Choose timezone
Your profile timezone:
pCT Machine Learning Monday Meeting
→
Europe/Berlin
Ralf Keidel
(Fachhochschule Worms (DE))
Contact
-
-
11:30
→
12:45
pCT Machine Learning Meeting
Convener: Ralf Keidel (Fachhochschule Worms (DE))-
11:30
Speakers: Alexander Schilling (HS Worms), Denis Petrov
-
11:45
Speaker: Alexander Schilling (HS Worms)
- Is the blueprint in Gabor's repository up-to-date? (https://github.com/pgpapp/pcT-ReadOut/blob/main/CarbonCarrierWithCables_B.png)
- No one knows if it is up to date. (Moved to Thursday meeting)
- The blueprint states a height of 165.9 mm for the chip area. This does not make sense in any of my computations. The bottom part goes out to 83.15 mm, while the top ends at 83.05 mm, leading to a height of 166.2 mm. Is it just a mistake in the blueprint or am I missing something?
- Seems that no one knows if the code from Gabor or the drawing from Ton is correct. (Moved to Thursday meeting)
- In the blueprint, the vertical position is asymmetric. When rotating the second half layer, there is an offset of 0.1mm, increasing the total height to 166.3 mm. Is this how it is interpreted or is the vertical offset from the center not being rotated with the second half layer?
- Sounds good, think that the single alpide strings are shifted downwards (behind/in front of the cable)
- What readout rate should we assume for conversion (5µs)?
- Not sure about it, from a practical standpoint 10µs, because 5µs would oversample our readout.
- How to combine multiple readouts in the same layer and what are the foundational theories for this? Helge simply sums the edep and averages the position, is this fine for now?
- Tricky part, don't know it exactly, he summed it up. -> Can work with the simple estimation
- When a readout happens at a point where there is overlap, do we simply duplicate the edep to both chips? This would magically create energy, but it may be negligible compared to what is lost in the aluminum absorber.
- With a higher amount of energy it is okay. -> We should work on a better simulation setup.
- Viljar/Helge: we could lose half of our clusters on the border region, therefore we could have a problem there too.
- Think it would be best to ignore everything regarding the border regions for the statistics, because the cluster could show us not the full energy.
- What happens to hits in areas not directly covered by pixels? Should we still create a cluster and activate the pixels that are reached by that cluster or simply ignore the hit?
- For now we ignore it and duplicate the hits. Later if the simulation is more specific we can talk about it again.
- (Denis) Asks for plans to rewrite the simulation -> We need to talk with Dieter about that.
- There is technically an air gap between front and back chip layers. If the particle comes in at a certain angle, it may hit only one of the chips despite the position being in an overlapping area. Should this be accounted for or do we ignore the air gap and assume the chips to all be in the same z-position?
- Same as above, ignore for now, wait for more accurate simulation setup
Conclusion: We need to clarify the measurements in order to create a more accurate simulation setup and then most of the other questions are clarified
- Is the blueprint in Gabor's repository up-to-date? (https://github.com/pgpapp/pcT-ReadOut/blob/main/CarbonCarrierWithCables_B.png)
- 12:00
-
12:10
Speaker: Pierluigi Piersimoni (University of Bergen (NO))
http://ionimaging.org/llu2021-overview/
Event: "The 7th Annual Loma Linda Workshop", August 2-4, 2021
Deadline: a week before the event (sic!)
General status update & Machine Learning stuff & Differential Programming (A combined talk with 3 presenters?)
Duration 30 - 45 Minutes per contribution
Informal discussions after the talk.
- 12:25
-
11:30
-
11:30
→
12:45
