Shenzhen Dongweixin Technology Co., Ltd
Address: 2nd Floor, Building 1, No. 3 Xintang Road, Xintian Community, Fuhai Street, Bao'an District, Shenzhen
Service Hotline: 0755-33160359
Accessories Sales Hotline:+86-15914016961 Miss Xian
Email: dwxsmt@126.com
Website address: www.dwxyamaha.com
Bank of Deposit: Industrial and Commercial Bank of China Shenzhen Baomin Branch
Company account: 400009253910083508
Account Name: Shenzhen Dongweixin Technology Co., Ltd
|
|
 Your Location: Home - View news |
|
| The Ultimate Challenge of AI Computing Power Hardware: SIPLACE SX Redefines the Boundaries of High-End SMT Processing |
| Addtime：2026/8/16 10:43:34 Clidk：320 |
The ongoing wave of generative AI continues to drive the iteration of computing power infrastructure. AI server motherboards, GPU accelerator cards, and high-speed interconnection backplanes are increasingly characterized by a wider weight span of components, growing PCB dimensions, and continuously escalating assembly reliability standards. The coexistence of giant BGA computing chips, high-density micro passive components, and irregular-shaped power devices on the same production floor poses a series of process challenges for traditional placement equipment, including uncontrolled placement pressure, large-board warpage, and the difficulty of balancing high speed with precision. Against this backdrop, the ASMPT SIPLACE SX modular placement platform—spanning hardware architecture, motion control, vision inspection, and production line digitalization—has redrawn the upper limits of SMT manufacturing processes for AI computing hardware.
The most prominent manufacturing challenge of AI computing motherboards is the mixed placement of ultra-lightweight micro components and heavy-duty chips weighing hundreds of grams. A single high-end computing board simultaneously contains 0201 micro components and large GPU package modules weighing close to 650g. The SIPLACE SX achieves compatible production through its modular placement head combination solution: the CP20 multi-head module is responsible for high-speed placement of high-density small components, ensuring baseline production capacity; while the TWIN VHF heavy-duty placement head, equipped with a dynamic pressure closed-loop control system, can adjust placement force in real time, preventing solder ball collapse or chip cracking on expensive computing chips due to excessive pressure. The built-in dynamic motion optimization algorithm optimizes acceleration curves for large-mass components, mitigating inertial deviation caused by high-speed movement and ensuring long-term stability of placement accuracy.
Warpage deformation of extra-large thick boards is another major pain point in computing board production. The high deadweight of long-size PCBs and internal stress within multi-layer boards can easily cause board bending, directly leading to BGA cold solder joints and poor contact. The SIPLACE SX features an upgraded heavy-duty transport track, paired with an adaptive intelligent support pin system that automatically plans support points based on PCB layout, dynamically compensating for board deformation. The optional 3D laser coplanarity inspection unit scans the solder ball contours of chips prior to placement, weeding out defective devices in advance and preventing batch quality incidents from the outset, meeting the zero-defect manufacturing requirements of server hardware.
The rapid iteration pace of computing hardware products demands frequent line changeovers, while also accommodating R&D trial production and mass production simultaneously. The SIPLACE SX's native modular architecture supports flexible addition or removal of gantry units, allowing capacity to be scaled on demand. The accompanying uninterrupted waffle tray feeder solution enables non-stop replenishment of tray-packaged AI chips, effectively reducing idle-time losses. Leveraging the SIPLACE WORKS intelligent manufacturing software, it supports rapid import of component coordinate data, shortening NPI debugging cycles and adapting to the fast-paced update rhythm of the computing hardware industry.
End-to-end digitalization capability meets the full traceability requirements of high-end computing hardware. The equipment natively supports HERMES and CFX industry communication standards, enabling seamless integration with MES, AOI, and intelligent warehousing systems. Placement coordinates, placement pressure, vision images, and material information are all stored in real time, creating a complete process record for each individual board, satisfying the stringent quality traceability standards of data center and high-end server applications.
Industry analysts believe that future AI computing hardware will continue to move toward higher integration, larger form factors, and heterogeneous packaging. SMT processing is no longer simply about component placement, but rather a complex system engineering integrating precision control, mechanical management, and digital traceability. Through its four core capabilities—heavy-load placement, deformation compensation, flexible modularity, and full-process data control—the SIPLACE SX removes the bottlenecks in integrated production of micro components and heavy computing chips, helping high-end EMS manufacturers break through the barriers of computing hardware manufacturing, and consistently supporting the large-scale construction of global AI computing power infrastructure.
|
| |
| |
|
|
|
