# 2021 Log <!-- styling tools --> <style> .caption { padding: 0px 50px; color: grey; font-size: 16px; } .temp { color: grey; font-style: italic; } .greyblock { background-color: #f5f5f5; padding: 20px; border: none; border-radius: 6px; margin: 10px; padding-top: 7px; } .figma { margin: auto; width: 100%; height: 700px; } .figmab { margin: auto; width: 100%; height: 450px; } .center { display: block; margin-left: auto; margin-right: auto; width: 50%; } </style> ## 10.19.21 This is some stuff I did today. I'm gonna try making this my home page in Brave. Now I have this log set and my home page. ToDo: 1. Get jitterate code working over all relevant attenuations 2. Export relevant histograms... 3. Work on plots in jupyterlab ## 10.20.21 Continuing to work on jitterate code. 1. See that all attenuations are working. 2. Make sure the fitting bounds are working. 3. Get the multiprocessing to work. 4. DCRmin. I need the format. And I need to compress the text. I'm having a hard time coming up with reasonable error bounds on mu and sigma for the histograms I fit to the jitterate data. All I want are some error bars. They don't have to be terrific quality, they won't be that useful in the final paper. They will just explain some of the jagged nature of some of the plots. Options: 1. some fancy math with the outputs of scipy.optimize.least_squares() 2. some fancy math with the inputs and outputs of scipy.optimize.curve_fit() 3. use a bootstrap method. relevant links: https://stackoverflow.com/questions/35690937/errors-on-a-gaussian-histogram-curve-fit-using-scipy-curve-fit https://stackoverflow.com/questions/14854339/in-scipy-how-and-why-does-curve-fit-calculate-the-covariance-of-the-parameter-es/17508924#17508924 https://stackoverflow.com/questions/49130343/is-there-a-way-to-get-the-error-in-fitting-parameters-from-scipy-stats-norm-fit https://web.physics.utah.edu/~detar/lessons/python/curve_fit/node1.html more advanced fitting library that might be useful: https://lmfit.github.io/lmfit-py/builtin_models.html#example-3-fitting-multiple-peaks-and-using-prefixes curve_fit: https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.curve_fit.html least_squares: https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.least_squares.html#scipy.optimize.least_squares general advice on getting errors after fitting: http://herschel.esac.esa.int/hcss-doc-15.0/load/dag/html/Dag.Sec.SF.Error.html if method 1 or two would work, I think they'd be the simplest. But method 1 gives nonsensical values. ## 10.21.2021 TODO: - [x] Read into the QKD papers again. Are they worth my time? - [x] Message Boris, ask for DCRmin template - [ ] Try revising DCRmin. Pull out words wherever possible. ## 10.22.2021 Lots floating around my head today. I finished the lathe safety quiz. The mill safety quiz is due two days from now, on what I assume is Sunday. So I should do it today. - [x] **mill quiz** I have a meeting with maria this afternoon. Discuss DCRmins status. - [ ] **DCRmin revisions.** During the INQNET meeting today I'm reminded how much they need to understand clock locking and and recovery. I need to make a presentation that outlines what I believe is the best course of action for them. They should: - send an ultra-low power clock that is recieved by SNSPDs. Rep rate rather low. Like 1 MHz. - As SNSPD tags come in, figure out what virtual faster clock they should line up with. Or, if they really want to anyclock (which they only should if I locks to both the awg and qutag) then they can do that. In that case, take use the modulo operator of any incoming timestamps to convert them to histogram space. That's what they should do. Always rely on a phase locked clock. If they are working with a mode locked laser, then they make have to operate at a weird rep rate, and that would probably writing a custom PLL in software. Anyway, I can't focus on that today. Today, focus on the DCRmin revisions, and mill quiz. ### 5:30 Meeting: > [time=Fri, Oct 22, 2021 4:47 PM] #### DCRmin 1. DCRmin received good reviews. Currently revising it to conform with the suggestions. <div class=greyblock > Reviewer 1: The authors report a new scheme for free space optical coupling to superconducting nanowire single-photon detectors. Very impressive performance is reported with near unity efficiency, low timing jitter and very low dark counts. The latter is achieved by carefully selecting a single input optical mode and implementing innovative filtering inside the cryostat. <b> The references are mostly appropriate; I have found one earlier paper on free space coupling in SNSPDs which should be cited Bellei Optics Express 24 3248 (2016) - this current work addresses many issues encountered by the authors of the 2016 paper. </b> This work is of strong interest to the optics community. I am happy to recommend the Memorandum for publication in Optica. </div> <div class=greyblock > Reviewer 2 : The proposed Memorandum by Mueller and coworkers reports on the realization of a free-space coupled SNSPD. While free space coupling usually comes with noise or compromised efficiency, the present implementation enables to combine all metrics simultaneously: large efficiency, low noise, and low jitter. This is obtained via a subtle engineering of the spectral filtering, cold environment and optical coupling. The achieved result is an important step for a variety of applications and will be of keen interest for the broad audience of Optica. I do have two minor comments.<b> i) is the use of differential SNSPD important in this demonstration? ii) the efficiency reaches about 70%, still far from highly-efficient SNSPD realized recently. Is it a compromise to achieve the presented result or is it only limited by the SNSPD in use in this specific demonstration?</b> </div> ___ #### Presentation about clock recovery - I would approach clock distribution a fundamentally different way than fqnet is currently attempting - Lautaro is building clock rate reduction boards. IMHO they're unnecessary - I would not rely on anyclock boards much, if at all - Use software Phase Locked Loop (PLL) if the clock is to be derived from SNSPD tags. - [notebook demonstrating PLL](https://colab.research.google.com/drive/1sKrHggHZX1TvFr9nIfykhSPJlRUwAgcp?authuser=1) <img src="https://i.imgur.com/ovVrIwC.png" alt="test" width="500" class="center"> ___ #### Jitterate (and PPM) - Have all the data for both projects. - Currently making plots and writing. - Focusing on Jitterate for now to push it over the line. ___ #### March Meeting - Abstract about high rate entanglement distribution. - My membership is renewed. My previous abstract: :::info With high efficiency $(>70\%)$, sub-Hz dark count rate, and timing jitter below 15ps, the latest differential-readout Superconducting Nanowire Single Photon Detectors are best used with novel light sources to demonstrate their true potential. We report on the compelling abilities and applications of a high rate modulated mode locked laser testbed paired with these high performance SNSPDs. By synchronizing a variable repetition rate telecom-band mode locked laser with high speed lithium niobate modulators, we generate ultrashort laser pulses with extinction ratio beyond 60 dB. Detecting these pulses over a high loss channel with low timing jitter opens the door to previously undemonstrated Pulse Position Modulation (PPM) on a 20 GHz clock, a promising advance on the path towards high rate deep space optical communication.We demonstrate PPM encoding characterized by channel capacity metrics that approach theoretical limits. This testbed and detector infrastructure also forms the basis for ongoing high rate quantum commutation research. By pumping a pair generation crystal with the modulated laser we aim to demonstrate entanglement distribution with repetition rates beyond 5 GHz. With this, we set the foundation for practical high rate entanglement-based QKD and teleportation. ::: :::warning Future quantum networks will require robust and high rate sources of photonic entanglement. In order for distant network nodes to properly measure and transmit quantum states at high rate, they must employ low jitter single photon detectors and high brightness sources of entangled photons. We demonstrate high rate distribution of time-bin entangled photon pairs in an architecture that allows for high scalability in a multi-node quantum network. With 14 ps FWHM jitter SNSPDs, we are able to measure the correlated states of on a 4.xx GHz clock and using an 80ps time bin seperation. With this high expirement rep rate, we are able to With low jitter (14ps FWHM) Superconducting Nanowire Single Photon Detectors (SNSPDS), time-bin entaglement may be observed at a high rep rate and with ultra-short time delay interferometers (80ps) ::: ## 10.25.2021 Todo today: - [ ] revise DCRmin, prepare for 9:30 meeting - [x] write abstract - [ ] jitterate coding - [ ] banner image for DCRmin... ### Meeting 1. Jittrate coding for plots 2. Clock distribution 3. For jitterate: 1. don't use error bars derived from fits at all 2. only analyze data with certain number of counts per histogram 3. consider using a raw fwhm instead of a fitted one. So I should be writing my abstract for march meeting about PPM rather than entanglement distribution. :::warning **20GHz pulse position modulation with a low jitter Photon Number Resolving Superconducting Nanowire Single Photon Detector** The emergence of high performance Superconducting Nanowire Single Photon Detectors (SNSPDs) will be pivotal to the development of future quantum and photon-starved classical communication systems. New detector types achieve low jitter, high efficiency, and photon number resolution all in a single device. We demonstrate the use of a differential readout Niobium Nitride SNSPD in a 20GHz pulse position modulation (PPM) communication protocol. The detector uses differential readout with a method to cancel timing jitter caused by the random photon absorption location on the nanowire, as well as impedance matching tapers which efficiently couple energy out of the nanowire. The tapers also induce photon number dependent effects on the rising edge of the readout pulse. These effects give the detector photon number resolution, which is of particular useful in certain quantum optics and quantum communication experiments. However, when recording the pulses with a constant voltage Time to Digital Converter, photon number effects on the pulse rising edge induce correlated offsets on the perceived arrival time of few-photon optical pulses. We study these effects in detail and device a method for cancelling them out, by splitting the RF pulse and triggering on it at two voltage levels to measure photon number character. These corrections are necessary to achieve low jitter in the PPM demonstration, and any other application that requires low jitter time tagging of optical pulses with varying photon number. Overall this investigation shows an integrated approach to photon-starved communication that fully exploits the lower jitter of a new SNSPD type while managing its complex manifestation of photon number resolution. ::: Emerging types of Superconducting Nanowire Single Photon Detectors (SNSDPs) have simultaneous jitter, high efficiency, and photon number resolution. ## 10.26.2021 - Finished revisions to DCRmin for now. Got an email last night at midnight saying revisions are due in 48 hours. So I'd assume the deadline is midnight on Wednesday. I'd like to get it submitted a day early (today) if possible. ToDo: - [x] DCRmin image - If I need my blender files for this, maybe I should do it at home this afternoon. If only I had those files on my new home server... gotta finish setting up nextlcoud. And buy hard drives - [x] revise abstract - [ ] Jitterate coding. Get the exponential modified gaussian working. Can I just do a really fast bootstrap method? - all I need is np.random.choice(a, size=len(a)). Run the fit a few times. Take the variance of the mus and the sigmas, and that's the error values. - [ ] GD and T quiz is coming up. ## 10.27.2021 - [x] March meeting stuff - [x] Get DCRmin submitted! I have the image, just waiting on Boris - [x] GD&T quiz - [ ] When the DCRmin paper is accepted, then send it to Emma again. Maybe the inqnet system for clock recovery should be designed in such a way that the discrete clock could be removed later. That means try to extract a clock from the data, and maybe compare it with the discrete clock. ## 10.28.2021 - [ ] Jitterate writing and coding. - [ ] exponentially modified guassians in the main code - [x] FWHM measure - [x] simple bootstrap? - [x] GD and T quiz Remember, I'm only 2 days away from monday meeting! ## 10.29.2021 Yesterday I started the mill part for Me113, and did the GDandT quiz. I was thinking about the best way to do bias the Jitterate calibrations. Depending on your goal, you may want the maximum number of tags to fall in some predefined bin, or you may want the minimum standard deviation of corrected counts. So it's reasonable to set the centerpoint for a distribution offsett from the peak it you're trying to get both the peak and some exponential falloff inside some time window. How do I define the "max counts contained" measure? Feeling pressure to show more results in the Friday meeting. Maria would like to see this as a 2021 paper. That's a goal to shoot for. End of day: 1. Morning: Jitterate 2. Afternoon: Jitterate ## 11.01.2021 ### For Monday Meeting: <div class=greyblock> <h4> Outline </h4> <ol> <li> DCRmin accepted </li> <li> March meeting abstract submitted </li> <li> Changes to Jitterate </li> </ol> </div> #### 1. DCRmin accepted Thanks for the help everyone! Image I submitted for the thumbnail: <img class=center src="https://i.imgur.com/saX7KYk.jpg"> <style> .block { display: block; } .center { display: block; margin-left: auto; margin-right: auto; width: 60%; border: 5px solid #e6e6e6; border-radius: 10px; } </style> #### 2. March meeting abstract: :::warning The emergence of high performance Superconducting Nanowire Single Photon Detectors (SNSPDs) is pivotal to the development of future quantum and photon-starved classical communication systems. These detectors achieve low jitter, high efficiency, and photon number resolution (PNR) all in a single device . We demonstrate the use of a differential readout Niobium Nitride SNSPD in a 20GHz pulse position modulation (PPM) communication protocol. The detector uses differential readout as well as impedance matching tapers for efficient coupling of energy out of the nanowire. Furthermore, the tapers induce photon number dependent distortions on the rising edge of the readout pulse. Therefore timetagging these pulses at a constant voltage convolves photon arrival time and photon number information. We demonstrate a method for measuring RF pulse slope that may be used to (1) measure photon number per optical pulse and (2) extract a low jitter photon arrival time measurement that is uncorrelated with photon number. Our photon-starved communication demonstration therefore fully exploits the lower jitter of a new SNSPD type to run at high rate, while managing its complex manifestation of photon number resolution. ::: #### 3. Jitterate ##### Last time: 1. Don't worry about error bars 2. Consider not fitting to a distribution, just use density estimation and FWHM 3. Consider fitting to an exponentially modified guassian I made an exponential fitting function, but it's center and falloff parameters seem [unhelpful for my needs. ](http://localhost:8888/lab/tree/fitting_with_errors_before.ipynb) Now using **median** for the distribution center point. This seems best for guaging the center of mass of a somewhat asymmetrical distribution, while also ignoring outliers  FWHM regions are in green. Blue lines are medians. #### ToDo: 1. Send emma final version of paper. Doesn't have to be the optica typeset version. 2. Forward invoice to Emma and Boris. **Avoid planning to write papers after you graduate.** ### Today: 1. Code for Jitterate 2. Journal Club (This thursday!) 3. Thoughts: Sam asked for a plot of detector efficiency vs wavelength. We don't currently have a nice way of making that, but we could with the tunable CW laser and maybe some changes to the attenuator script. Would need to add network control to the laser. Laser user guide. SCPI-like commands seems to be the way to go. (page 30) https://www.newport.com/medias/sys_master/images/images/hc2/h2c/8797022519326/6600-User-Manual-v4.pdf I can modify the SCPI python code I have for the AWG. #### Some comments for the Curvenote team 1. A small UX comment. 2. Different users should have something like color coded cursors like in google docs. So you can see what block or section one person is editing, so you know to avoid that section. ## 11.02.2021 ToDo: 1. Reading for journal club 2. Jitterate - How am I going to do the bootstrap FWHM? It takes too long... 3. Sorption cooling By 10am: An outline of my talk. What can I talk about and what can I not? Sorption cooler: Steady state: Tpump: 4 K T4K: 2.8 K T1K: 0.8 K T40K: 30 K Tswitch: 14 K why does the 1k stage go above 2.8 k? It goes all the way up to about 4k Think about a cold glass of water. The water that condenses on it surely raises the temperature of the water. The condensation process itself is a heat load. When that heat load goes away, then the pot [the glass of water] is at it's native temperature. The temp of whatever it is connected to. Before: T1k is at 800mk, pot is evaporating into the pump(charcole). T4k is at 2.8k (as it always is) 1. Heater switch turns off (Hs turns off, blue). This drops temp of little absorber connected to the gas-gap switch v 2. Gas inside the switch condenses into the absorber, leaving the region bewteen 3. The switch becomes a POOR conductor 4. Once that is done, you can heat up the pump to evaporate off the Helium. Heat up to 40 K with a heater (Hp (red) turns on for a short time). 5. Evaporation from the pot suddenly stops as the vapor pressure above increases. The pot is no longer being cooled. So conduction of heat from the 4k stage will warm it. 6. As vapore pressure increases, the temperature at which the helium will condense increases. Eventually the temperature of the pot is cold enough, and Helium condenses into the pot. Or it condenses into the condenser which is the tube between, and trickles down due to gravity 7. The condensation is a source of heat. Think about the condensation on the outside of a cold glass of water. It's probably raising the temp of the glass right? 8. The temperature of the 1k pot rises even above that of the 4k stage (2.8k) because of the condensation. 9. As time goes on, there is less and less gaseous helium that has not been condensed. The rate of heat added to the pot goes down. The pot approaches the temperature of the 4k stage (2.8 k). When it gets sufficiency close to this temperature, you can be sure most of the helium has evaporated off the charcole and condensed into the pot. Because if the rate of condensation was still high, there would still be heat load on the pot. (this is why T1k must reach about 2.8 k for the program to consider the condensation process finished. If there's an amplifier on the 4k stage that makes it slightly warmer, then it will never reach the set point. It will always think that the heat load is from condensation). 10. With T1k at 2.8 K, the program signals the switch to close; to become a GOOD conductor. It does this by warming the absorber in the switch with Hs. 11. With the switch a good conductor, heat flows out of the hot pump and into the 4k stage. You can see a momentary increase in 4k stage temp because of this. 12. Evaporation begins. Hs stays on indefinitely to keep the switch filled with gas and a good conductor. Who's triggered? https://www.youtube.com/watch?v=3Fx5Q8xGU8k&t=1744s Working on journal club today Finished machining the Mill part for Me113 ## 11.03.2021 1. More work on journal club today 2. Did some small stuff for the machine shop class 3. I'm really trying to learn more of the guassian state formalism. What Sam and Nikolai have been using for modeling HOM and teleportation ## 11.04.2021 1. More research about gaussian formalism and P-representation. Reviewing my old Quantum Electronics notes. Presented journal club at noon today. Explained how a sorption fridge works. Sam had a PNR meeting after. Went to the Physics colloquium at 4pm. Presenter was one of the founders of PsiQuantum. ## 11.05.2021 Today I have an 11:30 meeting with Rowan Cocket at Curvenote. Thinks to talk about. - What are your immediate goals? - What do you think are the first reasons people will try out curvenote? - I'm expecting the notebook to be the ground truth. But it seems like the curvenote plugin is modifying and choosing versions for my notebook. - Maybe say something like "are you sure you want version controll handled by curvenote?" If the user say's 'No' then that assumes they're doing their own version control with git, or they're just not advanced enough to care about such things. - Would it be advantageous to strip certain figures made by jupyter from their html block so that you can provide more control to the user, as if the figure was imported just like a .jpg (sizing, scale) 2itc internal computer consu myst - markdown format that adds metadata. very difficult to nikey case,s neuron interactive video authoria. ## 11.7.2021 ### Meeting: 1. Jitterate - Some info on using curvenote. - Sorry about inconveniencing people with a new platform. It's not so amazing yet, some parts may be frustrating. I'm interested in it because of the vision of its creators. - They're trying to better define the open format of the future of publishing. That is, what comes after PDF. A what advantages it will have. Rapid reuse, reproducibility, better managment of code, etc. 3. I'm planning on presenting to the inqnet team this friday about clock distribution and clock recovery. I might try a proof of concept of clock distribution with SNSPDs. 4. Journal club too a lot of time. I spent a lot of time learning about gaussian formulation, related to Sam's modeling. 5. **!! I need to write 3 sentences about DCRmin.** 6. ToDo: 1. Remove the afterpulsing in my illustrations 2. Try to include the corrected and uncorrected histograms for the final data figure(s). 3. Potentially combine the array data and the single pixel data onto one data-figure. ## 11.8.2021 Have a decent looking data plot for the single pixel. Though the group wants to see the histograms with and without correction. So i might have to remake it into two plots, and/or combine it with the array data. ## 11.09.2021 I made a demonstration of clock locking with an SNSPD. I used my jitterate datasets, where the rep rate of the (modulated) laser is 537 MHz. It's a weird sort of signal to lock the clock to. It actually shares more in common with quantum data than a dedicated clock signal. I wonder if I will need something like this anyways. Because the laser rep rate is going to be much higher than the dead time of the detector anyways. So I will be in the same regime with the high clock rate entanglement project. The jupyter notebook is called clockDemo. Histogram of RF channel (what was originally used as the 'clock'):  Around 10ps jitter on this channel is what I've observed before. Which implies the jitter of the locked lock is much better. A few ps. This is for the tungsten sciliside detector I'm using for jitterate. The jitter histogram has that long tail, because of the effect jitterate plans to cancel out.  ## 11.10.2021 Got new dataset plotting. I already re-ran all the scripts so that I could get another statistic related to visualizing the FWHM numbers.  Also one of the last projects for the machining class took some time. How to more accurately dimension drawings. ## 11.11.2021 New figure for the high rate entaglement project, loosely adapted from Sam's figure in the Sync paper.  ## 11.12.2021 Showed some slides the the friday inqnet meeting. General consensus is that it's cool, but not very scalable because the high quality clock only exists in software. Would be better to have a device that timetaggs SNSPD pulses and outputs a 10MHz sync. New plot showing how you can use an SNSPD to measure itself: ## 11.15.2021 Last week: 1. New jitterate data plotting - How should these figures be layed out, what figures should we keep/discard 3. Clock Locking, high rate entaglement - new drawing (see above). - Conclusions from inqnet meeting 4. Other small things - solidworks tutorials for the machining class ### ToDo: full width at tenth max: FW1/10M full width at one hundreth max: FW1/100M Look into putting in the high rate corrected/uncorrected single pixel dataset. ### Entanglement distribution ToDos 1. programming for swabian - clock recovery with clock divider board - support socket input/output for talking to the voltage source 3. getting everything running with a beamsplitter 4. program the voltage source Spent a little time today 'working on the web-swabian upgrade' ## 11.16.2021 Back in Pasadena - [x] Look into adding 1/100 max for the single pixel. - Done. Added new lies to the plots I'm building. - Spent a while today on the final quiz for the machining class. Super annoying quiz. There are questions that require you to sift though several pages of tables about hole and shaft interfaces, tolerances, machining operations, and so on. - Eventually I decided some questions I was going to just guess. ## 11.17.2021 - Finally finished the final quiz for Me 113 - Still adding and updating the jitterate plot today.  Had the nist meeting today. Was very interesting to hear about what goes into an aprac proposal, and how difficult they can be to get. Also, how we and other scientists and nist try to figure out what nasa is interested in by submitting proposals and leaning from the comments. Call it a morbid curiosity, because I would hate writing a full proposal just to get a few comments about it. ## 11.18.2021 - new layouts for the fitterate figure. also spend some serious time this morning learning about TextFrame... Working on making the plot with the 3.5GHz dataset. Looks like I'll need to make another version of ClockTools that makes less assumptions about the rep rate of the laser. Had journal club today. Presentation about Low kinetic inductance superconducting MgB2 nanowires. ## 11.19.2021 Spent some serious time today getting a time tagger script working for Sahil. Using the gating virtual channel. It ended up working well enough, though he has to learn a lot more about the software if he's going to use it himself. Worked on getting the lase locked at 4 GHz. Next comes the SHG, and SPDC. ## 11.22.2021 ### For Meeting 1. Jitterate - Mostly still doing work on the data figures. - Show layout options - Added the Full Width at 1/100 max. Data is clean enough it appears valid over the whole range of count rates. - Added the high rate plot, show different options - Added linear scaling inset axese - Show layout plan for the array dataset 3. DCRmin - No word back yet from the managing editor. 4. Entglement distribtion - Did a little prep while helping Sahil. Will do more today and tomorrow. ToDo 2. look at DCRmin comments from Boris, the accept proofs ## 11.30.2021 1. Jitterate - Assemble the array figures - Writing! FOCUS. 2. DCRmin?? Is there anything to do? - I think it's in production. 3. EntangleRate - Laser should be locked at ~4 GHz. 3.13 pm: I will try to pull out some data for what happens to the jitter profiles when you discard data instead of correct for it. -> Got most of the code done. I'm including count rate metrics for when you discard 2ndary tags less than 50ns later, less than 100ns later, and less than 150ns later. -> I'll make some graphs that show how fast you lose cout rate using this version. You may start losing count rate very quickly with dB... Because you'll enter a regime where potentially distracting thought: I want a web based way of demonstrating the line shown in the figure, with draggable callbacks. ## 12.01.2021 Will give a few slides at the nist meeting about jitterate. Focusing on that for now. ## 12.02.2021 Some things about curvenote. 1. I won't keep using it much longer for this paper. - They still have a lot of work to do. Plently of editing and collaboration features that overleaf has that they don't have yet. 2. For now, I'm am explicitely refernces figures without links. ! Remember! I'm planning to have a picture of of the detector reset process in the figma document for figure 2. Either the scope image or a decoded .npy file I think I took. Look out for this. **--> done** ## 12.03.2021 I should read the PNR G2 Walmsley paper, try to figure out how they determined those weirdly good values of P(1|2) and P(2|1). **Work on the PNR paper SNSPD paragraphs. Looks like Boris's part is already decently written** ## 12.06.2021 1. ~~Presentation for Keshav~~ ~~2. PCR curves for boris, subtract 3.3%~~ 3. ~~PNR scripts for Sam, remove the Timetagger dependency~~ how do you seperate two projects conceptually in a folder structure if they use the same large dataset? I want each folder tree for each project to be portable. I don't want to have two copies of the large dataset in two locations. ## 12.08.2021 I've probably spent too much time doing a thoughouh commissioning of the differential fiber coupled detectors. I did sets of PCR curvers at 5 wavelengths, and on DIFF-3 I tried to do a finer grained scan of wavlength at 0.36V bias current, with measurements at min and max polarization-determined efficiency. On monday I tried to communicate with the tunable laser over ethernet and with SCPI-like commands. There's very little info on the internet about how to work with this laser. Couldn't find any evidence of support for VISA, or the VISA NI MAX tool. I tried a scpi-over-ethernet script I have from the keysight AWG, but it didn't work. For one, I don't know what port to use, and the AWG script requires a port. Anyway, I spent too long trying to get it working. Then on Tuesday and Today (wednesday) I've been playing around with the PCR curves. Also got the PLL data organized for Keshave yesterday. I NEED to spend the rest of the day writing. Got the SNSPD meeting at 1:30 though. I was also trying to watch the FORCE11 conference on tuesday. **I need to learn how to use dataframes** ToDo: 1. Learn about dataframes 2. learn about orchid id. 3. keey doing pomodoros! ## 12.09.2021 I'd like to get some entaglement distribution progress done. With sam in a rush, best to focus on time tagging infrastructure ### Initial Plant for Swabian Entaglement Software 0. show the PLL can work we 1. ll when run as a fu 2. 3. nction multiple times. Do this on my computer. 4. A matplotlib plot that updates in real time and shows the jitter of the laser tags with and without a phase locked clock. - I can use the pause feature for this. I should demo this functionality for myself again. Ch. 1: clock Ch. 2: snspd tags. ## 12.13.2021 ### For meeting, from last week 1. Comissioned the two fiber coupled NbN detectors. 2. Writing Jitterate, - look at overleaf 3. Working on swabian real time codebase ### ToDo: 1. New authors for the jitterate paper - Rogalin, Ryan 2. Talk about next week about 3. Work on Sam's paper! I need to revise quite a bit. Explain the variation in pulse height without a picture... ## 12.15.2021 Thinking about clock synchronization clockkit: https://github.com/camilleg/clockkit https://stackoverflow.com/questions/36327805/how-to-install-gcc-5-3-with-yum-on-centos-7-2 Something in my repeated phased locked loop code (object oriented) is still not working the same as the default version. Need to debug.