How Does Liquid to Air Cooling Work?
The nVent Liquid-to-Air (LTA) Sidecar Heat Rejection Unit (HRU) is designed to enable AI deployments without chilled water. This fully integrated liquid-to-air heat rejection system supports up to two racks of liquid-cooled IT equipment in the absence of a Facility Water System (FWS). For accelerated computing, nVent’s LTA solution plays a critical role in liquid cooling platforms, circulating liquid through racks in a closed loop, effectively dissipating heat from chips.
Features & Benefits:
- Hot Swap & Serviceability
- Minimizing Failures & Leaks
- Controller Features
- Enablement.
Liquid to Air Sidecar HRU Specifications
The nVent Liquid-to-Air (LTA) Sidecar Heat Rejection Unit (HRU) is a completely integrated liquid-to-air heat rejection system that enables up to two racks of liquid-cooled IT equipment when a Facility Water System (FWS) is not available. The LTA Sidecar comes in a 48U rack with footprint of 600mm (24 inches) wide by 1200mm (48 inches) deep.
The LTA Sidecar comes standard with 14 hot-swap fans in N+1 configuration, a reservoir pump unit (RPU) with N+N hot-swap pumps, a hot-swap controller, a hot-swap concurrently maintainable redundant filtration system, 6 x N+N redundant power supplies, integrated leak detection, leak-free hose connections, and an LED Light Path status panel. Active components such as fans, pumps, power supplies, and liquid connections for filling and draining the system are all located on the front side of the Sidecar HRU for easy access.
Support for GB200 NVL36 and GB200 NVL72 rack configurations with a single GB200 NVL36 rack supported by a single Sidecar HRU (1:1) and a single GB200 NVL72 rack supported by two Sidecar HRUs (2:1) in parallel
LTA Sidecar supports GB 200 NVL72 configuration in a 2:1
Liquid to Air Sidecar HRU Applications and Compatibility
Integrating liquid cooling into existing data centers can be achieved through specific configurations utilizing NVL36 and NVL72 racks. These configurations are designed as reference for the AI Chipsets. Using an LTA sidecar, you deploy AI Chips in existing data centers with minimal infrastructure swap outs.
Why Deploy Liquid Cooling without Facility Water
Deploying liquid cooling without relying on facility water offers significant advantages, such as avoiding the complexities of facility water infrastructure and reducing operational costs. This approach allows data centers to leverage alternative cooling solutions while maintaining high performance and reliability.
LTA sidecar supports GB200 NVL36 in 1:1 configurations
Are LTA Sidecars Scalable?
LTA Sidecars provide a scalable solution for liquid cooling in data centers. Their modular design allows for easy expansion as cooling demands grow, making them a flexible choice for data centers looking to enhance their cooling capabilities without extensive renovations.
Enablement
- Enables liquid–cooled IT equipment in air–cooled environments
- Localized, closed liquid cooling loop without dependency on any facility liquid infrastructure
- Wide environmental support range
- Standard 48U IT rack form–factor
- Performance optimized for low CFM/kW heat rejection performance
- Support for treated water or glycol mixtures up to 30%
- Stainless steel piping
Hot Swap and Serviceability
- Toolless hot–swap, redundant pumps, fans power supplies, temperature & pressure sensors
- Toolless hot–swap system controller
- Hot–swap, redundant filters
- 25–micron or 50–micron filtration options
- Flexible hose connections allowing installation flexibility
- Top & bottom feed liquid connection points to IT equipment
- Cleanable, reusable filters reduce operating costs
- All hot–swap components are serviceable without interrupting cooling
- Pumps, fans, power supplies, and fill & drain liquid connections located on the front of the unit for ease of service and accessibility
Why Choose nVent for Liquid Cooling?
Liquid cooling is vital for scaling AI, as it effectively manages the heat generated by high-performance computing systems. This approach improves reliability, lowers energy usage, and accommodates the heavy computational requirements of AI tasks.
nVent & NVIDIA Collaborate to Advance High Performance Computing
nVent is proud to collaborate with industry experts to develop innovative technical designs that tackle the most challenging high-density compute cooling issues. With over 15 years of experience in liquid cooling, nVent’s deep technical application experience and global presence enables creating efficient solutions that ensure optimal performance and reliability for advanced computing environments for our partners.