Minutes (by Mary Bishai, BNL):
Requirements: Ettore
Marzio comments on the light collection requirements being different based on HD vs VD differences in active masses. Bo, Mary and Alberto M. comment that the HD active mass is greater than 10kton. Alberto notes the difference between HD and VD active volume is really only around 10%. Ettore notes the difference is not a big issue for the design of the photon detector.
VD PD design: Flavio
Alberto: What is the power budget and how does it compare to the heat loss through the cryostat?
Answer: I have assigned a few kW power budget - around 5kW maximum
Carla: Requirement on xArapuca tile, the present tile is about 8x the surface of the previous one - how close we are with the megacell to the limit of the power budget and is there room to split into two cells wrt to the power budget?
Answer: We can - in particular if we adopt the analog CE transmision - also if we go for two channels instead of 1 with the digital. This has to be looked at more carefully. The supercell mentioned here from the electronics readout is already split in two but not optically. The power needed for the SiPM is really a minor change in the power budget compared to the electronic readout. The number of SiPM has been increased in the megacells compared to the HD. It can be further increased that implies cost increase but only minor power increase.
Jae: Question on ordered options for VD PDS slide - these three options could have significant impact on the HV system. The last option doesn’t include PD mounted on the cathode? When is the right timing to make the decision.
Answer: Last option (membrane mounted) is the lowest risk but may not meet minimum requirements and will need modification to FC
Kettell: I am a bit of concerned about the strategy as shown here and it is possible we cannot prove our optical readout R&D solutions by CD1 review. We need a backup option that could meet minimal requirements that would meet CD1 review if the R&D hasn’t panned out.
Simulations: Laura
Marzio: For the cathode only solution you are violating the minimum requirements Ettore showed.
Answer: Yes for Ar only but it passes requirements for Ar+Xe - if we use the requirements for the HD unchanged and not come up with separate requirements
Mary: Slide9 : the minimum LY covers only a small volume, so the effective active volume would be slightly less but you have the higher LY everywhere?
Answer: The LY of 11 photons/MeV is for Ar+Xe and doesn’t meet the requirement from the HD - it will be even less with Ar only. We expect at least 40% increase in LY with Xe doping
Alberto: Slide9 - What is the dependency on the transparency?
Answer: Even with 100% transparency you will still not meet the 20photons/Mev requirement. Even with 20% of coverage. This one is in the range of 14%
Jae: This simulation has reflection on anode and cathode?
Answer: It has 25% reflection on the anode side
Marzio [couldn’t hear comments]
Response to Marzio from Laura:
Marzio, the difference in the number of PDs over the cathode and the laterals comes from 2 facts: 1. the cathode arapucas sees both the upper and lower volumes. In the simulation we presented, we were looking at the upper volume alone. When including the lower volume, the number of PDs over the laterals have to be doubled whereas the number of PDs over the cathode remains the same; 2. The extra PDs over the laterals in the calculation (2*320) vs. 768 comes from the two smaller sides also being instrumented with PDs.
Development of analog optical: Sabrina
Francesco: It is true that the system of Darkside meets the specs of DUNE except for the dynamic range - but keep in mind that we cannot get the system of the shelf - they only developed a technological solution that works for a handful of shelves so a large scale system doesn’t exist yet.
Marzio: What is known about the lifetime of these component in the cold - Darkside can replace their systems - we cant?
Answer: There is not any information - most of them are not made for cold so we have to test them
Dario: This is the same for all components working for cold - including the digital option.
Kettell: If you would be very optimistic about R&D what is the timeline would you expect?
Answer: We are aiming for a working device in time for the cold box test in the short time scale… several tests are proceeding over the next few weeks but we don’t have a timetable yet.
Flavio: I should mention longevity tests were not required for the 2020 tests. Not included in the plans for the present activity that is why there is no info
Phillipe: What are the requirements on the linearity of the transmission? What is the goal for the difference between LED linearities?
Answer: Linear transmitter for the full range is the goal - we are also thinking of studying a non-linear transmitter that covers the full dynamic range. Darkside aim for perfect linearity within the range in their usage so they don’t have to calibrate from LED transmitter to the other. Now focusing mostly on analog transmission and it may be that the 14 bit is not what we end up with for the digitizer - this was inherited from the HD.
Power fiber concept for Analog: Bill
Kettell: You assume the arapuca and transmission are separate boxes?
Answer: Yes
Kettell: How much of this have you tested in the cold?
Answer: The entire chain on the left hand side of slide 7 has been tested in the cold. Readout chain on the right hand side has mostly not been tested in the cold
David: Slide 4: How many fibers per transmitter/Arapuca?
Answer: It doesn’t change power consumption that much with 2 fibers per Arapuca vs 1 which is what we are designing for.
Flavio: The 4 arapuca modules you see on Slide 4 are the ones to be located into 1 cathode frame. These 4 will be grouped together into 1 electronics box with 1 fiber per arapuca.
Digital Cryo link option: Alan and ADCs: Dave C
Kettell: Dave can you say something about the lifetime of FPGAs for 20-30 yr expts?
Answer: BNL did some studies 10 years ago - the core logic is very likely to work well for many years. There has not been that sort of testing for the high speed digital I/O. We intend to do is use the FPGA as a data concentrator and then use a serial output to a laser driver - those links havent been tested for longevity in the cold.
Yasar: When do you expect cold tests for the QIE 14 to start?
Answer: This week we will start the cold tests.
Gustavo: Regarding the design of the FE of the ADCs - we have done extensive use of the active ganging - there is going to be an Op Amp that we have already tested in the cold that we will use for the active ganging that complements the use of the ADC. We don’t have a long history of using this for many years - but we do have some history of using them in several tests. Power consumption was already measured by Andre and the power consumption didn’t change that much. The FPGA was used in cryo CAPTAIN has been validated by other people in the Netherlands for quantum computing down to 40K - with some problems but that was fixed by adjusting some voltages. So there is some evidence that FPGA will survive for some time in the cold.
Digital power Arapuca: Bill
Kettell: Electronics box total dominated by 5V - something like a 100A at 5V?
Answer: Assumption is running all op amps of of 5V so 800W total.
Dave Newbold: For the timing thing you mention - the timing alignment requirements as per the DUNE requirements are quite demanding so timing takes up quite a lot of power and you don’t want to put that in Lar and you can save power by moving some of the timing to the top of the detector. It may be good to talk to the timing team.
Francesco: [Question about noise due to clocks?]
Answer: Ripple will probably cancel out if it is a 100 MHz clock for digitization since it is in phase with the readout.
We recently tested a phase-lock loop with the FPGAs with an external clock and showed it work.
VD Photon detector tile mechanical design: Dave
Kettell: This end of August date gives you two arapucas that are fully functional with wavelength shifting bars?
Answer: Yes - one with analog electronics for transmisssion and one with digital.
Kettell: Slide 3 can you explain the difference in efficiency of the number quoted of 3% and the numbers on this slide?
Answer: None of these are on 60x60 module not 10x40 - the idea is to point out is that we are not very far from our goal. We certainly havent demonstrated the efficiency on a module that large.
Carla: The 3.5% efficiency was performed by alpha scanning of the cell on 20cmx7.5cm wide tile - we will test soon 48cmx9.5cm cell. From simulation we expect the same yield and efficiency because the optical path is not changed. What could be a difference is that in the 60x60 the optical path changes significantly and you cannot put your SiPM a distance of 20cm. And so by properly sizing the concentration of the chromoform inside the wavelength shifting plate we hope to keep the same efficiency but this has to be proven.
Marzio: Does Xe change the simulation? What kind of optimization are you doing on these dichoric filters - are you optimizing for 135nm? If you can optimize you can get rid of the primary stage of wavelength shifting…so it seems the filter optimization on the long term path is the most important thing - you gain in efficiency you reduce SiPMs you gain on everything including cost and power and reliability…etc. Maybe you need a small group to help and push this forward.
Answer: Xe only changes the primary wavelength shifting stage, here we are only considering the secondary wavelength shifting stage. If we skip the primary stage we have to change the design of the dichroic filter, but we havent demonstrated that. I don’t have knowledge of the changes on the window with the dichroic filter. We are working on optimization of the filter for 135nm with the company developing the filters. For the cold box test we are still working with the primary. If there is someone who can push the design and technology it would help a lot.
Marzio: Does the efficiency scale with the number of SiPMs?
Answer: To first order yes. If you can double the efficiency from the dichroic filter stage that will have a big difference.
Marzio: Why not go to 60x30 instead of 60x60? If you use the smaller cell size you can keep the same number of SiPMs and therefore the same power but increase your efficiency.
Answer: We will test the collection of the 48cm long reading on the short side - then this will allow us to calibrate the MC simulation. The 60x60 option has been simulated in terms of efficiency and compared to smaller size based on the assumption in the simulation on the attenutation length - they show reduction of 20% in efficiency compared to 60x20. This needs to be confirmed with tests on the 48cm.
Carla: We are keeping in mind this option of two smaller cells reading them out as a single module which keeps the cost almost the same.
SiPM ganging board: Dante
[Had another meeting to go to - so missed the end of the meeting]
--------------------
Dr. Mary Bishai
Senior Physicist Office: 3-160
Brookhaven National Laboratory Tel: +1 631-344-4877