Modular multi-TB/s readout for the CMS HGCAL Data Acquisition System

by Martim Rosado

Thursday Sep 17, 2026, 4:00 PM → 5:00 PM Europe/Berlin

Seminarraum EG + zoom (IPE)

The HL-LHC upgrade at CERN forces the CMS detector to replace its endcap calorimeters with the new High-Granularity Calorimeter (HGCAL). This new detector will measure, with unprecedented high-resolution, the position, timing, and energy of charged particles generated in the LHC collisions. Event data read out from the 6 million active channels of the HGCAL Front-end Electronics (FE) electronics are transmitted to the Back-end Electronics (BE) electronics through O(10^4) optical fibres with heterogeneous data loads. The HGCAL BE DAQ system, dimensioned to read the detector out at an average event rate of 750 kHz, interfaces these fibres and is implemented in 96 Serenity boards hosting VU13P FPGAs. The average 1.9 TB/s of processed data are then transmitted to CMS’s central DAQ.

This talk presents the HGCAL BE DAQ system, detailing the development of gateware circuitry, its integration in the Serenity EMP framework, and validation with test beams in CERN’s North Area (SPS beam). The elementary processing element of the HGCAL BE DAQ, the capture block, has been integrated in the EMP framework and tested in a Serenity board to ensure its compliance with the required average event readout rate and size. Furthermore, the novel implementation of asynchronous FIFOs present in the capture blocks using the URAM cells facilitated a 30% reduction of BRAM cells in the target VU13P device, freeing them for other uses. A readout circuit solution is also presented, allowing to load-balance the data transmission to the system's outputs in a configurable manner, minimising the overall detector’s deadtime. 

These developments culminated in the first full-scale prototype of the HGCAL BE DAQ, integrated in the EMP framework and able to comply with all input fibre data loads present in the HGCAL with a single gateware design. A modular approach allowed to also use the developed circuits into multiple testing and detector production systems.

---------------------------------------------------------------------------------------

Martim Rosado has concluded his Master’s degree in Electrical and Computer Engineering from University of Lisbon in 2022. Since 2023 he is an Electrical and Computer Engineering PhD student working in new the High-Granularity Calorimeter of the CMS detector. His work lead to key contributions to the development and deployment of the HGCAL Back-end Data Acquition system, to be implemented across 96 AMD UltraScale+ FPGAs. In addition to FPGA development, he has also strong interest in microelectronics and parallel computing.

 

Zoom: https://kit-lecture.zoom-x.de/j/63503980416?pwd=EuEsPxvoVDM1TC7mIhtIqKasU03bNL.1

martim_rosado_hgcal_readout.pdf