MINUTES EURIZON Task 4.1 Topic 2 meeting: measurements October 2022
Pariticipants:
Simone Liuzzo
Lukas Malina
Bianca Veglia
Elaf Musa
Joachim Keil
Ilya Agapov
Thorsten Hellert
Nicola Carmignani
Lee Carver
S. White
Eurizon not yet approved (no official communication about it).
Comment on first version for our part: "Business as usual"
Revised document was submitted. No direct input was requested from our side.
DISCUSSION:
TbT measurement requirements
Beam with feedbacks off
N.C. suggest to use mag shaker. smooth sine wave excitation.
signal will be averaged. so there will be non non uniform shift bunch by bunch.
N.C. Comments that B.Roche and K.Scheidt mentioned that the shaker are easier to set up.
sinusion not a step function. each bunch will see a different phase. using only 20-30 bunches this will not be a big problem.
L.M. beam is stable with feedbacks on. N.C. there is some effect of current, but we will stay at low current per bunch. K.S. said that 10mA is the best current for BPM signal (above BPI) if we keep small exitation, this will not be a problem. Else, go to 5mA, 30 bunches. still low current per bunch, no significan change in optics.
L.M. BPMs gain for corect current. SL., NC. = YES
K.S. said that if we do measurement with stored beam BPM precision is much better.
L.M. 2 different electronics. How to trigger at the same time? How to synch BPMS? Worth checking ADC data read out.
N.C. There are procedures and functions in matlab to get the BPMs to the correct TbT setting. tryed only once a few MDTs ago (Sep2022). We can chekc synch. we will have signal of 320 BPMs in the correct order.
L.M. any machine safety issue? N.C. we can loose the beam. The shaker strength is not sufficent to drive the beam to the chamber.
L.M. rest is on the software side.
N.C. A meeting just with interested people for this MDT. deeper discussion on software.
S.W. will we do simulations? N.C. YES.
Meeting Wednesday 5th October
N.C. simulations before measurement. May be a bit short time.
S.W. also during restart. Restart will be shorter due to machine reqiurements. May be one day less.
N.C. can we do a similar MDT in PETRA?
L.M. not known yet. J.K. not sure as well. Must ask to operation people.
Extreemum seeker requirements
B.V. looked at the code in OCELOT. Looking at the theory in one dimension. Algorithm works. Exploring parameters. Looking for the data format that is needed and if it is possible to adapt it.
Apparently there are some limitations, as choice of bounds. May not be straight forward to obtain the optimal results. Use directly OCELOT to interface with the machine.
Next Thursday at 14.00
S.L. send to B.V. vectors of skew, sext, oct. B.V. will send code to try in simulator.
L.C. why not asking for access to ebs-simulator ? S.L. will ask for cybervisitor access.
if shifts are given, people interested to join:
Bianca, Lukas.
outcome of machine learning (CANCELLED)
L.C. basic toolkits introduction . Idea of workflow, how to approach it. Nicola Lina and Lee add conversation about : what is a good machine learning problem to try? Not clear. What would be a good approach with our machines? What other do: use ML model to speed up optimization (4h-> 30 min).
I.A. Workshop in Brookheaven. Will send invitation. Interesting programme.
https://www.bnl.gov/mlaworkshop2022/
L.C. will continue investigation and use our contacts (as B.Nash). N.C. confirms
L.C. one problem identified: I.E. linac-booster. Required a lot, but still to figure out out to use ML. there is work to do.
practical details for MDTs request during restart Thursday 20th to Monday 24th october
define list of MDTs for which to request time during October EBS restart (for example):
8h shift on lifetime optimization MDT using “Extremum Seeker”
and/or 8h shift on TbT optics measurements and correction preliminary studies
Would you accept to work during nigth shifts (23-7)? YES.
Would you accept to work Monday USM-Like nigth shifts (23-3)? YES.
define list of people participating in person to get request entrance (remote participation will be possible) BIANCA, LUKAS.