Login

K-modulation application for PSB and PS

→ Europe/Zurich
Zoom

Zoom
Hide

Present: Fanouria Antoniou, Hannes Bartosik, Andrea Calia, Jean-Francois Comblin, Gian Piero di Giovanni, Michi Hostettler, Delphine Jacquet, Bettina Mikulec, Tirsi Prebibaj, Giovanni Rumolo, Frank Tecker

The meeting started by briefly presenting the typical PSB and PS use cases for k-modulation and highlighting potential synergies for a dedicated application.

The subsequent discussion focussed on better understanding the requirements and brainstorming for a potential solution. It was agreed to organise a code sprint as next step in the process to look into the work already done by Tirsi and estimate the time required for development. Furthermore, specifications for the application will be drafted and the direct involvement of the OP software team should be discussed. Notes of the discussion can be found below.

  • Michi proposes to follow an approach similar to the Multiturn application, having a common interface with the possibility to incorporate machine specificities into the different building blocks of the application (excitation, acquisition, analysis)
    • the analysis part could become common, whereas excitation requires knowledge about machine-specific hardware
  • Frank mentioned concerns regarding the machine setup and how to incorporate that into the application
    • measurements in the PS required very good coupling correction
    • even for measurements on the flat bottom statistics of several cycles was required and outliers had to be filtered manually
    • the start and end of the measurement is cycle specific
    • Hannes asked whether Frank could provide some of the data from 2018 to identify the circumstances that required manual outlier cleaning in order to foresee that for the application and Frank agreed to help on that aspect
    • it was also suggested to keep machine settings (such as coupling correction) and tune measurement settings independent from the application 
  • Delphine asked for clarification whether the measurements should be performed cycle-by-cycle or within a cycle
    • Tirsi explained that there are two different use cases, requiring both types of measurements to be possible
    • Measurements within a cycle require the quadrupoles to be excited with an arbitrary function (most of the time sinusoidal), whereas constant offsets (varying from cycle to cycle) have to be added to the k-settings when measuring during the collapse of the injection bumps
  • Another point of the discussion was related to the analysis approached whether PSB and PS could profit from a very similar approach
    • analysis should indeed be very similar, especially if acquisition is based on data from the BBQ FESA class 
  • Bettina stressed the importance of having either the application or a script-based approach ready for the beam commissioning (i.e. early December) and that this approach needs to be tested with simulated data.
  • Jean -Francois asked about the missing steps to make Tirsi's application operational
    • Hannes and Tirsi explained that the application only controls the simulations for the moment, the interface to the machine hasn't been developed yet
  • Hannes explained that the new PSB tune control was set up in such a way in LSA to have an independent control of QDE3 and QDE14
  • The point was raised whether this application could profit from any other available application
    • the LHC application seems to be very special and difficult to profit from
    • there's no application in the SPS
  • The question was asked to Andrea, Delphine and Michi whether PyQt could be used instead of Java
    • Indeed, both approaches (PyQt and Java) are now CO supported, but Andrea explained that there's more experience available on the Java side
    • Nevertheless, it could be a good use case to develop this application in PyQt to build up expertise and a common framework
  • Jean-Francois stressed that analysis code conversion from Python to Java should be avoided
  • Andrea proposed to have a modular layout so that different blocks can be developed in parallel without blocking each other
There are minutes attached to this event. Show them.
Powered by Indico v3.3.9-pre