search

JeffersonLab / sim-recon

Simulation and Reconstruction for GlueX
9 stars 14 forks source link

copy DCDCHit factory into new factory with Tag "Calib". remove the QF… #1079

Closed zihlmann closed 6 years ago

zihlmann commented 6 years ago

… bit timing cut

and the drift time cut from this Taged factory and move these two cuts into the new DCDCHit factory. Add a pruing algorithm to the DCDCHit factory that removes hits from the list that are correlated with a hit that has overflow on the same connector.

zihlmann commented 6 years ago

This change follows a similar approach as is done in the TAGGER hit factory. cuts to remove hits from the CDC hit list are applied after the main hit list is created using the tag "Calib". In this way expert code can access the CDC hits before cuting some away depending on the timing quality, timing and correlation with saturation hits withing the same connector/hvboard/preamp. The QF bits are added to the DCDCHit container.

gluex commented 6 years ago

Test status for this pull request: SUCCESS

Summary: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/tests/summary.txt Logs: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/tests/log

Build log: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/make_CDCHitFactoryUpdate.log Build report: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/report_CDCHitFactoryUpdate.txt Location of build: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate

sdobbs commented 6 years ago

A couple quick comments before merging:

1) Can we make all of these large arrays vectors instead? 2) Can the 3.5 ns timing cut for correlated hits be a parameter as well?

nsjarvis commented 6 years ago

Hey Beni,

Are you sure that the cut should start at exactly 3.5 samples? I thought it looked more like 2 to 3 samples.

While you're doing this, you might want to add some code to cut out the impossible events where many channels on the same hvb have hits in the same time-sample.

Naomi.

On 11 March 2018 at 15:39, Sean Dobbs notifications@github.com wrote:

A couple quick comments before merging:

  1. Can we make all of these large arrays vectors instead?
  2. Can the 3.5 ns timing cut for correlated hits be a parameter as well?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/1079#issuecomment-372142715, or mute the thread https://github.com/notifications/unsubscribe-auth/AO7hVu83vJWtDQUfk9umb14QqUpbVBTPks5tdX2GgaJpZM4Slx_P .

rjones30 commented 6 years ago

Hello Beni,

I just updated my sim-recon, and then tried running the evio-hddm plugin to generate random events for mixing with MC to simulate background. I just want to make sure that you are not cutting out those early hits in the random events, the very ones we are trying to inject into the Monte Carlo to test for the rate-dependence from early hits shadowing true hits in the event. I see new messages showing up in the output from JANA when I run now with the evio-hddm plugin, that I never used to see in the past.

-Richard Jones

On Sun, Mar 11, 2018 at 10:48 PM, nsjarvis notifications@github.com wrote:

Hey Beni,

Are you sure that the cut should start at exactly 3.5 samples? I thought it looked more like 2 to 3 samples.

While you're doing this, you might want to add some code to cut out the impossible events where many channels on the same hvb have hits in the same time-sample.

Naomi.

On 11 March 2018 at 15:39, Sean Dobbs notifications@github.com wrote:

A couple quick comments before merging:

  1. Can we make all of these large arrays vectors instead?
  2. Can the 3.5 ns timing cut for correlated hits be a parameter as well?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/ 1079#issuecomment-372142715, or mute the thread https://github.com/notifications/unsubscribe-auth/ AO7hVu83vJWtDQUfk9umb14QqUpbVBTPks5tdX2GgaJpZM4Slx_P .

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/1079#issuecomment-372179491, or mute the thread https://github.com/notifications/unsubscribe-auth/AHeFWKygF7fBILPoGM-lFIYoa-PWnKy5ks5tdeHmgaJpZM4Slx_P .

nsjarvis commented 6 years ago

Hi,

It is to eliminate the repeat pulses caused by the electronics. They occur approx 24-32ns after an earlier large pulse.

I did something similar to produce https://logbooks.jlab.org/entry/3544091

Naomi.

On 11 March 2018 at 23:11, Richard Jones notifications@github.com wrote:

Hello Beni,

I just updated my sim-recon, and then tried running the evio-hddm plugin to generate random events for mixing with MC to simulate background. I just want to make sure that you are not cutting out those early hits in the random events, the very ones we are trying to inject into the Monte Carlo to test for the rate-dependence from early hits shadowing true hits in the event. I see new messages showing up in the output from JANA when I run now with the evio-hddm plugin, that I never used to see in the past.

-Richard Jones

On Sun, Mar 11, 2018 at 10:48 PM, nsjarvis notifications@github.com wrote:

Hey Beni,

Are you sure that the cut should start at exactly 3.5 samples? I thought it looked more like 2 to 3 samples.

While you're doing this, you might want to add some code to cut out the impossible events where many channels on the same hvb have hits in the same time-sample.

Naomi.

On 11 March 2018 at 15:39, Sean Dobbs notifications@github.com wrote:

A couple quick comments before merging:

  1. Can we make all of these large arrays vectors instead?
  2. Can the 3.5 ns timing cut for correlated hits be a parameter as well?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/ 1079#issuecomment-372142715, or mute the thread https://github.com/notifications/unsubscribe-auth/ AO7hVu83vJWtDQUfk9umb14QqUpbVBTPks5tdX2GgaJpZM4Slx_P .

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/ 1079#issuecomment-372179491, or mute the thread https://github.com/notifications/unsubscribe-auth/AHeFWKygF7fBILPoGM- lFIYoa-PWnKy5ks5tdeHmgaJpZM4Slx_P .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/1079#issuecomment-372182399, or mute the thread https://github.com/notifications/unsubscribe-auth/AO7hVlHZZdvKhM75utDxUn3vb60suaabks5tdedNgaJpZM4Slx_P .

zihlmann commented 6 years ago

Hi Sean, in principle yes to both questions. to 1) do you say that the push_back() function is as fast as just asking for a junk of memory. to 2) it is not a 3.5ns cut it is cut of 1.58ns cut around t=-3.58ns

zihlmann commented 6 years ago

Hi Richard, yes they are cut. the early hits seem not to be an issue. if you do not want them to be cut you can change the ccdb table TOF/timing_cuts which has 2 parameters -60 and 900. representing the cut of the timing window.

zihlmann commented 6 years ago

Hi Naomi, the correlation peak is aruond -3.5 cdc_correlationpeak.pdf bins.

sdobbs commented 6 years ago

Richard,

We do want to reject these hits from the reconstruction, so The HDDM output should probably be pulling CDC hits from the Calib tag. Actually we might want to have an even lower level tag for the CDC and FDC where no quality cuts are applied...

On Mon, Mar 12, 2018 at 8:08 AM zihlmann notifications@github.com wrote:

Hi Naomi, the correlation peak is aruond -3.5 cdc_correlationpeak.pdf https://github.com/JeffersonLab/sim-recon/files/1802644/cdc_correlationpeak.pdf bins.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/1079#issuecomment-372288188, or mute the thread https://github.com/notifications/unsubscribe-auth/ABIJag7J8hg0tJb4w1ofOS3KoKHiBY61ks5tdmU3gaJpZM4Slx_P .

rjones30 commented 6 years ago

Sean,

Yes, no QF cut. We want a level of hits that includes anything that can saturate the fadc hit readout on that channel, and cause it to drop any subsequent hits. How would you like to do this?

-Richard

On Mon, Mar 12, 2018 at 8:40 AM, Sean Dobbs notifications@github.com wrote:

Richard,

We do want to reject these hits from the reconstruction, so The HDDM output should probably be pulling CDC hits from the Calib tag. Actually we might want to have an even lower level tag for the CDC and FDC where no quality cuts are applied...

On Mon, Mar 12, 2018 at 8:08 AM zihlmann notifications@github.com wrote:

Hi Naomi, the correlation peak is aruond -3.5 cdc_correlationpeak.pdf https://github.com/JeffersonLab/sim-recon/files/ 1802644/cdc_correlationpeak.pdf bins.

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/ 1079#issuecomment-372288188, or mute the thread https://github.com/notifications/unsubscribe-auth/ ABIJag7J8hg0tJb4w1ofOS3KoKHiBY61ks5tdmU3gaJpZM4Slx_P .

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/JeffersonLab/sim-recon/pull/1079#issuecomment-372296129, or mute the thread https://github.com/notifications/unsubscribe-auth/AHeFWI5GToC12EQGLSzW16r5YohY25M8ks5tdmzZgaJpZM4Slx_P .

gluex commented 6 years ago

Test status for this pull request: SUCCESS

Summary: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/tests/summary.txt Logs: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/tests/log

Build log: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/make_CDCHitFactoryUpdate.log Build report: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/report_CDCHitFactoryUpdate.txt Location of build: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate

gluex commented 6 years ago

Test status for this pull request: SUCCESS

Summary: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/tests/summary.txt Logs: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/tests/log

Build log: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/make_CDCHitFactoryUpdate.log Build report: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate/report_CDCHitFactoryUpdate.txt Location of build: /work/halld/pull_request_test/sim-recon^CDCHitFactoryUpdate