Minutes of meeting held on 17 Dec 2020
==========================

slides are at:

https://beamdocs.fnal.gov/AD-public/DocDB/ShowDocument?docid=8907

(a) Tan gave an update of the work done so far:
    (1)  Improving neutralization in LEBT: work to continue
    (2) Penning source: Installed in test stand under rough vacuum.
    (3) Laser collimation: installation postponed until covid19 tier3
    is over.
    (4) Realignment of PreAcc: results unclear because of source problems.
    (5) Klystron testing: 4 spares tested good, 2 bad, 1 on test
    stand, 1 more to test.
    (6) Flat injection: Work to continue
    (7) Lattice studies: see today's talk by Jeff.
    (8) Adiabatic capture: work to continue
    (9) 2nd harmonic: Cracked shells repaired. Modifications pending.
    (10) Wide bore cavities: Modification to the tuner needs to be
    done. Sims completed.
    (11) Garnet loss improvements: Received type A garnets. Winding of
    solenoids stuck at TD.
    (12) Mode 2 dampers and coupled bunch mode dampers: Mode 2
    measurements continuing. New damper board measurements taken.
    (13) GMPS machine learning: Firmware work continues. Gathering
    data for ML.    
    (14) Paraphase controller: COMPLETE!!!! Operational.
    (15) LLRF Phase and radial feedback: Work continues. LLRF theory
    being worked on.
    (16) Plan is to continue having have 50 kV RF in situ tests in
    2021.
    (17) BTL to L11 changes: PIPII will install additional collimators
    in the BTL line to remove halo before the foil.
    (18) RF phasing experiments show that there are possible problems
    for anti-phasing between A & B stations. Investigation continues.
    (19) Transition crossing: Chandra has formed a group to look into
    this. Will collect data for simulation for high intensity when
    Booster is completely tuned up.
    (20) Magnet girder test: will continue in Jan 2021.
    (21) 20 Hz infrastructure test in 16 Sep 2020. Scale of the test
    is small. More tests in 2021.

(b) Dave Hixson gave a talk on the LCW upgrade cost required for PIPII
  (1) Current system is maxed out to Booster from CUB.
  (2) Dave is always trying to get more flow. Especially RF stations
  requires the flow. One improvement was to remove the Booster LCW
  from  the ramnant SY water.
  (3) Flow is  dominated by Booster RF.
  (4) Load wise is hard to say which device is dominant.
  (5) Present heat load is 2400 kW at CUB heat exchangers.
  (6) The present design max is 5000 kW based on 12 deg F delta.
  (7) The cost:
      (i) Some numbers were calculated in 2015 for PIPII.
      (ii) In 2020, Jonathan Hunt at FESS re-evaluated these numbers.
      (iii)  Work at CUB includes upsizing pumps and pipes.
      (iv) $1.6M cost estimate for materials and construction.
      (v) Design is $145k.
  (8) Note that FESS is also doing an infrastructure project and they
  will be upgrading.
       (i) Things may change at CUB: Presently 6 pumps used for
  Booster, perhaps there'll be 3 pumps after upgrade.
  (9) Tan asks AD has to pay or FESS? Dave H. doesn't know. This will
  need to be worked out between AD and FESS.
  (10) Kent asks whether Booster 95 pipe in the tunnel needed to be
  changed? Dave doesn't think so. Most of the problems are with RF
  stations upstairs. This needs to be looked at though.
  (11) The last 3 slides on current system and cost are the most important.

(c) Jeff talked about the development of SVD and PCA for measuring
beta functions with TBT.
  (1) PCA is also called SVD.
     (i) Basic idea is to find true principal signals from several
     detectors. SVD is used to find the principal signals.
 (2) For lattice measurement from TBT,  PCA is to find cos sine like
  in horz and vertical signals from the BPMs.
 (3) In SVD, the low order values of the singular values are cut so
 that only the princpal values remain.
 (4) This method is model independent except for normalization.
 (5) One important thing to do is to weed out bad BPMs which affect
 the results.
  (i) Check for good/bad bpms by comparing phase advances with MADX
  model. Any discontinuties/jumps will indicate a potentially bad BPM
  and will require more investigation.
(6) Jeff was worried about coupling affecting the measurement.
  (i) Tan says to decouple and then doing the measurement because coupling
  makes it difficult to disentangle.
(7) Finally, Jeff discussed aomaly detection w.r.t. a "golden" lattice
using an ML autoencoder.
   (i) Making such a detector may allow daily ops to recognize
   problems with the Booster lattice and its cause.