Scanning Tips and Tricks

From LHEP Wiki
Jump to navigation Jump to search

Scanforth

  • follow the tracks from the primary vertex
  • at all stopping points, open a volume
  • volume: 3 up, 5 down, 7x7 mm2 (max 7x7 mm2 at the 5th plate down)

About Brick Number

Example you have to scan brick 36423

in the scanning scripts put 36423 but in Sysal (ChangerStatus.txt file) the brick number MUST be ONLY the 4 last number : 6423

Create Prediction (for the ScanForth for ex)

manually create prediction for the scan forth: Create Scanforth Prediction

or here:

{
#include "scan.C"
EdbScanProc        sproc;
sproc.eProcDirClient="/terabig/scan/MICROSCOPE3/ONLINE"; // work directory
EdbPattern pred;
// track stopped at pl24	
pred->AddSegment(1,44480.9,16508.4,-0.123,0.04,32); // last BT from the track in linked_tracks.root (-> affine transformation applied!)
EdbAffine2D *aff = new EdbAffine2D(0.998787,-0.003071,0.000637,0.995767,326.291107,-348.701843); // AFF from Pl24
int id[4]={36423,24,1,300};	
aff->Invert(); // to "unapply" AFF transformation
pred.Transform(aff); // to "unapply" AFF transformation
pred.Print();
OverwriteErrors(pred,50.,50.,0.,0.6,0.6);
sproc.WritePred(pred,id); 
}

Screen Function

All the scanning processes should be started in a screen function in order to be able to access from everywhere. In addition, it's very easy to kill an process when it hangs (just kill the screen).

- screen -S myscreen : create a new screen
- screen -r : reattach to the screen
- screen -x : additionally attach to the screen
- inside the screen in order to deattach: Ctrl A D
- inside the screen to close: exit

Avoid Odyssey Crashes

Start process with a remote desktop and when it's running, close the remote desktop.

rdesktop -g 1272x930 -k de-ch -u scanner lheppc9


Scanning efficiency issues

Dear Francesco,

this happens when you hit the tracking timeout. In this case, ZERO tracks
are generated.
Try increasing the timeout in SmartTracker8.xxx
Cheers

Cristiano

----- Original Message -----
From: "Francesco Di Capua" <fdicapua@mail.cern.ch>
To: <opera-scanonline@cern.ch>
Sent: Friday, July 18, 2008 10:38 AM
Subject: sysal with high fog


>
> Dear All,
> we received a brick (n.117472) related to the CC event 22877818 with very
> high level of fog,
> we located and reconstructed vertex in intercalibration areas (see
> picture),
> in order to have a reasonable scanning speed the intercalibration areas
> was done with high treshold (900,840) and slope tolerance to 0.4 rad.
> Quality of data is not very good, low efficiency and a lot of views with
> few tracks found (see surface plot).
> Trying to do a better quality volume for vertex analysis, we tried to
> decrease tresholds and increase slope tolerance to 0.6, but we found a
> strange behaviour of Sysal, the scanning seem to start but not microtracks
> at all are found during data taking.
> Do you have any suggestion for parameters to be used in this difficult
> conditions?
>
> Thanks
> Francesco
>
>


Track momentum calculation

Letter  from Frederic:
Dear all,

I add a new function in check_vertex_totalscan.C which allows to 
estiomate the momentum of a track (using MCS from Magali study). You 
just have to call the function PMCS with the trackID as argument. 
Exemple below:

void PMCS(int IDTR)
{
   //EdbTrackFitter tf; // create track fitter object
   EdbMomentumEstimator tf;
    tf.SetParPMS_Mag(); //set parameters for detheta=f(theta) (angular 
dependance of BT resolution)
     tf.eDT0=0.002;      //Change a parameter
     for(int i=0; i<gAli->Ntracks(); i++ ){
       EdbTrackP *t = gAli->GetTrack(i);
       float p = tf.PMS_Mag(*t);
       float pmin=tf.ePmin;  // 90% CL minimum momentum
       float pmax=tf.ePmax;  // 90% CL maximum momentum
       if (IDTR == t->ID())cout << p << " - " << fabs(p-pmin) << " + " 
<< fabs(p-pmax) <<endl;
     }
}

Track momentum calculation for scanback or scanforth track

 A new script /terabig/scan/macros/testP.C do the momentum calculation of a track using data 
 from scanback or scanforth. It works in the same way as explain above. (see comments inside
 the scripts for details).

Vertex between 2 BT (or tracks)

 A new script /terabig/scan/macros/vertex_BT.C allows to calculate the vertex position between 2 BT.
 It returns the vertex position and the IP (the closest distance between the 2 BT).
 In the same script you have another possibilty to estimate the vertex position between a known BT 
 ( that you have to specify) and all the BT from  a cp file (that you have to specify).

IP between a vertex and a BT (or track)

 The script /terabig/scan/macros/vertex.C allows to check if a BT from a cp file 
 can be attached to a known vertex that you have to specify.
 The output is the IP of the BT w.r.t to the vertex.

Change placement of emulsion on stage

to do this, you have to change two settings in the opposit way

  • in PlateChanger, change in Stage Homing the Stage limiter offset on X and Y
  • in Machine default of the Stage, change the XYreference (in opposit way)