Energies of empty hodoscope counters are not truly zero

[frankboyu]

In GlueXPseudoDetectorTAG.cc, the hodoscope energy was determined as (as they should)

HODO_CHANNEL_EMIN[i] = hodo_parms[i]["xlow"] * endpoint_calib + endpoint_energy - endpoint_calib;
HODO_CHANNEL_EMAX[i] = hodo_parms[i]["xhigh"] * endpoint_calib + endpoint_energy - endpoint_calib;

This formula was used for all tagger counters, including hodoscope counter 128-178 that are not used in experiment. For these counters, their scaled_energy_range was set as zero in CCDB. But the Delta E = endpoint_energy - endpoint_calib term is still present, which ranges from 1-3 MeV for the 2021-11 dataset.

Then later there are two conditional statements that check the min and max energy as

while (energy < HODO_CHANNEL_EMIN[i] || HODO_CHANNEL_EMIN[i] < 1.)
++i;
while (energy > HODO_CHANNEL_EMAX[i] || HODO_CHANNEL_EMAX[i] < 1.)
--i;

Both of the latter ones (HODO_CHANNEL_EMIN[i] < 1. and HODO_CHANNEL_EMAX[i] < 1.) seem to be designed to bypass the empty hodoscope counters. But this would fail when Delta E is larger than 1 MeV.

We have seen the effect of this issue when analyzing the 2021-11 dataset. Half of the runs have Delta E = 3 MeV. As a result, photons in the low-energy hodoscope counters failed to be tagged.

We proposed to add an extra conditional statement to make the energies covered by those empty hodoscope counters to be truly zero

HODO_CHANNEL_EMIN[i] = hodo_parms[i]["xlow"] * endpoint_calib + endpoint_energy - endpoint_calib;
if (hodo_parms[i]["xlow"] == 0)
HODO_CHANNEL_EMIN[I] = 0;

HODO_CHANNEL_EMAX[i] = hodo_parms[i]["xhigh"] * endpoint_calib + endpoint_energy - endpoint_calib;
if (hodo_parms[i]["xhigh"] == 0)
HODO_CHANNEL_EMAX[I] = 0;

[rjones30]

Hello Bo, Yes, that 1 MeV check is wrong, there is a bug in this search in GlueXPseudoDetectorTAG.cc that I did not see until you pointed it out. Here is my proposed fix: ***@***.*** HDGeant4]$ git diff diff --git a/src/GlueXPseudoDetectorTAG.cc b/src/GlueXPseudoDetectorTAG.cc index 274b283..0c7bf4f 100644 --- a/src/GlueXPseudoDetectorTAG.cc +++ b/src/GlueXPseudoDetectorTAG.cc @@ -122,9 +122,9 @@ int GlueXPseudoDetectorTAG::addTaggerPhoton(const G4Event *event, if (energy < MICRO_LIMITS_ERANGE[0] && energy > MICRO_LIMITS_ERANGE[1]) { int i = MICRO_NCHANNELS * (energy - MICRO_LIMITS_ERANGE[0]) / (MICRO_LIMITS_ERANGE[1] - MICRO_LIMITS_ERANGE[0]); - while (energy < MICRO_CHANNEL_EMIN[i] || MICRO_CHANNEL_EMIN[i] < 1.) + while (energy < MICRO_CHANNEL_EMIN[i] || MICRO_CHANNEL_EMIN[i] == MICRO_CHANNEL_EMAX[i]) ++i; - while (energy > MICRO_CHANNEL_EMAX[i] || MICRO_CHANNEL_EMAX[i] < 1.) + while (energy > MICRO_CHANNEL_EMAX[i] || MICRO_CHANNEL_EMAX[i] == MICRO_CHANNEL_EMIN[i]) --i; if (energy >= MICRO_CHANNEL_EMIN[i] && energy <= MICRO_CHANNEL_EMAX[i]) { micro_energy = (MICRO_CHANNEL_EMIN[i] + MICRO_CHANNEL_EMAX[i]) / 2; @@ -134,9 +134,9 @@ int GlueXPseudoDetectorTAG::addTaggerPhoton(const G4Event *event, if (energy < HODO_LIMITS_ERANGE[0] && energy > HODO_LIMITS_ERANGE[1]) { int i = HODO_NCHANNELS * (energy - HODO_LIMITS_ERANGE[0]) / (HODO_LIMITS_ERANGE[1] - HODO_LIMITS_ERANGE[0]); - while (energy < HODO_CHANNEL_EMIN[i] || HODO_CHANNEL_EMIN[i] < 1.) + while (energy < HODO_CHANNEL_EMIN[i] || HODO_CHANNEL_EMIN[i] == HODO_CHANNEL_EMAX[i]) ++i; - while (energy > HODO_CHANNEL_EMAX[i] || HODO_CHANNEL_EMAX[i] < 1.) + while (energy > HODO_CHANNEL_EMAX[i] || HODO_CHANNEL_EMAX[i] == HODO_CHANNEL_EMIN[i]) --i; if (energy >= HODO_CHANNEL_EMIN[i] && energy <= HODO_CHANNEL_EMAX[i]) { hodo_energy = (HODO_CHANNEL_EMIN[i] + HODO_CHANNEL_EMAX[i]) / 2;

…

-Richard Jones

On Fri, Jun 19, 2026 at 1:23 PM Bo Yu ***@***.***> wrote: *frankboyu* created an issue (JeffersonLab/HDGeant4#238) <#238> In GlueXPseudoDetectorTAG.cc <https://github.com/JeffersonLab/HDGeant4/blob/master/src/GlueXPseudoDetectorTAG.cc#L98>, the hodoscope energy was determined as (as they should) HODO_CHANNEL_EMIN[i] = hodo_parms[i]["xlow"] * endpoint_calib + endpoint_energy - endpoint_calib; HODO_CHANNEL_EMAX[i] = hodo_parms[i]["xhigh"] * endpoint_calib + endpoint_energy - endpoint_calib; This formula was used for all tagger counters, including hodoscope counter 128-178 that are not used in experiment. For these counters, their scaled_energy_range was set as zero in CCDB. But the Delta E = endpoint_energy - endpoint_calib term is still present, which ranges from 1-3 MeV for the 2021-11 dataset. Then later there are two conditional statements <https://github.com/JeffersonLab/HDGeant4/blob/59ae979d7c2a81c3f43596b7b360a271486ef87c/src/GlueXPseudoDetectorTAG.cc#L137> that check the min and max energy as while (energy < HODO_CHANNEL_EMIN[i] || HODO_CHANNEL_EMIN[i] < 1.) ++i; while (energy > HODO_CHANNEL_EMAX[i] || HODO_CHANNEL_EMAX[i] < 1.) --i; Both of the latter ones (HODO_CHANNEL_EMIN[i] < 1. and HODO_CHANNEL_EMAX[i] < 1.) seem to be designed to bypass the empty hodoscope counters. But this would fail when Delta E is larger than 1 MeV. We have seen the effect of this issue when analyzing the 2021-11 dataset. Half of the runs have Delta E = 3 MeV. As a result, photons in the low-energy hodoscope counters failed to be tagged. We proposed to add an extra conditional statement to make the energies covered by those empty hodoscope counters to be truly zero HODO_CHANNEL_EMIN[i] = hodo_parms[i]["xlow"] * endpoint_calib + endpoint_energy - endpoint_calib; if (hodo_parms[i]["xlow"] == 0) HODO_CHANNEL_EMIN[I] = 0; HODO_CHANNEL_EMAX[i] = hodo_parms[i]["xhigh"] * endpoint_calib + endpoint_energy - endpoint_calib; if (hodo_parms[i]["xhigh"] == 0) HODO_CHANNEL_EMAX[I] = 0; — Reply to this email directly, view it on GitHub <#238?email_source=notifications&email_token=AB3YKWE5JX6PHNNXOXRB63T5AVZHVA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBXGAZDIOBUGMZTJJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB3YKWEKMOGDZPKZBD526K35AVZHVAVCNFSNUABEKJSXA33TNF2G64TZHM4DEOJTGA4TQNR3JFZXG5LFHM2DOMBSGQ4DIMZTGSQXMAQ> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/AB3YKWGMUHY3IVCC2UIMB6D5AVZHVA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBXGAZDIOBUGMZTJJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2KUZTPN52GK4S7NFXXG> and Android <https://github.com/notifications/mobile/android/AB3YKWDLQO3IRCALJTDOPYL5AVZHVA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBXGAZDIOBUGMZTJJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2K4ZTPN52GK4S7MFXGI4TPNFSA>. Download it today! You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[frankboyu]

Hi Prof. Jones,

Thanks for the quick comment! I think your approach is definitely more elegant and permanent to solve this issue.

I do have one small concern regarding testing the equality of two double valued numbers via HODO_CHANNEL_EMAX[i] ==
HODO_CHANNEL_EMIN[I]. Maybe we can try something like (HODO_CHANNEL_EMAX[i] - HODO_CHANNEL_EMIN[I]) < 0.01 ?

I know it's not ideal to introduce another small number for comparison. But the reason I have this concern is these values did change by a tiny amount during running. For example, I printed these edges for some counters and one of them changed from 8286.35 to 8286.3459148717.

I am not sure about the reason. Maybe because of the multi-threaded running or something else. But I guess this is hard to avoid as the values of these arrays are determined at run time and therefore can't be declared as const.