Data acquisition at Belle

Running a data acquisition system is fun! Despite the amount of data that Belle detector is daily producing, and extremely high impact of Belle physics results to the world, the Belle data acquisition system has been pretty simple, and has successfully completed its mission over 10+ years, on June 30, 2010.

Thanks to the excellent idea to use Q-to-T (charge-to-time) technique to convert the analog signal to timing information and then later on read them out by a single type of time-to-digital converters (TDC) almost everywhere in Belle, maintenance work has been quite reduced. At the later stage, they were gradually replaced with new TDCs based on a system called COPPER, but the concept is unchanged. I was working on the fast and slow control of the Belle data acquisition system.

Fast control means to distribute the trigger signal to every readout system and wait until they are ready for the next event, and it happens at 30 micro second after the trigger. This is not very fast for the recent high energy physics standard, but it still works at Belle, thanks to the excellent KEKB beam condition. For the COPPER system, a new timing distribution system was developed using a serial link technology, which allows us to build a compact and scalable system. The system eventually controlled about 200 modules.

Slow control means to initiate the data taking processes (i.e., start a run) that are running on hundreds of computing systems nowadays. The time scale is order of one minutes. Since there are many unexpected troubles happening everyday, the system could be very flexble to handle them. I have developed a network shared memory (NSM) that is the base of the slow control system.

Listed below are public material that are relevant to my works on data acquisition system at Belle.