3.1 Introduction and Overview

The DELPHI detector has been equipped with an automated system for monitoring and controlling technical aspects of the experiment, such as high voltages and gas supplies, for reporting and acting on changes in the status of the detector or its environment, and for maintaining the safety of the equipment.

This Slow Controls system should be distinguished from the Data Acquisition system (DAS; see section 2.10), which is responsible for the digitizing and recording of each physics event -- the products of the electron-positron collision. The emphasis of the Data Acquisition system is on efficient triggering and fast readout, since the electron and positron bunches cross every 11 $ \ensuremathbox{\mu\mathrm{s}}$. In contrast, the Slow Controls reacts to events that can take from seconds to hours to develop, but is more concerned with reliability, particularly due to its safety requirements.

The overall structure of the DELPHI Slow Controls system is summarized in figure 3.1.

Figure 3.1: Diagram of the overall structure of DELPHI Slow Controls system represented by the example of the high voltage control of the Outer Detector.
\includegraphics[width=\textwidth]{sc-structure.ps}
As can be seen, the system is highly modular and highly distributed with many programs running on both high-level (VAX) and front-end (G64) processors.

The DELPHI Slow Controls operator makes use of two main graphical displays, shown in VAXstation windows. The status display gives a colour-coded representation of the state of the various detector partitions. These states are defined in the State Management Interface (SMI) (see section 3.8.1), which is a hierarchical set of objects representing different aspects of the detector as seen by the Slow Controls system. SMI is also responsible for passing commands (either to the whole of DELPHI or to a particular detector partition) down to the appropriate subsystems which act on them. During 1994 an even higher level of SMI-based automation, called Big Brother (see section 3.8.3), was added to coordinate the actions of the Slow Controls and the Data Acquisition systems with the states of the LEP machine. The error message display shows outstanding anomalies in a textual form, grouped according to detector partition. These messages are handled by the Error Message Utility (EMU) (see section 3.6).

Both SMI and EMU show conditions determined by the Elementary Processes (EP) (see section 3.5.1), which are the lowest-level VAX control programs. The Elementary Processes are also responsible for handling SMI commands, logging state changes to the Status Update Database (see section 3.7.2) for use by the offline analysis, and providing a route for occasional expert intervention, using a user interface, HIPE (section 3.5.2).

The Elementary Processes communicate over ethernet (using the Remote Procedure Call (RPC) protocol) with the front-end control and monitoring microcomputers, the G64 crates (see section 3.3). The G64s monitor and control a variety of different types of hardware using digital and analog monitoring and control devices. Most high voltage supplies are controlled by an intelligent CAEN high voltage unit [80] (see section 3.3.1), which is in turn controlled and monitored by the G64.

The unified gas system, which controls and monitors the flows and mixtures of gases supplied to various parts of the detector, and the GSS safety system, which monitors the detector and its environment for hazardous conditions, use different structures (see sections 3.9.1 and 3.9.2), but are integrated with the rest of the Slow Controls at the SMI and EMU level. These software links are complemented by a system of hardwired interlocks.

A brief description of the DELPHI Slow Controls system has been given previously [81]. In this chapter I give a detailed and considered description of the systems employed. A less technical version has also been published [82]. For detailed descriptions of the gas and safety systems the reader is referred to separate publications [83,84]. The slow controls of the other three LEP experiments have been described elsewhere [85,86,87].

Tim Adye 2002-11-06