Beam Telescopes at the DESY II Test Beam

Beam Telescopes at the DESY II Test Beam Adrian Herkert on behalf of the DESY II Test Beam crew BTTB13, 19 May 2025, Valencia | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 2 Introduction ●The DESY II TB facility provides three beam telescopes, one at each beam line ●Can be requested when applying for beam time (Note: Different telescope models at different beam lines) ●Tracking at the DESY TB: –Relatively low beam energies: Between 1 and 6 GeV ●Good track resolution requires telescopes with low material budget (i.e. thinned monolithic sensors) –Beam structure and particle rates: ●Mostly one electron ‘at a time’ with ... ●… integer multiples of ~ 1 μs in between ●Average rates up to few times 10,000 particles/s (depending on energy and beam line) See talk on DESY II TB: S. Ackermann, Mon., 14:10 | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 3 ●In operation since ca. 15 years ago ●Large global user base; available also at CERN, ELSA, and SLAC (or TRIUMF?) ●MIMOSA26: –Active area: ~ 2.0 cm x 1.0 cm –Pixel size: 18.4 μm x 18.4 μm –Rolling shutter readout with 115 μs period ●Track pointing resolution down to ~ 2 μm possible ●Two rolling shutter cycles read out per trigger event → long event time (and no hit time information) ●Synchronization with TLU via handshake, receives trigger + trigger ID, sends busy with length 115 μs to 230 μs Currently available beam telescopes (1/2) MIMOSA26-based EUDET-type beam telescopes (TB21, TB24) | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 4 Telescope operation and DUT integration ●Integration of user devices with EUDET-type telescopes is handled via Trigger Logic Unit (TLU) and EUDAQ2 ●To keep the EUDET-type flexibility for DUT integration and stay compatible with already existing integrations, it’s preferable to have new telescopes integrated again with / in: ●In short: Enables hardware synchronization of different DAQ systems via distribution of common trigger and / or clock ●See previous talk ●Modular software framework to control data taking with multiple DAQ systems ●Provides GUI, basic runcontrol functionality, online monitor, interface with analysis software (Corryvreckan), etc. AIDA(2020) TLU EUDAQ2 | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 5 ●In user operation since summer 2022 ●Integrated with AIDA2020 TLU and EUDAQ2 → Compatible with existing EUDET-type integrations ●ALPIDE: –Active area: ~ 3.0 cm x 1.4 cm –Pixel size: 29.24 μm x 26.88 μm –Here: Triggered readout of a short time interval O(1 μs) ●Track pointing resolution down to 3 μm possible ●TLU integration uses synchronous mode with trigger ID; busy signal with constant duration of 20 μs Currently available beam telescopes (2/2) Adenium: An ALPIDE-based prototype (TB22) Mager, M. (2016) ALPIDE, the Monolithic Active Pixel Sensor for the ALICE ITS upgrade. Liu, Y. et al. (2023) ADENIUM – A demonstrator for a next-generation beam telescope at DESY. | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 6 Performance of Adenium ●Extensively tested in user operation by now and very well received: –Easy transition coming from M26 telescope –Stable operation –Performance parameters given by use of ALPIDE: Efficiency > 0.99 for each layer → > 0.99^6 ≈ 0.94 total ~ 1 track per trigger event Measured track pointing resolution vs beam energy Correlation in comparison with MIMOSA26 ALPIDE M26 | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 7 An upgraded common beam telescope (1/3) ●A large global user base is relying on availability of EUDET-type telescopes at different TB facilities ●Issues regarding future operation of current M26-based EUDET-type telescopes: –Deprecated components (system will reach end-of-life at some point) –Difficulties in keeping up with increasing requirements in detector R&D ●Task in AIDAinnova (European funding project; successor of EUDET) aimed at developing an upgraded version of a common beam telescope compatible with existing EUDET-type DUT integrations –ALPIDE chosen as best available sensor option (production-grade thin MAPS with small pixels and fast readout architecture) –Adenium was first prototype ●Issues not regarding performance but maintainability ●Will be kept in operation for now but there will not be more copies ●Had to start afresh to develop production version Introduction | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 8 An upgraded common beam telescope (2/3) Towards the production version Each layer as standalone device; hosts its own SoC, runs its own operating system → expensive (& updates tedious) One central SoC SoC glued directly to readout board (main custom hardware component, 6x per telescope) → no flexibility SoM Design at USTC → no access to PCB design files and firmware Design by staff at DESY electronics department → full access, guaranteed long-term support Firmand software developed from scratch Firmand software implemented in Caribou Adenium Production version | Beam Telescopes at the DESY II Test Beam | Adrian Herkert | 19 May 2025 9 An upgraded common beam telescope (3/3) Design ALPIDE on chipboard with edge connector as provided by ALPICE collaboration Custom PCB: Essentially adapter from edge connector to DisplayPort Power connector 12V; telescope layers powered over DisplayPort Zynq UltraScale+ (SoM: ME-XU1-15EG2I-D12E; same as for Caribou 2.0; on commercial baseboard) Custom ‘hub’ PCB: Interfaces telescope layers and TLU with SoC (instead of CaR board) ●40 MHz clock and T0 from TLU for synchronization ●Triggered readout ●Busy as long as event buffers on ALPIDE or FPGA are full ●Current bottleneck for rate capability: CPU reading from FPGA buffers → Implementation of DMA ongoing Corresponds to average rate of ~ 10,000 events/s