MicroBooNE Discussion 20Aug2010¶
Met with Eric Church, Ron, Kurt to discuss DAQ work for FY11.
Slow controls solution -
based on NOvA (Andrew Norman).
Glenn at Kansas state is doing this work. He is starting to look at EPICS.
currently based on miniboone, which has no slow controls and a very simple
process for reading and storing data.
1000 channels in each of 10 racks. Each rack has one CPU node with one
PCI-express card for reading in the data.
1.6ms drift window, 3 windows kept for a total of 4.8ms
2 words / wire = 16 bits / wire
15Hz windows * 4.8ms * 2bytes/wire * 1000wires * 2MHz sampling = 2.88e8 = 288MB/sec
compression: down to 10% of original
write rate: 5MB/sec for full system.
when I multiple to get the 288, it looks like only one rack to me.
I am not sure where the 5MB write comes from yet, seems as though it takes into account trigger
or something else.
Must see Eric's TDR section for the actual numbers.
Timeline and help requested¶
CD-2: end of the year
pull miniboone: summer 2011 (miniboone says this is too early)
beam at microboone: mid 2012
Our 1/3 FTE starts around 11/1/2010. Until then we will read the TDR section and the
front-end electronics sections.
We will help answer questions about slow controls and EPICS integration and setup.
9/1/2011: one complete test stand and maybe two that show readout and processing,
EPICS system in place, fake target / fake assembler (to feed the DAQ software components).
1/2012: mostly functions DAQ components, integrated and tested.
Kurt mentioned that it looks similar to the NOvA DAQ problem, but at a simpler scale.
The current event builder / logger is a simple c program. We should consider using the
NOvA software for this purpose.