More FDC issues?

[T-Britton]

This sample included all of Richards flags including:
DCAY 0
LOSS 2
MULS 0
DRAY 0

protons at 4 GeV and 15 deg. look equivalent but at 4 degrees:

HDGeant4 missing hits on track

very different wirebased trackingFOM

without multiple scattering we expect the tracking to overestimate the errors, thus causing a pile-up at a CL at 1, as it does with hdgeant. hdgeant4 does not.

You can get access to all of the reconstruction plots via the sept 11 meeting page, many differences to be seen

[zihlmann]

At the wirebased fit stage there is a tendency to loose the more downstream hits mostly of package 4.
However at the time based stage some of the hits are recovered in the Geant3 version but not in
the Geant4 version. This indicates that the FDC timing is not the same in geant3 and geant4.
This clearly shows that there is an issue with the wirebased tracking algorithm.
the plot below shows the wirebased fit result and the hits/pseudohits that were used.
Note that all hits in package 4 and most of package 3 are not used anymore. They were at
the candidate stage.

The second issue with the geant4 simulation is that quite frequently the event has no FDC hits
at all!!!!

[rjones30]

I traced this issue down to an inconsistency in the way fdc chamber layers were being indexed between the g3 and g4 hits libraries. My tests show that the track reconstruction efficiency for hdgeant4 simulation output is comparable to that from hdgeant, once this fix is in place. Experts should confirm this, but I believe this issue can be closed.
-Richard Jones

[T-Britton]

Great news! I’ll run 2 sets to compare with master branches this week.

[jrstevenjlab]

Thanks Richard! Just to be clear, is this fix already on the current master? -Justin

…

On Jan 1, 2019, at 11:52 AM, Richard Jones ***@***.***> wrote: I traced this issue down to an inconsistency in the way fdc chamber layers were being indexed between the g3 and g4 hits libraries. My tests show that the track reconstruction efficiency for hdgeant4 simulation output is comparable to that from hdgeant, once this fix is in place. Experts should confirm this, but I believe this issue can be closed. -Richard Jones — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_JeffersonLab_HDGeant4_issues_66-23issuecomment-2D450742906&d=DwMFaQ&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=cIyw2PFFtpuc0ZvlKFxF6UnQ9C9dChESxxwL_XrGKB8&m=wigQSZp6HRsf5-XBXWkvkgBqJSawPxr2fT3hcs414iA&s=s-2-vsodxRh_bL4b10qFCagM4wKovtX7xRyBi30VyBg&e=>, or mute the thread <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AMoSBspGVuAf6ezmniavyszqdNsR5NqVks5u-2D5I5gaJpZM4Wj9g2&d=DwMFaQ&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=cIyw2PFFtpuc0ZvlKFxF6UnQ9C9dChESxxwL_XrGKB8&m=wigQSZp6HRsf5-XBXWkvkgBqJSawPxr2fT3hcs414iA&s=zAaAiR3yy8JxDC7U1q7xv_GfNpiGYyFWmqtvK2_mBbE&e=>.

[rjones30]

Hello Justin, Yes, it is there. There are a couple of pull requests pending that are needed to make it work with G4.10.04, but with G4.10.02 libraries the present master contains the relevant fixes.

…

-Richard On Tue, Jan 1, 2019 at 12:01 PM Justin Stevens <notifications@github.com> wrote:

Thanks Richard! Just to be clear, is this fix already on the current master? -Justin > On Jan 1, 2019, at 11:52 AM, Richard Jones ***@***.***> wrote: > > I traced this issue down to an inconsistency in the way fdc chamber layers were being indexed between the g3 and g4 hits libraries. My tests show that the track reconstruction efficiency for hdgeant4 simulation output is comparable to that from hdgeant, once this fix is in place. Experts should confirm this, but I believe this issue can be closed. > -Richard Jones > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub < https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_JeffersonLab_HDGeant4_issues_66-23issuecomment-2D450742906&d=DwMFaQ&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=cIyw2PFFtpuc0ZvlKFxF6UnQ9C9dChESxxwL_XrGKB8&m=wigQSZp6HRsf5-XBXWkvkgBqJSawPxr2fT3hcs414iA&s=s-2-vsodxRh_bL4b10qFCagM4wKovtX7xRyBi30VyBg&e=>, or mute the thread < https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AMoSBspGVuAf6ezmniavyszqdNsR5NqVks5u-2D5I5gaJpZM4Wj9g2&d=DwMFaQ&c=lz9TcOasaINaaC3U7FbMev2lsutwpI4--09aP8Lu18s&r=cIyw2PFFtpuc0ZvlKFxF6UnQ9C9dChESxxwL_XrGKB8&m=wigQSZp6HRsf5-XBXWkvkgBqJSawPxr2fT3hcs414iA&s=zAaAiR3yy8JxDC7U1q7xv_GfNpiGYyFWmqtvK2_mBbE&e= >. > — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#66 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHeFWGzpjVS8Jc_7Dm4yFz4V-257U8q9ks5u-5RtgaJpZM4Wj9g2> .

[T-Britton]

A quick check of the above plots show them to be identical (within statistics). This issue is closed!