Page History

Attendees: Ken Lauer Vinod Gopalan Barry Fishler Alex Wallace Peregrine McGehee Zachary L Lentz Mitchell Cabral Nate Landis Robert S. Tang-Kong 

Summary of Last Meeting:

        LLNL Requirements:

• • Setpoint tool shall be able to write to multiple devices store various configurations of these values. 
  • The RRL devices shall have the ability to move through multiple configurations throughout the entire shot cycle, but does not require reconfiguration when shots are being taken. 
  • The tool shall provide the operator to check configurations on-demand.
  • There is no need for an atomic system as the time requirement to reach set configurations are in the hours range. 

        LLNL NIF System:

• • At NIF setpoints can be created on-demand.
    • Ability to continuously monitor the state/value/parameters of the device and setpoint.
    • Can write to multiple devices at once.
  • NIF has an Oracle database that is used to maintain setpoints and configurations. 
    • • Other systems have configuration files that live on local controllers.

        ATEF (Work In Progress):

• • SLAC-developed tool for testing that can be used to verify and check configurations.
    
    • Moving towards setting configurations as well. 
    • Configuration control has not been planned yet.
  • ATEF is a work in progress and is not complete yet.

        Goals coming out of WG:

• • Have device-level configuration and management pattern for all of MECU.
    
    • Find where we want to store configurations.
    • LLNL needs a very convenient means to verify configurations before firing.

        Lingering Questions:

• • How are other large-scale EPICS systems managing this problem? The challenge being configuration management.
    • Device-obstruction layer that sat on the client layer. 
    • SLAC can reach out to other teams to see how they are running this. SLAC's AD division is doing something similar (SCORE).

• In NIF there is a startup configuration 
  • Active setpoint. Constantly changes. 
• What are the qualities that we want in these setpoints:
  • Named sets of configurations
  • Need to know if the readbacks match up
    • Q: What is the threshold of comparison between the readback and setpoint?
  • List of PVs and their values
  • Q: Does autosave and restore really need to be apart of this tool?
    • Maybe not
  • Most of HE-LP laser device react quickly, but do not require time critical moves.  
    
    • They do have time critical items for stopping a shot (This item is more of an MPS topic). 
      
      • Same for RRL.
• Storing data: We have not settled on database/configuration store.
• SCORE: 
  • Article regarding score
  • LCLS implementation of SCORE
  • Java application with relation database (probably Oracle)
  • Elements in tool:
    • Setpoints names (PVs, not groups of elements). 
    • Values
    • Readback Values
  • SCORE is not something that LCLS wants to use because it's so different from what we currently have. Ken proposes we move this behind us. 
  • Ken's understanding that the SCORE application is an apply and verify. 
• Pheobus: Save-And-Restore
  • Takes snapshots of predefined list of PVs at a certain point in time and writes them back. 
  • May support Oracle.
  • Ken's opinion it is the most advance.
  • ECS tend to move away from control studio so we don't use Pheobus.

• PyDM:
  • Are we putting configuration tool at this GUI layer
    • Up for discussion

• We should capture and record the functionality and requirements in the system in Jama. 

•  Input LLNL requirements into Jama Barry Fishler

2/8/2023 Notes: