Transcript Document
EPICS @ FNAL With emphasis on ILCTA 7/18/2015 Geoff Savage/Margaret Votava FNAL 1 Current EPICS Projects Run II Experiments ILC Test Accelerator (ILCTA) @ FNAL D0 Controls. Vertical Test Facility (IB1) Horizontal Test Facility/Coupler conditioning (MDB) Beam test (NML) using photoinjector currently at A0 Other Proton Driver – EPICS only control system Nova – Seriously evaluating EPICS as control system for DAQ. Maybe also for controls itself? 7/18/2015 Will decide in next few months. Geoff Savage/Margaret Votava FNAL 2 ILCTA Controls Systems (!) ACNET (FNAL) EPICS (ANL) It’s what FNAL AD know and love Drivers/interfaces for FNAL devices already written (HRM) Cryogenics for MDB and NML. Maybe for IB1, but have large IFIX base. IB1 controls system At least partial (maybe complete) control system at MDB DOOCS (DESY) LLRF (Simcon 2.1/3.1 boards) supplied by DESY IFIX (Intellution) EPICS driver written, being debugged Digitizers supplied by DESY A0 is dominated by DOOCS devices. Photoinjector will move to NML being controlled by DOOCS Cryo system at IB1 ILCTA review committee in FY05 recommended picking one all were considered considered capable of meeting technical requirements Recommended DOOCS or EPICS with a slight perference for EPICS because more technical help available Have not yet achieved this (LLRF controller is DOOCS) 7/18/2015 Geoff Savage/Margaret Votava FNAL 3 But wait, that’s not all! Labview and MATLAB applications show up throughout “Will remain for forseeable future engineering development tools” not control systems. Want to limit scope to be used for analysis, but control 7/18/2015 Geoff Savage/Margaret Votava FNAL 4 What is DOOCS Distributed Object Oriented Control System http://tesla.desy.de/doocs/doocs.html Written by a small team at DESY over the last decade Control system for TTF and most likely choice for XFEL Cyro system is EPICS Features Newer than EPICS and therefore makes more modern architecture choices Uses RPC for communication path Devices are objects Includes DAQ support Includes a finite state machine Can talk to EPICS iocs Drawbacks Only user group is DESY, other than then ILCTA Not well packaged for development outside of DESY Support from experts is limited – not nearly as responsive as the EPICS community Much work done here and is getting better EPICS can’t talk to DOOCS devices (but this is a problem with EPICS …) 7/18/2015 Geoff Savage/Margaret Votava FNAL 5 EPICS Development Issues Development is done by a collaboration of users: No standards for package, quality control, support, etc for exetensions Multiple extensions with same functionality GUI support is primary example Larger scale developments/enhancements have no resources ANL (primarily) supports EPICS base (ie, front end support) Extensions (applications) are provided as needed – alarming, GUI support, archiving, etc Have a large todo list, but no organized/funded/supported todo-ers ANL is committed to APS support Several items on todo list are to implement features already available in DOOCS 7/18/2015 Geoff Savage/Margaret Votava FNAL 6 ILC Controls Digression Currently imagined (at least costed) as taking existing controls system as base with significant modificiation. EPICS and DOOCS are obvious candidates Way too early to make this decision Still may decide to write from scratch Global controls group (excluding LLRF) consists primarily of people from ANL, FNAL, SLAC, and a little from DESY add KEK and more DESY for LLRF ANL and SLAC are providing personnel help for ILCTA DESY providing equipment for ILCTA 7/18/2015 Geoff Savage/Margaret Votava FNAL 7 ILC Tests Areas IB1 MDB Will be EPICS, but using the DESY LLRF board Cryo either IFIX or EPICS Mostly TD Personnel Will most likely be all EPICS Cryo is EPICS Geoff working on EPICS PLC klystron interlock driver. Still using old HRMs speaking classic protocol (no EPICS interface, but a DOOCS one) Mostly same software personnel as proton driver (i.e., epics) NML – Under design/construction Will take photoinjector from A0 => photoinjector control will be DOOCS for at least the short term (next year or two) Cryo will be EPICS (modelled after MDB) Evaluating possibility of an EPICS -> DOOCS interface so applications can be either one. It’s an R&D playground in addition to a facility to test cryostat 7/18/2015 Geoff Savage/Margaret Votava FNAL 8 EPICS Collaboration Meeting Next week at Argonne Marginally too late to register Possible OK for US citizens or people who have been to Argonne before http://www.aps.anl.gov/News/Conferen ces/2006/EPICS/index.html 7/18/2015 Geoff Savage/Margaret Votava FNAL 9 FNAL EPICS Personnel About 10 people from all over the lab – most are new to EPICS < 6 months experience. Accelerator division Technical division Dennis Shpakov Particle physics division Sharon Lackey, Dennis Nicklaus, Paul Joireman, Kevin Martin Vladimir Sirotenkov, Fritz Bartlett, Geoff Savage Computing division Ron Rechenmacher, Luciano Piccoli, Kurt Biery 7/18/2015 Geoff Savage/Margaret Votava FNAL 10 EPICS? Experimental Physics and Industrial Control System www.aps.anl.gov/epics Getting started http://www.aps.anl.gov/epics/docs/GSWE.php A Collaboration A Control System Architecture A Software Toolkit 7/18/2015 Geoff Savage/Margaret Votava FNAL 11 EPICS Architecture 7/18/2015 Geoff Savage/Margaret Votava FNAL 12 EPICS Hardware Communication Think of EPICS as a way to get information into hardware registers Software layers Record support Device support Driver support (if needed) Records provide a consistent interface Long in/out, analog in/out Binary word in/out, bit in/out Much existing device support exists that can be reused, just ask Talk to your unique hardware with your own driver that you worked so hard on. 7/18/2015 Geoff Savage/Margaret Votava FNAL 13 Following Diagram for MDB Cryo - largest use so far – mostly done Modulator – done? Klystron interlocks Sequencing - Being developed Simcon (LLRF Controller) Using DOOCS, EPICS port in progress HRM (PD Ion source) Using DOOCS interface PLCs – in progress Geoff Savage/Margaret Votava Timing 7/18/2015 FNAL 14 7/18/2015 Geoff Savage/Margaret Votava FNAL 15 Modulator control using an existing driver. Call the driver from device support. Courtesy of Kevin Martin 7/18/2015 Geoff Savage/Margaret Votava FNAL 16 ADC Readout 12 channel 16-bit VME module VxWorks 5.4 on MVME 2434 Control/monitor with vme device support Data readout and initialization with C++ driver Connect data to epics with symbolic device support? 7/18/2015 Geoff Savage/Margaret Votava FNAL 17 PLC Communication AutomationDirect DirectLogic PLCs Implement network communication through add-on ECOM module DirectNET over ethernet protocol Serial communication requires more infrastructure Wrap the supplied linux driver in the ASYN framework ASYN provides standard interfaces between software layers Use ASYN device support Runs on Linux desktop Use for general purpose I/O and PLC logic 7/18/2015 Geoff Savage/Margaret Votava FNAL 18 Simcon board (LLRF) Board will be used at all ILCTA test areas for the near term IB1 – EPICS on MVME5500 running vxWorks 6.1 MDB – most likely ppc running vxWorks 5.4 NML – mix of DOOCS and EPICS SimCon 3.1 card 7/18/2015 Geoff Savage/Margaret Votava FNAL 19 SimCon 3.1 EDM GUI 7/18/2015 Geoff Savage/Margaret Votava FNAL 20 VME device support from vxworks 7/18/2015 Geoff Savage/Margaret Votava FNAL 21 Conditioning Control Use the sequencer to control the cavity conditioning sequence Implements state transition diagrams Channel access capable Need to adjust what parameters? http://www.slac.stanford.edu/comp/uni x/package/epics/sequencer/ 7/18/2015 Geoff Savage/Margaret Votava FNAL 22 Cryogenic Systems Uses APACS Interface to APACS using OPC Using OPC-EPICS gateway from Bessy Made the GUI look like the ACNET gui Started control via ACNET then transitioned to EPICS 7/18/2015 Geoff Savage/Margaret Votava FNAL 23 Gateways to other control systems EPICS IOC that would be a gateway to DOOCS. This is non-trivial - Ron ACNET - Vlad APACS - Vlad Labview - EPICS community 7/18/2015 Geoff Savage/Margaret Votava FNAL 24