Skip to content

Latest commit

 

History

History
320 lines (306 loc) · 21.3 KB

NOTES.org

File metadata and controls

320 lines (306 loc) · 21.3 KB
Raw

XXX switch to BCW binning?

Action items

  • [ ] Label all equations

Outline

Neutrino oscillation

Daya Bay experiment

Calibration

Time

Gain

Energy

Reconstruction

Energy

Vertex

IBD selection

General process; not specific cuts yet

Backgrounds

Accidentals

Li9/He8

Fast neutrons

AmC

Alpha-n

Reactor modeling?

Systematics

Covariance matrix (or pull term) construction

Results for nominal cuts

Cut optimization

Final result

Extra packages

  • latexsym
  • See RevTeX docs
  • textcomp
  • manyfoot
  • hyphenat
  • array/tabularx/supertabular/tabu: tabular*
  • ragged2e
  • changepage
  • sidenotes/marginnote
  • enumitem/paralist
  • titlesec/fancyhdr

Other classes

  • tufte

Answers

References

BibLaTeX

Emacs stuff

{AUCTeX} Subfiles package and reftex

NB: Looks like it’s not necessary to explicitly set TeX-master. AUCTeX and RefTeX are clever enough to figure things out.

Bibliography setup

;; (setq reftex-default-bibliography '("/home/mkramer/physics/thesis/biblio.bib"))
;; (setq org-ref-default-bibliography reftex-default-bibliography)
(dolist (v '(org-ref-default-bibliography bibtex-completion-bibliography))
  (set v (cdr (assq 'bib (symbol-value reftex-docstruct-symbol)))))
(setq bibtex-completion-library-path '("/home/mkramer/physics/thesis/refs"))
(setq org-ref-pdf-directory (car bibtex-completion-library-path))
;; (setq helm-bibtex-full-frame nil)
(add-hook 'pdf-view-mode-hook #'pdf-view-midnight-minor-mode)

cite:An_2017 ,An_2017 label:sdf

label:foo

ref:sdf

cite:An_2017

Analysis notes

Background histogram

Total height of stack has natural interpretation. Within each bin, the total height is divided up between components, such that equal contributors will have equal heights after unstacking and placing side-by-side. For unequal contributors, the heights (after unstacking) will reflect their logarithmic difference, i.e., each major axis increment corresponds to a factor of ten.

Kam-Biu’s idea

  • Take far site data
  • Use inverse square law to predict near site rate for θ13 = 0
  • Draw random subsample of near site data, matching predicted rate above
  • Ensure fit recovers θ13 = 0 and is stable against variations in cuts, repeat for other values of θ13

Notes on Li9

Regarding [Li9_2006]

  • The time constant T’ = 1/(R_mu + R_nu) applies when we’re considering only IBD candidates that are preceded “immediately” by a muon, i.e. requiring no intervening IBD candidate (so we’re throwing away cases where the IBD was preceded more recently by another IBD than by a muon). See reMarkable. If we ignore intervening IBDs and just take the time-to-last-muon for all IBDs, we just get T = 1/R_mu, regardless of the IBD rate.

Sources of images

References

Example theses

Big presentation on reactor neutrinos etc.

Jan 29, 2019

DocDB references

Backgrounds

Flashers

  • doc-7434 (Xin) - Summary (including bibliography)
  • doc-7537 (Xin) - Inefficiency
  • doc-7946 (Xin) - Inefficiency update

Fast neutrons

AdSimple

  • doc-7334 (center of charge; energy)
  • doc-7536 (templates)
  • email to Sam on 2020/8/4

Fitter

  • doc-7709 (rate fitter)
  • doc-8769 (toy MC)
  • doc-8774 (shape fitter)
  • doc-9999 (P14A update)

2015 unified nonlinearity model

  • docs 10330 10254 10260? 10268 10269 10274; history of unification 10061?
  • consModel_450itr.root is Soeren’s energyModel_march2015.root from doc-10330

2017 nonlinearity model update (what we actually use)

  • doc 11486, 11611, 11646 (Yongbo)
  • We use the root files from doc-11611
  • How do those differ from the ones in 11646?

Error budget

  • doc-11724 (P17B, Beda)
  • doc-10804 (P15A, Henoch)
  • docs 8774, 9999 (shapefit technotes)

Statistics

From Henoch

Misc

Supermisc

(Parallel?) fitting

SLURM

Neutrino history

Neutrino oscillation

Other experiments

June 1, 2020

Non-DocDB

DocDB

Daya Bay DAQ

Missing from DybPublications

Neutral kaon mixing

Software

Uncertainty propagation

Julia

Tab-completion in the REPL

Printing the whole thing

Plotting

DataFrames

UpROOT

LibreOffice from Python

Emacs

More bibliography management

External tools

org-ref/helm-bibtex hacks

Finding the bibliography

Instead of the below, could just set org-ref-default-bibliography

(defun my//ad/thesis-find-bib (orig-fun &rest args)
  (if (and (buffer-file-name)
           (s-prefix? "/home/mkramer/physics/thesis" (buffer-file-name)))
      '("/home/mkramer/physics/thesis/biblio.bib")
    (apply orig-fun args)))

(advice-add #'org-ref-find-bibliography :around #'my//ad/thesis-find-bib)

Formatting collaboration name in helm-bibtex

Should properly fix helm-bibtex and submit a PR (see #266, #267)

(defun my//ad/fix-bibtex-completion-author-abbrev (orig-fun &rest args)
  (-let* (((old-field entry default) args)
          (new-field (if (equal old-field "author-abbrev")
                         "author"
                       old-field)))
    (apply orig-fun (list new-field entry default))))

(advice-add #'bibtex-completion-get-value :around #'my//ad/fix-bibtex-completion-author-abbrev)

Multiple bibliographies (for figures, etc.)

DYB misc

Coordinates and baselines

The caption of Table 1 in the long paper is wrong. It says that X is due north. Actually, X is due “west” and Y is due “south” (right-handed coordinates), and the axes are rotated an additional ~8.5deg clockwise from “true” (i.e. Google Maps) north, so that D1 and D2 have the same Y coordinate (i.e., we get these axes by rotating the “standard” axes by 188.5deg clockwise). Also, every published diagram I’ve seen of the site layout is totally not to scale. The only trusted source is Table 1 (mislabeled axes notwithstanding). I’ve (partially) verified that the coordinates in that table give the proper baselines.

Minimum hit separation

An SPE pulse looks to be around 30 ns baseline-to-baseline (doc-6710). Meanwhile, misc_ana/TdcSep indicates that the minimum separation between hits is about 50~ns. Most “late light” hits arrive 75-125 ns after the early hit. Also see 1902.08241 (nonlinearity) and 1707.03699 (FADC).

Hit timing

Light is emitted stochastically with a fast (5 ns), medium (30 ns), and 5% slow (150 ns) component (doc-8732: Why Switch to Peak Charge?). Propagation, including reflection, can add up to 20 ns or so. See docs 8319 (Additional Sources of Non-Linearity), 8732, 8356 (Re-calculating the Gamma Energy Scale using Peak Charge), 6710 (me), 5612 (closely spaced hits), and Zeyuan’s doc-6926 (determination of time window; which hit to take).

Things to investigate

make_combined_spectra.C

  • line 176: why do we use [0] instead of [j]