Vertex Examine and Beam Spot Monitor Shift Instructions

Michiel Sanders
msanders@fnal.gov

2nd March 2009

Introduction

The vertex examine can be used to monitor tracking performance, the beam position at DØ, and the beam quality. In its present implementation, it runs the Level 3 tracking algorithms and it reads the tracks reconstructed by the L3 farm, it runs a simple dca-phi fit to find the beam position, and it runs offline vertex reconstruction algorithms. Moreover, it analyzes Luminosity data to find the z position of the interaction. Finally, it checks the validity of combinations of Luminosity trigger and/or terms.

The BeamSpotMonitor performs several tasks:

• Run the vertex examine during stores.
• Monitor the beam position and send alarms to the Significant Event System if the beam moves too far away from the reference point.
• Transfer the beam position values to COOR's name-value service from where they get picked up at the start of a run by L3 and L2STT.
• Transfer the beam position values to ACNET.

An extensive description of the examine and the beam spot monitoring tool can be found in DØNote 4755.

How To Run

The vertex examine runs by itself during global physics runs, during a store (the requirement is "In store" and a trigger configuration with the string "official/global" in the trigger configuration name). It automatically starts and stops at run transitions. The shifter doesn't have to do anything. The examine runs on d0ol77.

The histograms can be viewed with a GMbrowser that can be started with the following procedure:

1. Log in to d0ol77 (d0ssh d0ol77)
2. Type setup d0online
3. Type start_daq vertex_browser

This browser can be kept running at all times, even in between stores. It will automatically load the histograms of a new run.

The automatic starts and stops are handled by the BeamSpotMonitor. If this program dies (or when the vertex examine does not run during a physics run in a store), the Process Watcher will raise a "major" online alarm, of type MISSING-PROCESS (or WRONG_USER or NOT-SINGLETON).

The BeamSpotMonitor can be restarted with the following procedure:

1. Log in to any online node as user d0run (d0ssh -l d0run <node name>)
2. Type setup d0online
3. Type start_daq beamspotmonitor

In the very rare case where the BeamSpotMonitor and/or the vertex examine seem to be in a confused state, it may be necessary to manually kill the vertex examine process on d0ol77, then stop (stop_daq beamspotmonitor) and restart (start_daq beamspotmonitor) the BeamSpotMonitor. There should be no need to manually start the vertex examine and/or restart the GMbrowser. The logfile of the BeamSpotMonitor can be found in the directory /online/log/beamspotmonitor. Consult this file first before restarting the BeamSpotMonitor.

Saving histograms for pasting in the logbook

Under the "file" menu on the browser window, there are several options for saving the current page, or all pages to a gif or postscript file. Clicking the middle button on the top left of the window will save the current page to a gif file. The saved files will show up in the ~/VertEx_plots directory, when logged is as d0cap or d0cft, or else in /tmp/VertEx_plots_<username> on d0ol77. The plots can be included in the logbook with the normal procedure.

List of Histograms

The following histograms are available in vertex examine:

Good L0 z position:

Beam position from time info in luminosity detector. The spikes are ok, they are caused by a hardware feature.

Good fastz, (Anti)proton bit fired:

Beam quality bits. The majority of the events should have good fastz.

Bad gap south (north) bit:

Gap trigger quality plots. All events should be in the zero bin. If not, the south (north) LM bit is faulty. This results in high gap trigger rates. It is known that the Gap North is not working properly (yet), the experts are working on fixing it.

Reconstructed vertices x,y,z:

Position of all primary vertices (in cm) reconstructed from L3 tracks with offline vertex algorithms. Don't expect the distributions to be centred at zero. Also, the beam position can change from store to store.

Number of vertices:

Number of reconstructed vertices per event. The number of vertices per event will decrease with decreasing instantaneous luminosity.

Number of tracks attached to vertex:

Number of L3 tracks attached to each reconstructed vertex.

Beam x,y offset at z=0:

Position of the beam in x,y at z=0 (in cm). This should be approximately the same as the average primary vertex position.

Beam slope x-z, y-z:

Tilt of the beam in xz and yz planes (dimensionless, or cm/cm).

Number of tracks (L3Track):

Number of reconstructed tracks per event.

Track rsigned, z at DCA, phi, tan(lambda), q over pt, DCA vs phi (L3Track):

Track parameter distributions.
rsigned: distance of the track to the z-axis (beam line) at the point where the track is closest to the z-axis in the transverse plane. This quantity will show a double peak structure whenever the beam is not exactly centred on the z-axis.
z: z coordinate of the track at the point where the track gets closest to the z-axis in the transverse plane.
phi: phi direction of the track. This distribution should ideally be flat, but due the reconstruction of fake tracks in certain regions, and due to detector performance issues (eg, number of enabled SMT HDI's), it usually isn't.
q over pt: charge over the transverse momentum of the track.
DCA vs phi: rsigned as a function of phi. This will show a sine-wave shape if the beam is not centred at the z-axis.

Number of xy,z hits (L3Track):

Number of axial / stereo hits on the reconstructed tracks. Tracks with no stereo hits will show up at tan(lambda) = 0. Also, due to the larger abundance of fake tracks with a small number of SMT hits at high instantaneous luminosity, the average number of hits will increase with decreasing instantaneous luminosity.

Track phi CFT only (L3Track):

Track phi for tracks with exactly 8 axial hits, only CFT hits. If parts of the SMT readout are missing, and the CFT is readout completely, a bump will show up in this distribution, in the region where the SMT readout is missing.

Track phi not-stereo (L3Track):

Track phi for tracks with no stereo hits attached. A bump in this distribution points to the region where stereo hits are missing, e.g., a set of stereo fibres in the CFT.

Hit phi in (+-z) global layer 0,...,16:

Track hit phi for hits attached to reconstructed tracks. Layers 0,...,8 are the SMT layers (from inside out), split up in positive and negative z. Layers 9,...,16 are the CFT layers (from inside out). Dips in the track phi distribution can be correlated to missing hits in a certain layer.

Examples

Beam Position

The beam position is found using the correlation between the track dca and the track phi. For one fit, a sample of at least 30000 tracks is used. The fit is independent of the vertex reconstruction. It is assumed that the primary vertices lie on a straight line parameterized by

x = X_0 + A_x z
y = Y_0 + A_y z

The fit parameters X_0,Y_0 give the beam offset at z=0, parameters A_x,A_y quantify the beam slope in the xz,yz plane.

L3 Tracking

The Level 3 tracking algorithms run in the vertex examine and on the L3 farm use the CFT axial fibres as a starting point and includes SMT hits. Either 8 axial CFT hits, or 7 axial CFT with at least 3 SMT hits are required. If an axial track is found, stereo hits are searched for. At most 2 misses (with respect to the number of axial hits) in CFT stereo or at most 3 misses in global stereo are allowed. There is a cut on p_t > 0.5 GeV.

For the vertexing and fitting of the beam position, only those tracks are used which have at least 10 axial hits in total, and at least one stereo hit. The track histograms show all tracks, before these cuts are applied. This selection means that for a run which does not include the SMT, the track distributions will be filled, but no vertices will be found.

For the Experts

In the default implementation, no selection on trigger bits is made. Vertex examine will process data from the all stream (global physics runs). To process data from any stream (e.g., from a zero bias run), type start_daq vertex_examine anything. You can also select data from a specific L1 trigger. To get zero bias triggers only, type start_daq vertex_examine ALiveBX. For minimum bias triggers, type start_daq vertex_examine Afastz. (Trigger names as defined in global_CMT-8.41)

In Case of Trouble

First contact: Michiel Sanders

e-mail: msanders@fnal.gov
telephone: +49 89 28914137