IFR Fast Monitoring

 

Author: IFR group

Version 0.1

Preliminary

 

1. Introduction

In this document the Fast Monitoring of RPC will be described. IFR data are composed of strip hits and TDC measurements. They are independent streams of data produced by different hardware and so they require different and independent monitoring.

 

2. Fast Monitoring

The main goal of Fast Monitoring (FM) is to check the status of the detector during normal data taking and discover, as soon as possible, any hardware related problem.

Another goal of FM is the determination of quantities needed to configure the detector, e.g. dead or noisy channels. For the strips this information can be used during Feature Extraction and for the TDC it must be stored in the chip to save the event buffer memory.

Fast Monitoring is the storage of data in histograms during data taking in the Online Event processing (OEP) either before L3 selection reading the information of the Tagged Containers (TC) or after the Prompt Reconstruction (PR) using the Digis.

Fast Monitoring should run on the maximum number of events without affecting the performance of DataFlow.

The Distributed Histogramming Package (DHP), written by the OEP group, will be responsible for providing histogramming facilities. The histograms, 1-D or 2-D, will be filled during the data taking and stored in the database at the end of the run. Some histograms will be continuously monitored, by comparison with a reference or checking some confidence level. The monitor frequency must be defined. It could be at the end of the run, or each half hour. An alarm message can be issued if this comparison fails.

A limited amount of computation, like to work out the mean or the variance of the histograms should be possible at the end of the run before store them in the DB. On request of the operator all the histograms will be visible during data taking. N-tuples will be supported only for debugging purpose due to the difficulties to run the on distributed platforms.

The first group of histograms in Tab[1] is devoted to check the status of the readout electronics, and must be filled before the L3 trigger, with the information provided by the TC.

 (Hardware configuration:4 ROM, 2 ICC per ROM, 7 IFB and 4-5 ITB per ICC, 1024 strips per IFB and 96 Fast-OR per ITB, FEE = IFB or ITB)

 

Parameter

Plot type

Channels

# of plots

Event size (bytes)

1-D

100

1

Tot. strips Mult.

1-D

250

1

Tot. TDC Mult.

1-D

100

1

FEE Occup.

1-D

92

1

Mult./FEE

1-D

50

92

Mult./crate

1-D

100

8

Occup./IFB

1-D

1024

56

Occup./ITB

1-D

96

36

Occup/FEC S( 1)

1-D

96

92

Occup/FEC S(nh)

1-D

96

92

Occup/FEC S(nh)**2

1-D

96

92

 

Table 1: hardware plots

 

The last three plot, S(1), S(nh) and S(nh)**2 are useful to compute the mean and the dispersion of the multiplicity.

The second group of histograms, in Tab[2], can help to check out the status of the RPC chambers before the prompt reconstruction using information coming from strips and TDCs after the conversion of the TC's to Digi's.

(Barrel: 6 sectors, 19 layers and 12 FECs per layer per sector; Endcap: 4 half-doors, 18 layers and 24 FECs per layer per half door. The RPC module is an independent detector, i.e. in the barrel 3 modules are joined to form a chamber. Inner RPC: 4 sectors, 2 layers/sector 16 FECs/layer.)

 

Parameter

Plot type

Channels

# of plots

Zone

Mod. mult.

1-D

50

1

 

Mod. occ.

1-D

774

1

 

Strips mult/mod.

2-D

2-D

2-D

50 x 57

50 x 108

50 x 8

6

4

1

Barrel

Endcap

Inner

 

Table 2: Strip plots in OEP

 

More plots can be added to monitor the Gas, LV, HV distributions.

The reconstructed cluster size in the IFR is very sensible to some parameter of the detector like the gas composition. The information about the cluster size and the time spread of the cluster hits can be obtained only after the Prompt Reconstruction. In Tab[3] there is the list of requested cluster plots.

 

Parameter

Plot type

Channels

# of plots

Zone

Tot. clust. mult

1-D

100

1

 

Tot. clust. size

1-D

100

1

 

Clus. mult./mod.

2-D

2-D

50 x 57

50 x 108

6

4

Barrel

Endcap

3D clust. mult.

1-D

100

1

 

3D clust. theta angle

1-D

100

1

 

3D clust. phi angle

1-D

100

1

 

#hit layer/3D clust.

1-D

50

1

 

#digi/3D clust.

1-D

50

1

 

Tot. clust. T spread

2-D

100

1

 

Clust. T spread/mod.

2-D

2-D

50 x 57

50 x 108

6

8

Barrel

Endcap 

 

Table 3: Cluster plots in PR

 

The alternative solution to the 2-D multiplicity plots is 1-D plots for each chamber, in total 744 plots for hit multiplicity and the same number for cluster multiplicity are requested (1548 in total). The histograms can be filled during data taking and saved in the Condition Database at the end of the run.

The last group of histograms in Tab[4] is devoted to check the global performance of the IFR by plotting some variable of physical interest, computed with the parameters available in the online. Some plots will be redone in the offline with better calibration constants.

  

Parameter

Plot type

Channels

# of plots

Zone

Prob. to be muon

1-D

100

1

 

Prob. to be pion

1-D

100

1

 

# of neut. hadrons

1-D

20

1

 

layer efficiency

1-D

1-D

1-D

19 x 3

18 x 6

8

6

4

1

Barrel

Endcap

Inner

 

Table 4: IFR Particle Identification

 

For performance reasons there will be limitations on the total amount of memory reserved for histograms. Each subsystem will have up to 2 MB of space for his monitoring histograms, that translates to 500000 bins. In this documents IFR requests of the order of 200000 bins.


This page maintained by Maurizio Lo Vetere , ( lovetere@ge.infn.it )

Last modified: Mon Mar 30 18:32:52 CET 1998