RG-L

This is the logbook of Felix Touchte Codjo. It aims to capture all the instants of the ALERT experiment.

DATE → 25-06-06

(25-06-06 10:27:54) A good track in AHDC

event ~165 in run 21919_00040

DATE → 25-06-02

(25-06-02 09:55:07) AHDC occupancy that is not an issue

Here is the link to the JLab logbook: https://logbooks.jlab.org/entry/4386523

DATE → 25-05-27

(25-05-27 08:38:08) AHDC cables reconnected

Need to update the correspondance between dream channel, slot (or FEU) and the HV sector

DATE → 25-05-21

(25-05-21 04:47:27) My personal HIPO to ROOT

This block of code can be properly use to develop a GUI that plots the histograms of all banks following specified cuts

DATE → 25-05-05

(25-05-05 00:08:37) Raw hit cuts

The current dev version of coatjava, after 12.0.5t, does not reconstruct tracks. Example, in the run, 21494.00041, tracks are reconstructed with coatjava 12.0.3t but not with dev. What is strange is that, the cut on leadingEdgeTime is more restrictive in 12.0.3t compared to dev. All lot of things have changed since then : ahdc geometry; raw hit cuts are now load from a ccdb. I think we have a lot of mismatches in the reconstruction due to the update in the ahdc geometry. To confirm this, I have to run some simulations first. I planned to ask Mathieu how to do that. That will be a good opportunity to define a t0 for simulation.

DATE → 25-04-25

(25-04-25 12:32:05) A unified ALERT expert

from now, I'am the unique ALERT expert. I have the responsability the contact remote AHDC and ATOF expert on their personnal phone number.

(25-04-25 12:30:41) Target purge

Whitney proposes to purge the target at the beginning of each run. This can be not productive. So we (almost) agree to do 3 purges per day. One for each shift take over.

(25-04-25 12:23:20) AHDC reconstruction review

I closed this PR #599, as it highlights the fact that we need a whole review of the reconstrcution service. It's good, but it is not a priority. Let's analyse the data !

(25-04-25 12:04:37) Target purge

When we change the gas type, we should do a long 1 hour purge (6 auto purge of 10 minutes). When we don't change the gas type, a regular short (auto) purge is sufficient : to be done once per day. Let's say every day after the RC meeting.

DATE → 25-04-24

(25-04-24 17:46:00) Phi PreCluster issue

Phi in PreCluster is no longer consistent with the new geometry. Link to the issue #598.

DATE → 25-04-23

I think the issue comes from the fact that each call to drawOccupancy(), remove columns (and with, no auto widget like TextView_occupancy)

(25-04-23 18:54:34) Conversion DREAM channel to HV sector

Current algorithm. Warning: sectors are wapped following 1 <-> 2, 3 <-> 4, 5 <-> 6, 7 <-> 8. Please refer to GitHub.

(25-04-23 18:49:32) Identify HV sector with low gain in AHDC

With the help of Nathan, I went down in the Hall B and placed the corresponding HV in new power supply spot.

(25-04-23 18:44:58) Successfully reper HV sector in amon

I created in a script, amon/ressources/mapping, that generated C++ code, AhdcMapping.h and .cpp, from the AHDC translation table.

DATE → 25-04-20

(25-04-20 05:16:21) Occupncy tab in amon

A new tab dedicated to the occupancy study has been created. This should allow us to easily indentify possible broken wire.

DATE → 25-04-18

(25-04-18 13:26:28) AHDC occupancy withcuts

Run 21494. We add cuts a the stage of the reconstruction (HitReader). The values of these cuts are read from a CCDB, with a sligth change 180 <= ped <= 360. With these cuts we goes from 60 % to 6 % of occupancy in the first layer of the AHDC. The goal is to maintain a rate of 5 %. Study done over 65k events.

(25-04-18 13:23:20) AHDC occupancy without cuts

Run 21494. The occupancy in the first layer of AHDC is very high. We need to reduce this number before the reocnstruction.

(25-04-18 11:43:42) Issue in AHDC::track bank

We don't want to fit tracks with less that 3 points. That's the reason why we got weird value on px, py, pz : NaN and Infinity.

DATE → 25-04-16

(25-04-16 17:49:20) PR GitHub CI failed

The error is related to a ccdb connection. I should be fixed later automatically.

(25-04-16 17:37:41) Load row hit cuts of AHDC from CCDB

Pull request submitted #572

(25-04-16 16:09:23) CCDB for raw hit cuts

Nathan created a new CCDB table. link

(25-04-16 16:05:28) Correlation between cuts in amon

This study aimed to determine good cuts to reduce the AHDC occupancy at the reconstrcution stage, in HitReader. We plan to load these cut parameters from a CCDB.

(25-04-16 16:02:21) Add adcOffset cut in amon

I have also added a check button, to decide wether or not, we should hide the wires/hits of a given even that don't satify th cuts.

DATE → 25-04-14

(25-04-14 16:54:23) Target purge

It has to be done once per day by the AHDC/target expert in coordination with the Run Coordinator. Due to daq issue, I will probably need to end the recent run, "stop" the beam... Today, around 3pm, I did it with the help of Sangbaek and Bryan. We set the system in auto mode.

(25-04-14 16:51:00) ALERT bank names nomalized

Renormalization of ALERT bank is effective. The associated PR has be appreoved and merged in the development branch. link

(25-04-14 16:47:44) Bin numbering fixed in mon12

A pull request has been submitted to fix the bin numbering issue in AHDC 2D histograms. Waiting for approval. link

(25-04-14 10:46:16) Fix bin numbering in mon12 for AHDC

The x axis range from 1 to 100 with 100 bins. The number of bins is not good at all. In this case, we only need 99 (the maximum number of wires on a layer).

DATE → 25-04-12

(25-04-12 22:09:16) Mismatch between geometry and reconstruction service for AHDC

I'm the responsible of this mismatch. :-(

That must be the occasion to prevent people to use hard coded definitions but only refer to the related service, in that case, the geometry one.

DATE → 25-04-10

(25-04-10 15:38:26) mon12 update

Now I have the lead on the modifications that should be done for the AHDC in mon12. The number of bins was hard coded in coatjava; so each time it changes, it should be modified in mon12. It will be better to load it from a ccdb.

(25-04-10 13:30:49) Extend AHDC tracks

The banks AHDC::Track, AHDC::KFTrack and AHDC::Hits have been extended. The pull request has also been accpeted. Note that, some tracks can share the same hit. In that sense, the last track overwrite the track id in the AHDC::Hits bank if it shares some hits with the previous track. 1) A solution could be to duplicate the AHDC hits each time a new track is found. 2) Or, as we don't have most of 10 hits (for example), we can multiply the track id by the same multiple (e.g 137 if the this hit is associated to the track number 1, 3 and 7). The issue with this last option is that we will not be able to store other quantities such as sum_adc, sum_resisuals or chi2. 3) We can create a AHDC::KFHits bank that is well fitted to store this track id. But at end, it rejoin the point 1).

DATE → 25-04-09

(25-04-09 20:36:32) AHDC::KFTrack issue fixed

The issue has been fixed by Eric. Pull Request

(25-04-09 15:46:09) Only one track is computed in AHDC::KFTrack

This an issue !

(25-04-09 15:06:56) Remove redundant track id in AHDC::Hits

The track id is the same for Track and KFTrack. So track_id_kf has been removed. The KF only recomputes some parameters of an existing Track.

(25-04-09 15:01:26) Add a specific track id for KFTrack in AHDC::Hits

Try two track id for Track and KFTrack

(25-04-09 12:56:20) Update AHDC::KFTrack structure

Do trackId in Track and KFTrack can be differents ?

(25-04-09 12:56:04) Update AHDC::Track structure

Do trackId in Track and KFTrack can be differents ?

(25-04-09 12:54:39) Update AHDC::Hits structure

I wonder if we need a trackId for each Track and KFTrack ?

DATE → 25-04-08

(25-04-08 17:23:09) State of AHDC::KFTrack

Only contains x, y, z, px, py, pz. Need to add more information such as : deposited energy, track id (also in AHDC::Hits), p_drift, chi2, sum of residuals...

(25-04-08 17:22:55) State of AHDC::Track

Only contains x, y, z, px, py, pz. Need to add more information such as : deposited energy, track id (also in AHDC::Hits), p_drift, chi2, sum of residuals...

(25-04-08 15:53:59) AHDC::KFTrack is filled

I thought we add an issue with the two definition of Hit (ahdc.Hit and "KF.Hit"), but it's not. That's why the residual is filled. Useful command line :

hipo-utils -dump output5.hipo -e AHDC::KFTrack -b AHDC::Hits -wait 1 | more

(25-04-08 15:47:00) AHDC HitReader issue fixed

I add to modify the rec/ahdc/Banks/RecoBankWriter.java. The time attribut is now filled.

(25-04-08 12:35:14) AHDC CCDB

The AHDC CCDB are well read

(25-04-08 12:38:54) AHDC HitReader

The doca and the time are normally computed.

(25-04-08 12:33:18) AHDC HitReader

In the AHDC::Hits bank, the time is not filled. I assume the KF Hit masks the Hit from HitReader. ISSUE TO BE FIXED !!!

DATE → 25-04-07

(25-04-07 15:17:48) Reconstructed data

Example of the AHDC::Track bank for the run 21317.00010. We removed some hits that do not satify these cuts PR #540

(25-04-07 02:14:56) CCDB for ATOF

I created a new branch check/atof-ccdb, I started to modify it. I got some good error when I used recon-util. In some cas, the order id is not accessible, I need to redefine a new key value

(25-04-07 01:27:37) Commissionning

Is a track ?

DATE → 25-04-06

(25-04-06 22:40:08) ATOF::tdc order

Run 21171

(25-04-06 22:39:48) ATOF::tdc component

Run 21171

(25-04-06 22:39:31) ATOF::tdc layer

Run 21171

(25-04-06 22:38:55) ATOF::tdc sector

Run 21171

(25-04-06 22:27:58) ATOF numbering

Definition of the sector, layer, component and order for the ATOF

(25-04-06 17:33:50) 3D view in AMON

Fisrt implementation of a 3D view in AMON

DATE → 25-04-05

(25-04-05 19:48:30) Study of the zero suppress effect

The previous logs show the consistency of the wftime with the different mode of ZS. The first channel (== wftime) refers to the time of the first sample (expressed in term of bin number) according this reference.

(25-04-05 19:24:09) Study of ZS effect, case ZS=1 && Zstyp=1

Run 21183, case with zero suppress type 1. Here the wftime varies between 0 and 49.

(25-04-05 19:17:41) Study of ZS effect, case ZS=1 && Zstyp=0

Run 21175, case with zero suppress type 0. Here the wftime is always 0.

(25-04-05 19:14:06) Study of ZS effect, case ZS=0

Run 21123, case without zero suppress. Here the wftime is always 0.

(25-04-05 18:49:30) PR approved (t2d implemented)

This PR has been approved yesterday. I just notice that, for the case of the ATOF, the key_value may not be the best one. I need to know the range of each sector, layer, component for the ATOF.

DATE → 25-04-04

(25-04-04 13:44:03) Reconstruction is executing

Issue with AHDC::Hits --> wire fixed. The numbers are reasonable now. I needed to tell hipo-utils that the bank definition has changed or re-decode the evio files with the new bank definition.

(25-04-04 11:05:22) CCDB for ALERT are successfully loaded.

The timeOffsets's error was due to a wrong key value. In the AHDCEngine, I also needed to specify yes or not, the reconstruction requires constants.

DATE → 25-04-03

(25-04-03 18:25:30) Try to load CCDB tables

Here is a screeshot of the error

(25-04-03 15:38:44) Implement time2distance

The implementation of the time to distance in coatjava is ongoing. Already done :

Modification the AHDC::Hits banks
- I added time = leadingEdgeTime - t0
- Now, doca refers to the time from time2distance calibration
- residual refers to distance from time2distance calibration minus the distance of closest approach computed at the Kalman Filter stage

(25-04-03 15:37:19) PR approved

The PR for the constant calibration loading has been approved. Compilation OK ! Now, need to check the execution on decoded files.

(25-04-03 14:02:22) PR open

A new pull resquest is open to load calibration constants for AHDC and ATOF reconstruction.

(25-04-03 09:22:09) Deal with overlap in mon12

Some histograms in time1D are also presents in 1D. But those of 1D are not well scaled. I prefer to conserve those of time1D. This needs to be improved in my PR or later.

(25-04-03 09:19:41) Update mon12

Delete timeMax histogram and add integral histogram in the time1D tab of mon12/AHDC. The histograms have also been reorganized.

DATE → 25-04-02

(25-04-02 16:50:50) Note, reconstruction (time2distance)

Here is a link to another detector. It gives an example of procedure to read calibration constant. link

(25-04-02 16:19:35) Note, reconstruction

We have two DOCA. One computed from the time, that needs t0 and time2distance and used in FreClusterFinder, and another one estimated by the Kalman Filter. Now the residual could be the difference between the two DOCA. Before, it was compared to the DOCA from simulation.

(25-04-02 15:55:53) clas12Tags/(gemc)/source updated

No need to update gemc by reducing the number of samples, a commit made by Raffaella already use coatjava notation in gemc. Link.

(25-04-02 15:08:46) Implement the doca in reconstruction, t0, time2distance, protection

Until now, the doca is obsiouly wrong. We need to use the t0 and apply a time2distance to the corrected code. We also need to distinguish real data and simulation data.

(25-04-02 11:51:17) Add adcMax hist in time1D tab of mon12

Fix count issue (e.g ToT) and add adcMax hist. Run 21042

(25-04-02 11:29:16) Add new tab in mon12 for 1D time histograms

Add 1D histograms for timeMax, leadingEdgeTime, constantFractionTime, timeOverThreshold, wftime. Run 21042

(25-04-02 10:40:58) Add a 2D hist for wf::time in mon12

Issue fixed. The cut on the last layer has been changed. Run 21042

(25-04-02 10:29:02) Add a 2D hist for wf::time in mon12

The first channel time, AHDC::wf --> time is now plotted in the AHDC time tab. Run 21042. We can notice an issue. The layer 8 is not filled. This is due to a cut made during the last pull request, I guess.

(25-04-02 10:26:03) Move ToT in charge tab of mon12

Because the timeOverThreshold is related to the charge, i.e adcMax or integral. Run 21042

(25-04-02 09:09:03) Connection to ifarm changed

Connection to the ifarm changed. Now, you are using a new pin for the double authentification. Guess what ?

DATE → 25-04-01

Assumption of the starting point to plot 3D figure

Starting point to plot 3D figure

DATE → 25-03-31

(25-03-31 11:39:33) Update of waveform display

Specific view in AHDC, Run 21242.00009 evt 2342.

(25-03-31 11:39:15) Update of waveform display

Now we use the decoding output from coatjava to display waveform in the event display of AMON

DATE → 25-03-30

(25-03-30 17:36:30) May be a cosmic

Run 21242.00009, evt 21983 (if nb. starts at 1)

(25-03-30 08:33:10) LeadingEdgeTime in run 21242

AHDC::adc --> leadingEdgeTime

(25-03-30 08:31:24) First channel time in run 21242

AHDC::wf --> time

(25-03-30 08:28:55) File decoded

The file are decoded and available in local (run 21236, 21241, 21242)

(25-03-30 07:41:05) jcache run 21236, 21241, 21242

These runs remain to be cp and decoded. Goal : test the new ModeAHDC and the first channel time.

DATE → 25-03-29

(25-03-29 12:01:22) ModeAHDC functional

I loooked at the run 21110 with ZS=1 && ZStyp=0. The run is small but we can see that the first channel time distribution is always 0.

(25-03-29 11:44:32) ModeAHDC functional

Simple note. Run 21162 is ZS=1 && ZStyp=1

(25-03-29 11:37:50) ModeAHDC fonctionnal

Here is plot of the AHDC:::wf time which store the first channel time of the AHDC pulse.

(25-03-29 05:41:36) ModeAHDC updated

Here is the associated leadingEdgeTime

(25-03-29 05:34:46) ModeAHDC updated

All time quantity are perfectly translated by the first channel time. Here is a distribution of the timeMax.

(25-03-29 05:09:33) firstChannel issue FIXED

As suspected, the time was overwritten. We were still using the MVTFitter. Now let's do a pull request.

(25-03-29 04:50:45) firstChannel issue

Here are the link fitPulses and dataList.

(25-03-29 04:48:11) firstChannel issue

I think I just find the error. What we call firstChannel seems to be modified in the method fitPulses() of DetectorEventDecoder. It is strange but we are still using the MVTFitter. To correct that I propose to comment the relevant bloc of code. Line 171

(25-03-29 04:09:37) Time from Bonus12

At the stage of the CodaEventDecoder, the firstChannel has reasonable values.

(25-03-29 03:50:55) Time from Bonus12

Here is the time (the first channel from DREAM manual) distribution. It is very strange, look like an ADC distribution. Probably not, I suspect this time to be modifed before reaching this stage. So I will debug like in old times, i.e do some println()...

Log Book