Download document () of 20
Send email
Download
You have exceeded the download limit
Leverage the Eaton rack emulator to verify leak-free piping and Cooling Distribution Unit (CDU) operation, check electrical input power wiring and delivery, and most importantly, simulate the pressure drop of the server rack to validate the CDU performance. The Eaton Rack emulator can verify CDU performance across rack heat loads up to 250 kW, covering current and next-generation AI server rack densities. Customers that use the rack emulator can expedite AI Data Center installations and performance validation, bringing more processing power online faster.
A rack emulator is a piece of equipment designed to test and verify the thermal and flow performance of a coolant distribution unit prior to IT equipment installation. A rack emulator sits in the same footprint as a rack and mimics the thermal and flow load output of a real rack.
The Eaton rack emulator is specifically designed to validate coolant distribution unit (CDU) performance before any IT equipment is connected, making it an essential tool for AI data center liquid cooling commissioning.
A rack emulator mimics the thermal load and pressure drop of a rack of IT equipment. Commissioning personnel can test and validate input power with the rack emulator as part of the liquid cooling commissioning process prior to AI rack installation. A series of high-powered heaters generate heat equivalent to the amount created by AI processors, memory and other electronics that would be connected to the technology cooling system (TCS) liquid cooling loop. A control valve adjusts to match the pressure drop the TCS loop would experience across the entire liquid system within the rack, from the liquid cooling loop, through quick disconnects, and out through the manifolds. In addition, commissioning personnel can test and validate input power with the rack emulator prior to AI rack installation.
CDU testing with a rack emulator follows a structured commissioning workflow. The rack emulator is connected to the CDU and liquid cooling loop in place of the server rack. Coolant flow, pressure drop and thermal performance are validated under simulated rack conditions across heat loads up to 250 kW. Any issues with piping, fittings, CDU settings, or flow rates are identified and corrected before IT equipment is installed, protecting servers from commissioning errors and reducing deployment time for AI data center builds. Commissioning teams can also verify leak-free piping connections and confirm that the CDU maintains the required supply temperature stability under full load before the first server is ever connected.
Eaton's rack emulator fully simulates AI server rack function and footprint, as opposed to traditional load banks that only force heat through a system. Traditional load banks cannot replicate hydraulic pressure drop, while the Eaton rack emulator validates both thermal load and hydraulic resistance for accurate CDU performance validation. Our rack emulator is a dynamic load that can imitate a server rack during operation. This enables commissioning professionals to ensure complex liquid cooling systems are correctly set up to meet performance requirements.
A traditional load bank generates resistive heat to stress-test electrical systems but cannot replicate the hydraulic pressure drop of a real server rack. This means a load bank cannot validate how a CDU or liquid cooling loop will perform under real rack conditions. An Eaton server rack emulator generates both the correct thermal load and the correct hydraulic resistance simultaneously, giving commissioning teams a complete and accurate picture of liquid cooling system performance before any IT equipment is installed. The rack emulator also validates electrical input power wiring and delivery a step load banks cannot perform in a liquid-cooled environment.
The Eaton rack emulator supports validation of CDU and liquid cooling loop performance across a wide range of server rack standards and processor platforms. Supported configurations include OCP (Open Compute Project) rack designs, NVIDIA GPU server racks including GB200 NVL configurations, AMD EPYC and Instinct processor platforms, Intel Xeon platforms, and custom high-density AI cluster deployments. The single-unit design mimics thermal and flow loads up to 250 kW, covering current rack densities and next-generation AI infrastructure where rack power is trending well beyond 100 kW per rack. Contact Boyd Thermal by Eaton to discuss rack emulator configuration for a specific platform or deployment.
Features |
Metric |
Imperial |
| Dimensions | 610mm x 1513mm x 2400mm | 24” x 60” x 94.5” |
| Weight (dry) | 770 kg | 1700 lbs |
| Electrical service | 380-480 VAC / 3 Ph / 50 and 60 Hz | |
| Thermal load | Up to 250 kW at 480 VAC | |
| Minimum flow rate | 70 LPM | 18.5 GPM |
| Maximum flow rate | 420 LPM | 110 GPM |
| Minimum pressure drop | 0.4 bar | 6 PSI |
| Maximum pressure drop | 2.75 bar | 40 PSI |
MyEatonFAQs
