INTT Weekly meeting

Thursday 9 Oct 2025, 20:00 → 23:00 US/Eastern

Meeting URL 

https://riken-jp.zoom.us/j/92149923535?pwd=sbLcakxQeRQrch8dnJfJ8g1NiEdvKa.1

20:00 → 20:20 More about silicon seeding

Speaker: Cheng-Wei Shih (National Central University)

INTT_2025_10_10_SiSeed_InttTimeMM2.pdf

- "Timely matched" means 2 INTT clusters have the same timing buckets.
-- No timing from MVTX clusters.100 bunch crossing resolution for MVTX
-- Silicon seed start finding cluster matching geometrically within
500 timing buckets.
- Timebucket = INTT_bco_full + INTT_hit_bco - offset - GL1BCO
-- In the seeds there is no more INTT BCO info, so it's hard to check
the timing of INTT.
- Since this is streaming, we still have plateau region from real
collisions for the red histogram (2 INTT clusters have same time
bucket).
- The second spike indicates of carry-over hit events.
-- If there is no carry-over hit event, there should be no second
spike in the seeds with 2 INTT cluster timely mismatched.
- Spikes can be understood, but the cause of bumps around the spikes
is still unclear.
- We can check the calorimeter-GL1 event combined data to see if there
is still bump in the timing.
- BCO range is from negative 100 to positive 400, but why negative 400?
- One INTT hit shown up in multiple events because this framework is
driven by gl1.
-- Other experiments use the heart beat to synchronize events, which
means 2 adjacent events never overlap.
- You can check other time bucket because it could be behavior of
background of each bunches.
-- This tail is unique in this time bucket.
- Only seeds which has confirmed INTT timing match can be trusted.
- Lop pT soft particles should be removed for this analysis. A large
delta-phi region will be removed. pT>500MeV cut.
- By applying right cuts, you can still use the seed for the INTT
detection efficiency.
-- You can use EMCal cluster association, which is naturally selecting
hard components.
- The large fraction fake MVTX clusters can be filtered out by adding
many conditions.

20:20 → 20:40 Offline QA

Speaker: Yuko Sekiguchi (RIKEN)

INTTmeeting_20251009.pdf

- What is the difference between the offline calibration and offline
QA frameworks?
-- The offline QA need to use DST, but not CDB tree.
- The calibration files are produced from raw DST container.
-- We already BCO QA section, so we can grab histograms.
-- CDB tree is useful to estimate the fraction of cold/hot channels.
Yuko will discuss with Joseph for a different way to show it in
offline QA.
- >80% is so low and will miss a section drop issue.
-- >90% good, 80-90% questionable, <80% bad last year.
- Clustering is using one global INTT hot/cold map.
-- If they only use global map, how useful the phi distribution?
- What is the procedure of making the offline QA in a systematic way?
-- They are looking at all runs? How is the tag used if they are not
looking at all runs?
- Hit carry over events in the offline QA plot possible?

20:40 → 21:00 INTT Publication Plan

Speaker: Dr Itaru Nakagawa (RIKEN)

251010_PublicationPlan.pdf

- Author should be as inclusive as possible, everyone who contributed
the construction should be included.
-- All of the BNL techs who involved in the INTT work should be
included in the paper. Double check with Rachid.
-- Different subsystem members will not be on the paper. Just acknowledge them.
-- This is true even if we show MBD-INTT correlation plots.
-- Follow the TPOT NIM author list.
- We can go ahead to publish even before some of the beam
commissioning plots comes out.
- Half entry/chip saturation issues are hardware issue and should no
be mentioned in the NIM paper
- Cluster size should be avoided because we have an unkwon issue.
- S/N ratio can be estimated from the noise hit rate from abort gap.
- Dead hot channels as a function of phi is more easy to understand.
- INTT hits inner vs outer plot should update with full INTT version.
- For vertex, use INTT half detector method, not MBD correlation (optional).
-- Include available plots as much as possible.
- Details should be left for the editing committee.
- Crossing angle dependence of vertex is uncorrected.
- Zvertex with field off data is more preferable.
- Cosmic event display
- X-Y vertex reconstruction (No preliminary yet because misalignment
is significant)
- Don't rush like 1-2 month deadline. It is too short. Wait for
important plots to be preliminary.
-- Deadline in a half year is minimum.
-- Should we wait for the misalignment to be finalized? Check the
updated alignment.
- Efficiency is something new and it may not be fit in the publication schedule.