A centralized control system for the DAΦNE Φ factory

Abstract

We describe the control system for the new DAΦNE Φ-factory under construction at the Frascati National Laboratories. The system is based on a centralized architecture for simplicity and reliability. A central processor unit coordinates all communications between the consoles and the lower level distributed processing power, and continuously updates a central memory which contains the whole machine status. This memory constitutes the machine database prototype. A simple message passing scheme built on a system of mailboxes takes advantage of high speed busses and of Fiber Optic interfaces. Macintosh II personal computers are used as consoles. The lower levels are all built using the VME standard. I. DAΦNE The DAΦNE accelerator complex [1] of the INFN Frascati National Laboratories consists of a two ring colliding beam Φ-Factory and of a 510 MeV e+/einjector for topping-up. (See Fig. 1). The project has been approved by the INFN Board of Directors in June 1990 and the engineering design has started in January 1991. Construction and commissioning is scheduled for the end of 1995. The luminosity target is ~1033 cm-2 sec-1. The main features of the Φ-factory are : electrons and positrons circulate in two separate storage rings and collide at an horizontal half-angle θx = 10 mrad (in one or two interaction points) in order to achieve high collision frequency without parasitic crossings; the novel design of the magnetic lattice is a 4-period modified Chasman-Green type, with a 1.9 Tesla normal conducting wiggler magnet inside the achromat. a crab-crossing option is contemplated (if needed). injection is obtained through a LINAC and an accumulator damping ring. Fig. 1 : The DAΦNE Φ-Factory layout II. SYSTEM STRUCTURE Fig. 2 shows the general architecture of the control system. Three levels are defined: PARADISE (PARAllel DISplay Environment) is the top level, implementing the human interface. Several consoles, built on Macintosh II personal computers, communicate with the rest of the system through high speed DMA busses and fiber optic links. PURGATORY (Primary Unit for Readout and GATing Of Real time Yonder) is the second and central level of the system. It essentially contains only a CPU and a Memory in a VME crate. The CPU acts as a general concentrator and coordinator of messages throughout the system. This does not constitute a bottleneck, since we use very high speed busses (MacVee[2]) and fast fiber optic links. A measurement of the throughput of these busses [3] shows performances more than one order of magnitude better than using a conventional LAN. Using a polling mechanism the CPU also checks the lower level units, which relay significant changes in the machine situation. The updated information is stored in the central memory, where it is accessible from the consoles and it represents the prototype of the machine database. HELL (Hardware Environment at the Low Level) is the third level of the system and it is constituted by many (about 60) VME crates distributed around the machines. Each crate is equipped with at least one CPU which performs control and readout of the related elements in the machine. Only significant changes in the parameters are transferred to the Purgatory, thus hiding useless information from the central processor. A first estimate of the system gives about 7000 channels to be controlled.