The CMS collaboration is building a new outer silicon tracker for the High-Luminosity Large Hadron Collider (HL-LHC). The tracker must survive a harsher radiation environment and handle many more simultaneous collisions. Part of the upgrade are ‘‘2S’’ modules. These modules pair two strip sensors to decide on the fly if a particle has enough sideways momentum — transverse momentum, or pT — to be interesting. Using that cut, the detector can discard about 90% of low-pT signals and send much less data to the Level-1 (L1) trigger, the fast hardware that first selects events to keep.

A 2S module is a compact assembly. It has two identical silicon strip sensors, electronics boards called front-end hybrids with CMS Binary Chips, a data concentrator chip, and a service hybrid that carries the optical link and power converters. Different module variants exist: sensor spacings of 1.8 mm or 4 mm, and designs with five or six cooling points. The tracker layout calls for thousands of modules: for example, one partition (TB2S) will use 4,416 2S modules. Production is distributed across 12 centers in Asia, Europe and North America.

Because these modules must work reliably in the detector, production includes strict quality steps. Each finished module goes through a set of tests and a so‑called Burn‑in: 16 to 18 thermal cycles between room temperature (20 °C) and the cold operating temperature (−35 °C). Electronic readout tests at both temperatures check noise and stability; cold noise is typically lower. A central software tool (POTATO) grades test outputs and local centers keep detailed records, for example with a database called GABRIEL at the Brussels site.

The paper focuses on practical problems found during assembly and how teams fixed them. Three representative issues are described. First, scratches from probes or jigs can cut sensors and make them leak current. If leakage stays below a stated limit when the module is biased at high voltage (800 V), the part can still be used; in some cases removing the bond wires of noisy strips improved the sensor behavior. Second, electric sparks occurred when conductive edges on a service hybrid were too close to the high‑voltage sensor edge. The production fix was to add a Kapton insulating layer between the sensor and the hybrid. Third, unseen conductive contamination on jigs or carriers caused early breakdowns in the high‑voltage tests. Teams treated these by using an ion blower, by applying a controlled high‑voltage bias at low humidity with a current limit, or by carefully wiping edges with electrostatic‑safe tissue and ethanol.