn't realize his meeting moved times. Going to get some parts done between 3-5:15.

CAD of mounting plates is in the main folder; images of designs are below.

Large rail mounting adapter for optical bench:

Small rail optical adapter plate to mount mirror:

Small rail adapter for optical bench:

Was able to get on the Rubin slack (Larom added me). Iain has not responded to my Jira ticketing issue email, so I followed up with him this morning. Note that I can begin the alignment procedure of the optics, locking things down up to the point where I need to place the mirror on the rail, so there is plenty to do even without machine shop availability right now.

Wall Mounting

Mounting the computer on the wall will look something like this: 

I will need to drill out the existing holes in the plate so they accommodate 1/4"-20 screws, and I can mount to the optical bench (I checked that the spacing works with the MB648). I am not sure if for depth reasons I would need flatter head 1/4-20 screws.

The KIM101 has a mounting plate compatible with optical benches that is already included. 

The LabJack holes do not line up with the optical bench. I can make a mounting plate for this, similar to what I have been making for translation stages.

I am also working on the slides for the image quality meeting.

Machined the adapter for the manual carriage, and also need to machine two of the support blocks, one of which I have started. I would like some shorter M3 screws so those counterbores are actually functional; I will look for them in the machine shop tomorrow.

• shorter M3 screws from machine shop

4.18

Iain directed me to rubinobs.atlassian.net to post my IHS ticket, which I was able to do. Anastasia replied that I now have Summit VPN access. I also received a note from SLAC IT saying I have been issued an SSO account, which seemed to be what the platform Elana sent me earlier (https://usdf-rsp.slac.stanford.edu) was wanting. I am still unable to log in, however, and have discovered that my issue earlier was probably because I was clicking the windows login button. SLAC said that even though the email they sent me told me to reply, the email address I replied to is not monitored and I should fill out a ticket at https://it.slac.stanford.edu in accounts and access; however, it repeatedly says that I have logged out when I try to. I cleared my browser cookies, and it then asked me to log in before telling me I had successfully logged out. I will submit a ticket for this as well.

I did machine two holders for the translation stage! They look like

They work! However, I discovered that my design assumes the relationship between the hole spacing on the stage is constant, were I to put these under both ends. As a temporary solution, I am using one of the blocks in the middle of the stage. I am using three BA2s to bolt the two optical benches together.

The PDU came today as well! Unfortunately, its mounting holes don't quite line up with the optical bench (only one does). I will CAD something and machine it such that this can mount to the vertical optical bench. This is not as time-sensitive as other things at the moment, but I will add it to the list. Alejandro said today that I will be able to work after hours if I let him know in advance and there is a shop monitor, which is very very helpful!

Elana added me to the AuxTel service planning channel, and I was able to begin watching the Jira ticket we put in for our installation at that time. 

I began the alignment procedure. Tomorrow I will need to start with aligning the elliptical mirror and the concave mirror so that light follows straight paths. I may need to shift the elliptical mirror and the plate beamsplitter to the left to make space for the camera to be at the focal length of the concave mirror. 

LIST

Urgent

• draw out full schematic for image quality meeting slides
• image quality meeting slides (presenting Friday May 19)
• ticket for ethernet for NUC computer
• SLAC account ticket
• FULL ALIGNMENT!

Slightly less urgent

• drill out computer mounting holes (MOVED)
• use scope to measure range of duty cycles out of LabJack (MOVED)
• conversation about observing strategy with declination + mounting conversation (Meghan, Chris, Elana) (MOVED)
• LabJack mounting adapter (CAD and machine) (MOVED)
• machine second support piece so that translation can have multiple supports (MOVED)
• improve collimation (now have to re-do because beam expanders are connected)  (MOVED)
• cut down Vance wedges (10 degree) (MOVED)
• learn AuxTel data analysis procedure with USDF account (MOVED)
• ND Filter Flipper (MOVED)
• Ethernet testing (MOVED)
• ask Chris what precise non-motorized slide he bought so you can add to BOM

(Unnecessary) Medium urgency contingent on slide not working

• get translation stage moving with motor controller
• machining to attach large motorized stage to optical bench

Very low priority (we are unlikely to have power in the dome)

• PDU setup  (MOVED)
• PDU mounting adapter (CAD and machine) (MOVED)

Moved all things with strikethrough to 4.26.

4.19

Put in the ticket to get ethernet access for the mini computer. Elana mentioned that DHCP address/wifi connection might be easier; this can be a backup plan, but we had wanted to use ethernet to prevent network connectivity from causing issues with the system.

• think about targets/tracking/non-rotating dome problems (Talk to Craig and Eric)
• figure out what angle we need to point (for heights)

Image Quality Meeting

I said that we were sending in a microsecond pulse (should have been millisecond), and I was not as clear as I should have been about the overall science that the system aims to achieve. I have edited the slides to make that clearer now, and other than that the meeting went well. Notes from meeting:

• Roger is going to investigate through what channels we would go in the future to procure scaffold mounting; he is also going to confirm the channels through which we get OKed for access and to bring in our platform and system
• Scaffolding is not feasible, so we will mount this lower
  • The telescope is not currently able to point horizontally, and the minimum pointing angle is ~18 degrees (we are calling it 20 to be safe)
  • Small-scale scaffolding could be feasible (because we are not going to the top of the dome), but this was somewhat unclear
• We will disable power to the motors that control the dome during alignment.
• People are saying that Aux Tel is a waste of time now that the camera etc. is coming, so logistically we should make it very clear we are pulling no resources from Rubin
• WE CAN BRIBE PEOPLE WITH FOOD!

Slides:

Discussions with Chris

We now have a mounting contingency plan for if the dome mounting does not work. Chris and I both calculated that the center of the setup will be at around 7 feet off the floor of the telescope (with adjustment for the beam sent ~1 foot lower than the center of the primary mirror), so either a) we have a mounting platform that is 3 feet tall, or b) we have a ground mounting setup on the top of a set of ladders (possibly with glorified lead screws to control the tilt of the optical bench). The stepladders should sit on a piece of plywood that we would have to purchase at the hardware store, as the floor is grates. The photos below show approx. calculation of heights:

Elana heard from #auxtel-service-planning that there is 400V power on the dome, so transformer setup on the dome itself is unlikely. In the case this doesn't pan out, we will use a long extension cord and a long ethernet cable to get power/network to the computer and power strip. This will make the turning off motor power in the dome essential, as the relative position of the cord will not change with the dome. We can also pick an angle at which to point where the declination is 100% vertical, meaning we can leave the dome in the same position for hours of measurement. However, some of these angles are at the zenith angle of the south celestial pole, which is high airmass. This strategy should be more fully fleshed out in a longer conversation with Elana and Chris.

We will need to optically attenuate the beam, which we can do both in the time domain and in the light domain using an ND filter. We will add an ND filter on a flip mount to our optical setup in front of the first beamsplitter cube. We will have an initial system configuration where we leave the beam on and don't attenuate it, and then when actually trying to take data we will attenuate. We can use the camera to calibrate the amount of light we send in to the main telescope, although we have to add back in the factors of two from additional beam splitting. Approximating the distribution as a gaussian with a height of unity, we know that the peak can have 100,000 photons, and the volume under the curve is 28 ~=30 (2*pi*sigma). So we can deliver a total of 3 million photons per beam per burst (100,000 photons per pixel per pulse). Procedurally then, we will have four 3x10⁹ - photon beams, and we will make sure we have measurable current.

We also discussed how to compute the system's angular resolution. My initial strategy of calculating the angle of a pixel was not exactly correct, as we have a blob of light coming into the camera. Notes from that are:

New todos added to priority list.

• Add another flip mount and ND filters to packing list

Chile packing discussion:

• Hat/sunglasses/sunscreen/lip balm/meds (UV high)
• Nalgene
• Towel (microfiber)
• Fleece
• Steel toed shoes
• Hard hat (or can use the ones they have down there)
• Headlamp (or can use phone flashlight)

I will be checking the pelican box with my luggage. It looks like I do have a checked bag included with my ticket. I should not have to purchase an additional checked bag. I should fill out the Chilean declaration customs form 24 hours in advance of my travel. Tools and gear (hose clamps) are in the shipping container next to Aux Tel (green duffel bag and toolbox). 

We agreed that we can do qualitative centroid-ing of the beam on the camera. However, we need to do qualitative attenuation calculations at some point (which can be done with the Thorlabs camera we already have, sans a few factors of 2!)

It is currently too bright for the camera to detect the beam, so I had to change the camera exposure time. I am still running into an interesting thing where the beam movement in the lateral direction shows up very distinctly on the camera, and in the other direction (up/down) it does not appear to move at all. Because we are using a mirror I don't think this would be a cancellation effect, but I'm not quite sure why this is the case.

Also, the beam only remains in the exact same spot on the camera if I slide the mirror to the exact same horizontal position, but this should not be that big of a deal (can be marked in multiple places and we can use the beam that comes off the back of the beamsplitter as a reference).

• update BOM

4.22

I now have an IPA account and access to the summit VPN. 

4.26

I also have a working USDF account as of yesterday. Want to do full alignment in the next three hours if possible (sans remoting in to the computer, which we will have to do but should be okay).

Today alignment info

Beamsplitter (#1): S/N 97103144, 1 (side to side), 2 (up/down)

Beamsplitter (#2): S/N 97103144, 3 (side to side), 4 (up/down)

Mirror (#3): S/N 97102990, 1 (side to side), 2 (up/down)

Mirror (#4): S/N 97102990, 3 (side to side), 4 (up/down)

Alignment notes:

• I am getting a bright dot and a dim dot on the face of the camera from the initial beamsplitter cube; one of them disappears when I block the light path to the mirror. I will count this one as the spot we truly want to measure.
• I am having trouble getting the spots that come back off the mirrors to get to the camera; the light really does not seem to like it for some reason. Going to try to get a single spot back from a mirror now.
• → I have realized that there is a relatively small range of motion inside which the beam will come back to the camera off the two mirrors. This just means I will spend more time doing coarse alignment for the mirrors!
• We are aligning around the beam from beamsplitter number 1. If that one is going to be moved, it should be moved first.
•  I think it's easier to position everything and bolt it down before connecting the wires to the Kcubes.
• I successfully have all four beams going to the camera. Woohoo!

Aligned beams!!

Note that I can change the exposure during alignment to get a more exact measure of the centroids. However, this works and is relatively painless. YAY!

Remote Desktop

Trying to remote in to computer now, but am getting Error code 0x204. Looked it up, which brought me to this link; I had to uncheck the box as described here:

1. On the remote windows PC, open "Run" and input "SystemPropertiesRemote.exe" to open up the System Properties screen.
   Make sure you have checked "Allow remote connections to this computer", and uncheck "Allow connections only from computers running Remote Desktop with
   Network Level Authentication" if your MAC pc is on the different network from the windows pc.

• I am still getting the same error code. I am connected to ethernet, so internet connectivity is not the issue.
• I tried using the ethernet IPv4 address instead of the wifi one, and this also did not work.
  • Going to switch back to the wifi one, try connecting the computer to Harvard Secure, and see if that allows it.
• It seems like the problem was that the IPv4 address had changed since the last time I connected! This should not be a problem with a permanent IP address in Chile.
• When I turned off the wifi, the computer no longer was able to connect, even though it is on ethernet... trying to re-connect to the device with the ethernet IP address.
• Ethernet still not working. Troubleshooting this, but it might be different issues than we would have in the dome.

I got into the setup window (which has the BIOS information) by alternating between Delete and F7, and now I have reset the computer to default to displaying what to press during startup. I also changed the computer so that the specified state it should go to when power is re-applied after a power failure (In the BIOS information, Changing State After G3) from G5 to G0. This will allow the computer to restart properly and we will not lose connection when we use a PDU.

Procedure to test ethernet

• Manually setting IP addresses to make our own little network (as a proof of concept)
• (192.168, 10.X, and 172.16-172.32 are private and non-addressable IP addresses)
• Disconnect computer from wifi and Harvard ethernet
• Change Manual IP on computer:
  • IPv4: 192.168.1.1
  • Subnet Mask: 255.255.255.0
• Plug into ethernet switch
• Disconnect Mac from wifi
• Set static IP on Mac (there is no DHCP server on the network - everyone on network needs to have a static IP address)
  • IPv4: 192.168.1.2
  • Subnet Mask: 255.255.255.0
• Plug USB to ethernet dongle into Mac
• Try to remote in to the computer (using static IP)

Even though no internet access, I should still be able to remote in to the device. This will be a test of whether it is a computer problem or something upstream that is the problem (a network/firewall within Harvard)

NOTE: I NEED TO RECONFIGURE DEVICES TO CONNECT TO DHCP AFTER THIS IS COMPLETE

If there is some problem with this, I can always go through the Google remote desktop browser extension, which allows your computer to be accessed from another computer that you are logged onto the browser for. This can pose a little bit of a security risk, as google doesn't let you create accounts without a phone number, and I would have to be logged on in the browser. 

Ethernet testing

• All settings on my computer and the Geekom computer seemed to be working properly
• I set up the windows machine to respond to ICMP requests (ping)

• When I turned on wifi on just my mac, using remote desktop and the ethernet switch worked!

My username is megha, password is McKay106! for remote login.

3 options for connectivity in Chile:

1. set a static IP (talk with network admin to make sure this works)
2. DHCP address with reservation from rubin (so that we always get assigned the same thing)
3. DHCP but they give us an FQDN

In Chile, I can duplicate the same profile I have for connecting here and put in whatever new static IP they assign. Also things will be easier to troubleshoot now that I have pinging enabled, and I can ping through Royal TSX itself!

Pulse Code

I did some testing with the scope of the code I am currently using for PWM/pulsing. It seems like the code works really well at PWM, but the length/number of pulses is somewhat variable with changes in the code:

I think I should go back a step and write different code for a single pulse and pulse modulation, especially now that we have the flexibility of ND filters. This page should be helpful in writing a pulse out. The clock has 80,000 divisions per second, so I should be able to turn on the laser for as short as 0.0000125 seconds (1.25x10^-5 s), so a millisecond pulse should be no problem.

I will look at this code in LUA and eventually command-line python. I first will test what 10 pulses looks like so that it's easier to see on the scope, and then I can scale down the number. I also will make a git repo so that I can track changes. 

ND Filter Mounting

I mounted another flip mount for the ND filters to thread into between the plate beamsplitter and the first beamsplitter cube. 

To put in the ND filter flip mount in the final assembly, I will want to increase the distance between the camera and mirror 4. This is more of a note for the real deal mounting, but I will still make sure I can get it to work with the current setup as well (as I need to pack exactly which components I will need).

Updated List

Machining

• drill out computer mounting holes
• LabJack mounting adapter (CAD and machine) (MOVED)
• CAD and machine second support piece so that translation can have multiple supports
• cut down Vance wedges (10 degree)(MOVED)
• PDU mounting adapter (CAD and machine)(MOVED)
• ANYTHING THAT NEEDS TO BE MACHINED when stuff doesn't fit together

Procedural

• use scope to measure range of duty cycles out of LabJack
• conversation about observing strategy with declination + mounting conversation (Meghan, Chris, Elana) (scheduled for Monday at 12pm)

System TODOS

• improve collimation (now have to re-do because beam expanders are connected)   (Will do in Chile)
• learn AuxTel data analysis procedure with USDF account (Can do on the way)
• ND Filter Flipper
• Ethernet testing
• PDU setup (MOVED)
• wedges to optical benches (MOVED)
• Make sure everything bolts together 
  
  • threading everything
  • turnbuckles/hooks

4.28

The holes to mount on the back of the labjack are 5.75" apart, so they would not fit nicely on an optical bench. Also, the holes are not large enough for 1/4-20 screws to fit into.

I think the ideal mounting scheme will probably be vertical (90 degrees counterclockwise rotation from shown in image above), with an adapter to mount whatever size hole that is to the 1/4-20 optical bench holes, and a zip tie around the bottom through other optical bench holes. Will cad the mounting adapter now. Note that the holes are 0.17 in diameter, so I could do 6/32 or 8/32 screws to attach.

#2 size holes are 0.082 diameter

4.29

I have confirmed that the turnbuckles and the hooks for the mounting assembly accommodate mounting our optical benches with no additional machined parts.

The wedges do look a bit more complex.

I machined computer mounting plate, but need to drill out the holes a bit more so that it works to put multiple screws in at the same time. I CADed the LabJack mounting adapter:

The mounting adapter for the Vance wedges, if counterbored, will have to be very thick. Also need to make sure the angles work in my head.

PDU

Going to play with the PDU for a few minutes, and CAD its mounting adapter.

• have to be careful when mounting the PDU that the circuit breaker does not face down

We did not find the 10 degree wedges, but Chris ordered 14 degrees and I will machine them. We also came up with a mounting strategy - I will counterbore angled holes from the bottom of the wedge so that the mounting can happen on top, and then will tap holes straight through so that things can mount through the bottom.

Questions

• 10 degree wedges location?
  • Will I do two adapter plates (one for the optical bench sitting on top, one for the right angle bracket supporting the wedges from below)?
• Steel toed shoes location? McKay 104?
• Machining access in chile? No.
• Is PDU something I should focus on? if so, same part in chile (with different capacity for power)?

We are going to secure the hooks with large washers so that we have a vertical degree of adjustment.

Requirements for code:

1. turn on and leave on
2. one pulse of duration x
3. pulse train at given freq with given duration

FINAL LIST

• Find steel toed shoes
• Fix computer mounting adapter (on mill)
• make labjack mounting adapter
• find 10 degree wedges
• mounting plan for 10 degree wedges
• Drill holes in wedges
• cut down wedges
• code
  • 1ms pulse
  • leave on continuously
  • modulation and length
    • input pulse length and frequency
• what combo of ND 200,000 photons
  • leave beam on continuously, expose for fixed time, and scale by ratio
  • what ND filter causes this not to be overexposed
    • 72% quantum efficiency - 72% of photons are detected and are the signal
    • 72% percent of the energy in the photon is turned into current- x Coulombs of charge
  • or exposure much longer than time

• PDU setup ?
• make PDU mounting adapter ? - last thing to be done

4.30

Worked on the software side of things today. I was able to get a single, very repeatable pulse down to microsecond accuracy. Trying to come up with a function for frequency /duty cycle that gives us maximal pulse range...

The duty cycle changes what percent of a pulse the laser is on, and the frequency changes the number of pulses per second. So with a given frequency f, the minimum length of pulse is 1/(100*f), and you can change the pulse length up to a point, where your maximum modulated pulse is 99*(1/f).

I will take the LabJack on the plane to play with this some more .

5.2

Wedges complete! Note that when mounting, the wedges should be first mounted 2" apart on the breadboard, then the breadboard should be mounted on the 90 degree bracket. It would be nice (if this is going to be done with everything already on the board) to suspend the center of the board while attaching the wedges to it.

5.3

Priorities today:

• 200,000 photon calibration
• PACKING everything and making sure it is organized
• LabJack will come on plane with me, or I will play with it Monday night

We tested the photon counts in the darkroom this afternoon. I set the camera up with a 1 ms exposure time, and we calculated that OD 5 (with 2 ND filters) will deposit ~10⁵ photons on the camera's surface.

Be ready to stay outside in freezing temperatures for a few hours!

5.7

Computer setup on Rubin networks

From the users network we can reach devices on the 170 network (we pinged 139.229.170.20 and got a response). My computer can thus be on a different network than the MOSShub computer.

IT initially planned to connect to the 160 network with a DHCP reservation and the FQDN of "auxtel-MOSS". We wanted to connect through ethernet, as it's more of a robust connection and I had thought there would be ethernet on the dome. Unfortunately, there is not, so I would have to go up on a stepladder to fix the connection should the computer become disconnected from wifi.

They put the machine under the local Rubin network, and then added remote desktop users under system → remote desktop → remote desktop users, for security purposes. This means the users have to have Rubin network connectivity. If users are remote, they will be able to remote into the machine with summit VPN.

We are encountering the exact same problem here with remote desktop where as soon as I remote in on wifi, the computer is kicked off the network. Because there is no ethernet in the dome this is now something we have to figure out.

• Same problem occurs when I log in and when people with Rubin administrator credentials log in, so it's not a network problem.
• They created a new local administrator on my machine (moss-ladmin) to test if that worked differently
• They have not had this same problem with an Intel Nuc that is in the main control room area, but it runs windows 10.
  • We tested on a different machine running windows 11, and it also worked. So it's not an OS problem.
  • We checked the system BIOS and there doesn't seem to be a security setting we can change
  • We could use their windows 10 machine, but Microsoft will phase 10 out sometime sooner than 11, and at that point the observatory will not allow its use for security reasons. So this would be an impermanent solution
• We are going to see if we can requisition one of the Intel nuc 12 pros that A2 has coming in.
  • She seemed to approve this!! Chris will bring another when he comes next week.

When the new computer comes, I will be able to set it up at night at the hotel, so I won't lose time during the day. Things to set up:

• download kinesis
• download thorcam
• download vscode
• download kipling
• check bios turn-off settings
• install python 3.12
• install labjack-ljm

We began setup of optical bench system today; we unpacked and assembled beamsplitter cubes and mirrors onto the tip/tilt mounts. We also put up the beam expander assembly.

When I get back to the summit I want to check that the light path goes along the optical bench - hopefully all components are secure on the optical bench.

• check heights
• check that light goes along row of holes on optical bench
• mount camera
• mount lmr75
• mount beamsplitter plate
• mount ND filters
• mount translation stage
• unpack computer
• spatial alignment
  • make sure that both flip mounts are oriented such that they will not cause problems
  • make sure the distance between camera and mirror 4 is large enough so that ND flip mount can get through

Code Edits

Changed the main script so that a plot shows up of the expected output from your inputs. This should help with visualization and which pulses we send in. Also began command line argument format. Updates are in the git repo.

5.8

Another table would be ideal today if we can find one that's clean...

Updated code! Was able to write PWM function and we now have command line functionality. The repo should be quick to pull from git once I have the new computer. Command line documentation is included in the file, but here is a snippet:

System weight

Freddy wanted the approximate mass of the system, so going to do that now.

MB648 - 15.18 lbs

APY002 - 1.05 lbs

BS031 - 0.9 lbs

BE20-532 - 1.32 

BE05-532 - .35

BB2-E02 - .25

CM750-500-G01 - 0.42 (not going to be mounted)

AP90RL - 2.25 lbs

Computer - 3.65 lbs

Labjack - 1.2 lbs

Kcubes - 1.26 lbs

Wedges - 9.9 oz = .61 lbs

Assuming lens mounts, screws negligible (0.05 lbs each essentially)

Approx. weight: a little over 50 lbs

2*15.18+5*1.05+3*1.05+1.32+.35+3*.25+2.25+3.65+1.2+2*1.26+2*.61 = 52.02

On-summit work

Tested code with laser diode - works as intended! Hooray. 

Finished mounting things. It does seem like there is a (very small) crack in the large spherical mirror, shown below:

Checking the beam off the table spacing now. It does seem like the beam is centered on that outside row of mounting holes, but this means that the edge of the beam is clipped at the corner cube. This may not be a big enough deal to fix, as we can mount the corner cubes hanging a little off the back of the system and the total photon count is what matters (losing a fraction will not hurt us). I want to ask Elana what she thinks but ultimately I think it is ok. 

→ Elana agrees that it is ok, and we can just use the clearance we have from the slots.

I will need to set up the computer now because to align I will ideally have ThorCam software, even initially.

→ I set up remote access to the computer with giving my computer a static IP and going through the ethernet switch, which I am now very glad we brought.

Have completed the full by-hand alignment! Updated to-do list (pre-scaffold):

• make sure all the piezos and motor controllers work (read: make sure they survived the journey)
• make sure flip mounts work
• attach wedges to bottom center of optical bench (see photo of attached wedges from before - 2" apart will fit in slots)
• set up new computer (detailed list of tasks above)
• harness fitting
• have Mario unlock the storage container
• calculate angular variance in this spread based off of pixel size
• commit latest code changes
• more 1/4"-20 mounting hardware
• re-locate short stubby wrench

When I repoint all the piezos to 1000, 1000 (yay they all work!), I get a spread like the one shown here. I want to calculate the actual spread in angle, so that we can determine what variance we will get by repointing all the beams up/down by the same piezo value to adjust for our wedges.

Realigning everything now, so it is aligned before I come back in tomorrow morning.

I think I want to just confirm the beams are aligned tomorrow before putting on the wedges. So the procedure tomorrow will be:

1. clamps from auxtel container (mario)
2. make sure beams aligned
3. unattach second optical bench, wiring
   1. take off camera and lenses that have to do with alignment (confirm)
4. suspend between table and stool to put on wedges
5. begin mounting on scaffold procedure
   1. mount one optical breadboard with hooks and hose clamps
   2. mount ap90RL
   3. mount breadboard with wedges on APR90L
   4. turnbuckles and things

5.9

Alignment test went smoothly! Beginning to take things apart.

Disassembled so that only required components remain; also unplugged all wires.

• shorten factory reset on labjack

Reconfigured mounting logistics with Freddy - we are going to use a Loctite nut on the back and thread through the front of the optical bench. This will both make mounting more painless and more robust, as vibration from rotation of the dome will not loosen the threading.

The mounted breadboard is at a significant angle already, so the wedges may be overkill for mounting. We will get up on the scaffold and test exactly what angle the 90 degree bracket rests at, but if it's something like 9-15 degrees already I would say the wedges are not worth using and we should mount directly to the bracket. 

Ivan has the new computer for me! He forgot to configure DHCP reservation though so we are dealing with that after lunch. Computer allows login with network username and pass (lsstcorp in 1password)

If you remote into the computer when someone else is logged in it will take a full 30 seconds to go through.

YOU MUST BE ON LSST-WAP TO CONNECT!

moss-ladmin login:

.\moss-ladmin
MO2024rubin

Need to go through IT to add remote desktop users

Pointing/On Telescope Alignment

We are basically able to visually align the beams using the opposite port to latiss, which allows us to see the pointing and I can hand-adjust the piezos. While the tabletop alignment was nice, this was reassuring that we can re-align using the telescope should anything happen. Photo of two beams on telescope wall:

WE HAVE THE BEAM ON LATISS!!! (in the very corner) Images on Rubin-TV → auxtel

Current pointing: 90.1, 25.9

In Rubin-TV: 90.14, 25.95

We are going to offset the beam (using script ATCS) in x,y (pixels) and we will see what this is in Rubin-TV

In x we are moving + 2000 pixels (+200 arcseconds in x) and 0 in y. (got an error trying to use offset atcs :

Error in run
Traceback (most recent call last):
 File "/opt/lsst/software/stack/conda/envs/lsst-scipipe-8.0.0/lib/python3.11/site-packages/lsst/ts/salobj/base_script.py", line 603, in do_run
  await self._run_task
 File "/net/obs-env/auto_base_packages/ts_standardscripts/python/lsst/ts/standardscripts/base_offset_tcs.py", line 206, in run
  await self.tcs.offset_xy(
 File "/net/obs-env/auto_base_packages/ts_observatory_control/python/lsst/ts/observatory/control/base_tcs.py", line 1145, in offset_xy
  bore_sight_angle = await self.get_bore_sight_angle()
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/net/obs-env/auto_base_packages/ts_observatory_control/python/lsst/ts/observatory/control/auxtel/atcs.py", line 1644, in get_bore_sight_angle
  angle = np.mean(azel.elevationCalculatedAngle) + nasmyth_angle + 90.0
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'elevationCalculatedAngle')

Re-took at same pointing just to confirm. We are dealing with slight movements of the system; probable that wind on dome has something to do with it, although it's unclear.

Set to 26 degrees elevation, no change in azimuth. Continuous beam on again. Result:

WOW WE HAVE MULTIPLE BEAMS!

We have very blobby beams. This is ok - we can work on collimating them with LATISS too. We are going to try to move in azimuth. We are at 90.1. We are going to move to 90.11. Result:

We are now going to try to collimate. 

Collimation

When I say up I mean rotation towards the dome, such that the line that the tick was at moves up in the vertical direction.

Turned the dial on the BE05 1 notch UP. Taking another image.

Image 19.

Trying 2 notches down now. img 20:

Seems smaller. going to try 2 more down. img 21:

going one up: img 22:

going one up so that it matches img 20. locking and moving to second beam expander (be20-532) dial.

Moving this second dial 1 up. image 23:

We think this looks smaller than img 20. We are going to move this dial up again (1 notch). Image 24:

We think this is definitely smaller! Going one more, again on large beam expander. Image 25:

Way better. Going up again. Image 26:

Better again!! Going to keep going one at a time. Image 27: 

This looks worse. I am going to return to the state of image 26 and lock. I also want to check that the blobbiness is not due to just thermal averaging, as we are currently taking a 5 second exposure with beams on continuously.

First, we are going to try a 1s exposure with continuous beams on. Image 28:

Because these look much better, we are going to try a 1/2 second exposure. Image 29:

These also look better. We are going to try a .1 second exposure. Note that we have had OD 5 ND filters for all of these tests. Image 30:

We are going to try 0.01 which might be past the exposure short limit of the telescope. (did not work - code got upset bc minimum allowed 0.1)

We will next try shortening pulse which should have the same effect. But we need to make sure we have the synchronization down. This will be image 31:

YESSSSSS! SO HAPPY! Note that we will get the fourth beam tomorrow by moving the entire system over.

We are taking out an OD filter - moving to OD 4 so that the counts for the 1ms pulse are a bit higher. This will be image 32:

Beams look larger because of the rubin scale according to Elana; radial FWHM according to image analysis is around 3 arcseconds.

We still want more counts, so we are moving to a new ND filter combo. We are trying OD 2.5. Still 1ms pulse, 15 second exposure time. Image 33:

Although they look super blobby again, we have better centroiding and quantitative analysis. We are going to try to get the third beam in the image so that we have some real data. We are still at 90.11 and 26. We are going to try azimuth changing to 90.12, and keep elevation. Image 34:

We think this is basically in the field of view.

Initial DIMM Testing

We are going to try to start differential image motion images. Images 35-42 are this initial testing. We did ./pulses.py -n 8 -d 0.000125 -f 0.125 -g in the command line this first time, intending 8 pulses, 1ms each, with 5 second exposure times.

These counts are an order of magnitude lower than the 30,000 Elana expected.

We also tried -n 8 -d 0.00125 -f 0.125, and the order of magnitude didn't really change (counts changed by a factor of 3). This was images 43 to 50. Unclear whether this is because my code is incorrect or there is some limiting readout factor.

Intel Nuc Setup

FQDN worked to login, while DHCP gave this error:

Using FQDN to test. Want to test that I can remote in to Summit VPN and then remote in to the machine (using Rubin-Guest network, which I know doesn't work to connect usually).

Confirmed connection does not go through on Rubin-Guest without VPN. Trying VPN now:

CONNECTED! Yay. This worked. Want to change the bios settings but can do that tomorrow - hopefully there are no power outages tonight and I can install all software from the hotel.

→ all software installed, need to get git working but easy to do remotely.

5.10

Duty cycle testing

It turns out that 1/80, not 1/8000, was the right duty cycle and I think Elana and I miscommunicated about how the frequency worked. This command-line prompt <./pulses.py -n 8 -d 0.0125 -f 0.125> should work to give us 1 ms pulses:

Checking the spacing now (way out zoom on oscilloscope):

we do have pulses every 8 seconds, and we have 8 pulses! (the tiny vertical lines are the pulses, 10s spacing)

Group meeting slides:

Chris explained phased readout quite simply; the shutters take about 100 us to move, and then expose for however long, and then readout happens while they are closed and takes 2 seconds. We are getting variability in peak height (see slide with initial 3d plotting), where the same amount of flux theoretically should come from all beams. Chris thinks this could be due to lateral motion and dust on the deuer , which attenuates the beams differently. We can test minimizing the effects of dome motion with shorter exposure. We also know all dome motion should move the beams together, while the effect we want to measure, thermal variability, should move the beams differentially.

Updated to-dos from group meeting:

• move setup laterally so that fourth beam hits the mirror
• Loctite screws in right angle bracket and on optical bench
• finish code debugging with Elana
• Finish mounting
  • nuc
  • kim101s
  • labjack
  • power cords
  • cable routing
  • plug everything in
  • record serial numbers
• data reduction pipeline that measures centroids and flux
• Focus telescope and adjust duration/ND filters -> focus and photons
• Play with exposure time vs. rate of photos vs. collection efficiency
  • ***Shortening pulse could freeze dome vibration motion (attenuation from dust). Test this!
• Scaling up: take 1000 images!

We are repointing this thing so after we moved the beams they are back on latiss.

Moving 25.7, 89.1 for image 10, 1ms pulse, 15s exposure,  (5/10/24). Nothing. We are going to keep the same pointing and try continuous exposure for image 11.

There is a brighter spot but it might be background; we are going to try one more time then align with the other port.

moving to 25.8, 89.1. image 12:

Super weird streak; we are goign to take another image here to see if it's repeatable.

New weird streaky thing. We are starting to think it's just moving a ton - going to take a shorter exposure on the camera.

moving back to m3 and 25.75, 89.1 so that we can have some spatial nasymuth port resolution. on wall alignment:

Got to 25.65 with the beams spatially.

moving to 25.55, 89.1 (tenth of a degree so full field of view), img 20:

Nothing.

Post meeting taking another exposure (image 21):

still nothing. We are now moving back to 25.65 and 89. Image 22:

Moving azimuth to 88.9 now (image 23):

Trying 89.2, 25.65 (image 24):

This is brighter, so we are going to try 89.3, 25.65 (image 25):

We are going to move to 89.25, 25.65 to try and keep searching. Image 26:

We are going to take an image with no beams so we see if this spot is actually what we are looking for. Image 27:

So I do believe that blotch is actually our light.

We are repointing to 25.75, 89.25. (all these were with continuous beam <./pulses.py -n 1 -d 100 -f 60>). Image 28:

Going to try 25.55, 89.25 (image 29):

Seems like that was the wrong direction. Elana is going to plot where we've been so we can have some direction as to where we go next.

Trying 89.1, 25.75 (30)

Trying 25.85, 89.1 (image 31):

Dropping to 5s exposure and going to 25.95, 89.1 (image 32):

Dropping to 0.1s exposure, going to 26, 89.1(img 33):

25.97, 89.12 (img 34):

25.98, 89.13 (img 35):

25.99, 89.13 (image 36):

going the wrong way - trying new:

25.96, 89.13 (image 37):

25.95, 89.14 (image 38):

25.95, 89.12 (image 38) 10s exposure, 1ms pulse (image 39):

25.95, 89.11 (trying to get four beams - image 40):

PSFs quickly:

We have four beams. Going to take data.

Images 41-48 are the first DIMM. 5s exposures, 1ms pulse each exposure, and 8 pulses/exposures in total. We have some image analysis data!!

Repeating this with images 49-56

exposure time 2s, number of images 20 tried 1ms exposure

./pulses.py -n 20 -d 0.025 -f 0.25

images 57-77.

weekend data analytics!

5.12

Tomorrow priority list:

• Test if beams in same spot with LATISS - EXACTLY THE SAME! WOOHOO!
• Figure out which beam is which
• Telescope focus sweep
• Make OD filter and pulse length decisions
• Take 1000 images ?
• Meghan mount all components and loctite and tighten everything
• Test remote operation
• Talk about night plan: general observing strategy, dome positioning, etc.
  1. note which nights are likely to have better weather...

Blocking beams such that if facing towards the dome and moss, going left to right. (mirror closest to beam expander first, then left to right to far mirror. should be 4,1,2,3). These are images 13,14,15,16.

Moving beam 4 so that the center is on the image. Turned the dial a little bit towards the dome facing the setup. We will see what that does to our image (17).

Beam moved off the image. So I am going to try moving it the other way a tiny bit. Image 18:

Beam still off detector.

We are going back to the other nasmyth port to align on that fourth beam the best we can. Will take time. It is worth this time to get all beams adjustably on the telescope.

There is a relative misalignment between the two nasmyth ports. We have to re-find the beams before we can align them.

Tried 89, 25.75. 

Tried 90

tried 88

Going back to 89, I believe elevation is wrong. Moving from 25.95 to 26.5. Moving to 25.5.

25.75.

25.65

89.1

88.95, 25.6

88.9, 25.6

______________________

starting 7:20 with

88.9, 25.8

other neysmith port: 88.9, 25.8

88.9, 25.6

25.7, 88.9

Collimation

Significantly changed the collimation of the first beam expander

DATA ANALYTICS

We looked at images (307-316) and we needed threshold of 500, box size of 100 to work.

For images 317-416:

Elana code is finding all four beams and nothing else in every image so far, and is able to centroid the beams!! Exciting. we are taking more data. One of the PSFs is garbage, but we will try convolving and see where that gets us.

Questions:

• what in the heck was happening with the focus and PSFs earlier?
  • ATMCS cRIO needed resetting? down
  • maybe primary not supported on pneumatics
• next steps:
  • measure/calculate/compare the distances between beam centers
  • take as much data as humanly possible
  • separate notebooks for quick look and full blown analysis (easy!)
  • making whole system remote
  • do while observing
• power law convolving with data analysis on fourth peak?
• why is PSF of fourth beam bad (beam closest to beam expander)? 
  • meghan hypothesis that maybe has something to do with light being clipped by edge of mirror

Matching sources and positions

Tuesday

1. Dome rotation test

Code improvements:

• better background subtraction?