US Liquid Cooling editorial · Reviewed · 3 min read
A rear-door heat exchanger places a heat-transfer surface at the back of a rack so heat carried by the rack’s exhaust air can be transferred into a liquid circuit. It is a different approach from placing cold plates directly on processors. Understanding that distinction helps buyers ask the right questions about airflow, liquid connections, and the facility.
Rear-door cooling should be evaluated as part of the installed rack environment. This guide offers a review checklist, not a recommendation for a particular door, server population, or cooling capacity.
Identify the rack and server requirements
Begin with the exact rack dimensions, installed equipment, cabling, and access requirements. Ask the rear-door supplier which rack interfaces are supported and whether adapters are needed. Confirm how the complete assembly is supported and serviced.
The server airflow arrangement remains relevant because the heat is carried to the rear-door exchanger through air. Ask the server and cooling suppliers to review the proposed combination. Do not assume that a door can compensate for any arbitrary server or rack layout.
Follow the liquid circuit
Identify the supply and return connections, the distribution equipment, and the heat-rejection destination. The door is only one heat-transfer element in that path. Specify who supplies hoses, manifolds, isolation provisions, and any required control or monitoring devices.
Clarify the fluid specification and the approved operating conditions for the door. If a CDU is included, show how its interface requirements align with the rear-door equipment. A complete diagram should distinguish the circuits and make every supplier boundary visible.
Review physical access carefully
A rear door must coexist with technicians, network cables, power connections, and the available aisle width. Ask for the required opening envelope and service clearance. Include the route and movement of any hoses during normal access and maintenance.
Review the layout with the people who will actually service the servers. A design can be thermally plausible but awkward operationally if routine work requires moving unrelated equipment. Photographs of a populated example installation can help the team identify practical questions, but exact drawings are still necessary.
Clarify active and passive elements
Ask which fans, controls, sensors, and power connections belong to the proposed model. Do not infer a control strategy from the generic term rear-door heat exchanger. Products differ, and the operating documentation should explain the actual arrangement.
Request the sequence for startup, shutdown, alarms, and maintenance. Establish what happens when the liquid-cooling service is unavailable and which server-side actions are expected. Those responses must be coordinated with the facility’s operating procedures rather than left to informal assumptions.
Compare performance on stated conditions
The quoted performance should identify the relevant air and liquid conditions. Ask the supplier to explain which assumptions are essential and how the proposed rack population was evaluated. A capacity number without its conditions provides little basis for comparing two systems.
The design team should also address the room’s remaining cooling requirements. Vertiv’s overview describes rear-door and direct-to-chip approaches in the context of hybrid air-and-liquid environments. That is a useful reminder to assess the full room rather than assuming that one component eliminates every other cooling requirement.
Prepare a practical acceptance plan
Before installation, agree on the documentation, inspections, and operating demonstrations required at handover. Include connection identification, service access, alarm visibility, and the maintenance procedure for the installed model.
Keep the final rack configuration with the cooling record. Future server changes, cable changes, or rack relocation should trigger a review of the original assumptions. Rear-door cooling is easier to manage when its mechanical, thermal, and operational interfaces are all recorded together.
Sources & further reading
- Vertiv liquid cooling options ↗
- ASHRAE data centers and telecommunications facilities ↗
- Open Compute Project cooling environments ↗
Explore the company directory → · All liquid-cooling guides →