\section{The Burst Alarm System at La Palma} {\bf Current status:} \par The Burst Alarm System {\it gspot} (Gamma Sources Pointing Trigger) is working in La Palma since last summer. It performs a full-time survey of the {\it GRB Coordinates Network} (\g) alerts~\cite{GCN}. Different satellite experiments send GRB coordinates to the \g which distributes the alerts to registered users. The Burst Alarm System is composed of a core program which manages the monitoring of the \g and the communication with the Central Control (CC). It also handles three communication channels to notice the shifters about an alert. It is a C based daemon running 24 hours a day on the {\it www} machine, our external server, in a {\it stand alone} mode. It does not need to be operated and is fully automatic. It manages network disconnections within the external net and/or the internal one. \subsection{The connection to the GCN} The connection to the \g is performed by {\it gspot} through a TCP/IP connection to a computer at the Goddard Space Flight Center (GSFC). This computer distributes the alerts from the satellite experiments through an internet socket connection. {\it gspot} acts as a server while the client, running at the GSFC, manages the communication of the data concerning the GRBs and concerning the status of the connection. \\ The format of the data distributed through the \g differ between the individual satellites and the kind of package. Currently, three satellites participate in the GRB survey: HETE-2~\cite{HETE}, INTEGRAL~\cite{INTEGRAL} and SWIFT~\cite{SWIFT}. The alerts include the UTC, the GRB coordinates (not always), error on coordinates (not always) and intensity (photon counts) of the burst. The first notices from HETE-2 and INTEGRAL usually do not include the coordinates. In few cases only coordinates are distributed in refined notices. The \sw alerts are predicted to arrive with coordinates between 30-80 sec after the onset of the burst. The error on the coordinates from the BAT detector will be 4 arcmin which is smaller than the size of one inner pixel of the \ma camera.\\ In case of an alert {\it gspot} stores the informations and enters an {\bf Alarm State}. The duration of the alarm state depends on the following parameters: \begin{itemize} \item {\bf Darkness of the sky}: Determined from the distance of the sun to the astronomical horizon of 108$^\circ$ zenith; \item {\bf Position of GRB}: The GRB equatorial coordinates are transformed into local horizontal coordinates. The resulting GRB zenith angle has to be smaller than 70$^\circ$. In the case that the moon is shining, this zenith angle limit is reduced to 65$^\circ$; \item {\bf Position of moon}: The angular distance from the GRB to the moon has to be at least 30$^\circ$. \end{itemize} If one or more of these conditions fail, {\it gspot} enters into a \text {yellow}{\bf Yellow Alarm State}. It means that the GRB is not observable at the moment. Currently the program does not calculate if and when the GRB will become observable for \ma. If all conditions mentioned above are satisfied, {\it gspot} enters into a \textcolor{red}{\bf Red Alarm State}, it means that the GRB is considered to be observable now.\\ In both cases (in \textcolor{red}{RED} and \textcolor{yellow}{YELLOW} alarm state) {\it gspot} establishes the communication with the CC and sends the GRB equatorial coordinates (RA/DEC J2000). For the communication with CC the format defined in~\cite{CONTROL} is used. In the same time the shifters and the GRB-MAGIC group is contacted in different ways as described in the next sessions. \subsection{The interface to the Central Control} An interface of {\it gspot} sends all the relevant information to {\it arehucas}. In the case of {\bf NO Alarm State} the standard packages, containing the main global status of the two subsystems, are continuously exchanged between CC and {\it gspot}. In the alert case {\it gspot} starts to send to CC special alert packages, containing information about of the GRB and the ''color'' of the alert. The exchange of the alert packages continues until the following steps occur: \begin{itemize} \item {\it gspot} receives from {\it arehucas} the confirmation that it has received the alert notice; {\it arehucas} must send the alert back in order to perform a cross-check of the relevant data \item the alarm state expire after {\bf 5 hours} \end{itemize} At the moment {\it arehucas} informs the shift crew about the alert and takes further steps only in case of red alerts. In this case a pop-up window appears with all the alert information received by the burst monitor. The operator has to confirm the notice by closing the pop-up window. He can decide to stop the current scheduled observation and to point the GRB. A new button in the CC will be displayed and allows to point the telescope to the GRB coordinates by pushing it. \subsection{GRB archive and emails to the GRB-mailing list} In case of an alert -- even if it did not contain the necessary coordinates -- the information is translated into ''human language'' and stored in ASCII files. At the same time an e-mail is sent to the MAGIC GRB-mailing list. \subsection{The GRB web page} The status of the GRB Alert System and relevant informations about the lastest alert are displayed on a separate web page. The page is hosted at the web server in La Palma. The address is the following:\\ \qquad \qquad http://www.magic.iac.es/site/grbm/\\ The web page updates itself automatically every 10 seconds. In this way the status of the Burst Alarm System can be checked by the shifters and from outside. \subsection{The acoustic alert} A further CC-independent acoustic alarm called {\it phava} (PHonetic Alarm for Valued Alerts) will be installed in La Palma very soon. It will provide a loud acoustic signal even if {\it arehucas} is switched off, so that persons in the counting house will be noticed about the alert situation. The signal will be on as long as {\it gspot} stays in alarm state, and in any case for a minimum of 1 minute. This device feature also a display with the status of the system and the alert. \subsection{Summary of alerts received until now} Since July, 15th 2004 {\it gspot} has been working stably at La Palma. It received about 100 alerts from HETE-2 and INTEGRAL, out of which only 21 contained GRB's coordinates. Time delays were in the order of several minutes to tens of minutes. The Burst Monitor can be considered bug-free since November 2004. From this moment we received the following two significant alerts:\\ \begin{tabular}{lllcccl} 19th & December & 2004 & 1:44 am & INTEGRAL satellite & Zd $\sim 60^\circ$ & Time delay 71 sec.\\ 28th & January & 2005 & 5:36 am & HETE-2 satellite & Zd $\sim 65^\circ$ & Time delay 73 min. \\ \\ \end{tabular} In both cases the weather conditions at La Palma were very bad. \subsection{Experience of SWIFT GRBs until now} According to the \sw homepage~\cite{SWIFT} the satellite detected 12 GRBs since mid December last year. The bursts were detected by chance during the comissioning phase. The satellite did not send the coordinates on time to \g. Anyhow, in the current sample are two bursts which in principle could have been observed by \ma:\\ \begin{tabular}{lllcc} 19th & December & 2004 & 1:42 am & Zd $\sim 65^\circ$ \\ 26th & December & 2004 & 20:34 am & Zd $\sim 52^\circ$ \\ \\ \end{tabular} \subsection{Comparison between the satellite orbits} Figure~\ref{fig:orbit} show the difference between the orbits of the \sw, \he and \ig satellite. The \sw and \he satellites are situated in a circular orbit with 20.6$^\circ$ respectivly 2$^\circ$ inclination. The revolution period of the \sw and \he satellite add up to 100min. The \ig satellite is situated in an highly eccentric orbit with a revolution period around the Earth of three sidereal days. \par It is difficult to make strong conclusions from the individual satellite orbit. The orientation of the satellite FOV is influenced by the scheduled targets. Hovewer, \sw is the satellite with the largest inclination and overlaps mostly with the FOV of \ma. This increases the chance to recive {\bf Red Alarm} from this satellite. \begin{figure}[htp] \centering \includegraphics[width=0.7\linewidth]{GCNsatellites.eps} \caption{Orbits of the \sw, \he and \ig satellites} \label{fig:orbit} \end{figure} \subsection{Routines to be defined} The Burst Alarm System is currently able to provide the minimum features needed to point and to observe a GRB. However, in order to improve our efficiency to point and observe GRBs, several procedures have to be defined: \begin{itemize} \item {\bf Yellow Alarm strategy}: The strategy to follow a {\bf Yellow Alarm} is not defined yet. In such a case the CC does not undertake any steps, except confirming the alarm notice to the Burst Monitor. We do not calculate if and when the GRB will become observable. It would make sense to check if during the period of 5 hours we could point to the burst. Then, the Alarm System should change to the {\bf Red Alarm State} at that time and allow the observation. \item {\bf Sequence of alerts}: How to deal with new alerts that are distributed during the time that {\it gspot} is in alarm state? Currently {\it gspot} locks its alert status until it exits the alarm state (see session 2.2). This feature was implemented to avoid any loose of the GRB information. Such a situation can occur when for example more than one burst alert is send before the shift crew starts the CC. To solve the problem we will change the {\it gspot} routine by implementing a list of the available GRB alerts. \par If more than one alert is present in the list, the program will weight the possible GRBs on the following criteria: (1) the total time of observability within the canonical 5 hours, (2) the intensity of the burst and (3) the time until the GRB becomes observable. The information of the best GRB will be send to the CC. \end{itemize} %%% Local Variables: %%% mode: latex %%% TeX-master: "GRB_proposal_2005" %%% End: