Developments in the way data centres are being cooled, such as hot aisle containment and modern air management solutions, typically allow for higher coil air-on temperatures and increased chilled-water setpoints, both of which allow for ambient temperatures to provide adequate cooling for a greater amount of time. Liquid cooling (including direct-to-chip) operates efficiently at elevated water temperatures, which extends free-cooling seasons and helps future-proof the cooling plant as IT loads rise.
The business case for free cooling is simple: more free cooling hours should mean fewer compressor hours, lower electricity spend and a smaller carbon footprint. The catch, however, is that a significant number of ‘mixed-mode’ installations don’t truly live-up to their claims, switching between modes in a binary way, with little or no time spent in partial free cooling.
When mixed-mode falls short
All free cooling chillers are marketed as supporting partial free cooling, where free cooling and mechanical refrigeration work together in order to deliver a smooth, efficient operating curve across the seasons; maximising low-cost ambient heat rejection and minimising compressor run time without compromising supply temperatures.
In practice, many systems spend very little time in a truly optimised ‘in-between’ state. Free cooling is effectively either fully on or forced off due to a conflict of fan speed requirements, typically happening during the shoulder seasons.
To extract optimal free cooling, condenser fans need to run faster to pull more air across the free cooling coil. But when compressors are running and the ambient temperature has reduced to the point that free-cooling becomes possible, the refrigerant circuit usually demands lower fan speeds to avoid over-condensing the refrigerant. If fans run too hard while compressors are active, the refrigeration cycle can move outside safe limits, increasing the risk of machine trips and even catastrophic compressor damage.
Most controllers will protect the compressors first. Fan speeds are pulled back to keep the refrigerant circuit stable, but that reduction in airflow also collapses the free cooling contribution. The result for owners is unsatisfactory: you’ve paid for free cooling hardware, but the controls prevent it being used for a meaningful share of the year.
Higher OPEX, reduced sustainability
When mixed-mode free cooling underperforms, compressors run far more than the design intent. That drives up chiller kWh, worsens PUE and inflates operating budgets.
In effect, you pay twice - capital for coils, controls and pipework that don’t deliver their potential, and ongoing energy for compressor hours that should have been avoided. Unnecessary compressor operation then consumes power capacity that could otherwise be allocated to IT load, limiting growth on sites where grid connection upgrades are slow or expensive.
Specify free cooling as a primary operating mode
A common mistake is treating free cooling as an optional enhancement to a conventional vapour-compression chiller. Owners can avoid this by making ‘free cooling first’ a core requirement at specification stage, choosing chillers designed to maximise ambient heat rejection whenever conditions allow, not only at very low outdoor temperatures. It’s important to understand how a unit will behave hour-by-hour across a typical year for your location and water temperature regime.
Adaptive controls
For effective mixed-mode operation, the differentiator is an optimised control strategy married to appropriate hardware. Owners should look for designs that can modulate free cooling contribution without forcing the entire refrigeration circuit to follow a single, conservative set of limits. The goal is to keep the plant stable and predictable for IT cooling, while still extracting as much ambient capacity as conditions allow.
One practical approach is to use isolating valves across condenser coil banks so that, when ambient conditions are favourable, redundant parts of the refrigerant circuit can be taken out of play. This allows fan speeds to increase to maximise heat rejection through the free cooling coils without destabilising the vapour-compression cycle.
Go Glycol free (if you can)
Glycol is often introduced into chilled-water circuits for freeze protection, but owners should be aware of the trade-offs. Glycol increases fluid viscosity, which raises pumping energy, and it typically reduces heat transfer efficiency, both of which compromise PUE. It also introduces additional handling, monitoring and environmental considerations over the life of the site.
Where designs can avoid glycol (for example through integrated, glycol-free free cooling architectures), owners can reduce losses in terminal unit heat transfer and simplify long-term maintenance.
What owners should ask for: annualised performance, not just a datasheet
Owners should request annualised modelling for their location, load profile and supply/return temperatures, with clear demonstration of when the unit operates in full free cooling, partial free cooling and full mechanical mode, and what that means for kWh, cost and carbon.
Two chillers may both claim free cooling at similar ambient temperatures yet deliver very different real-world results.
An owner’s checklist for making free cooling deliver
When evaluating or retrofitting free cooling plant, owners and their design teams can reduce risk by building these requirements into specifications and factory acceptance testing:
• Demand annualised energy modelling: request a clear breakdown of full/partial/mechanical operating hours and annual kWh at your setpoints and climate, not just peak-efficiency figures.
• Prove genuine mixed-mode capability: ensure the control philosophy avoids ‘all-or-nothing’ behaviour and can sustain partial free cooling without compromising refrigeration safety.
• Prioritise stability and resilience: ask how the system manages low-ambient operation, prevents outages, and protects compressors while still maximising free cooling contribution.
• Minimise parasitic loads: evaluate fan and pump energy across the year and consider choosing glycol free solutions.
• Require transparent controls and data: ensure the chiller exposes mode status, fan limits and efficiency KPIs to the BMS so performance can be tracked and optimised over time.
• Future-proof temperature regimes: select equipment that remains stable and efficient at elevated chilled-water temperatures and can support the transition to higher-density and liquid-cooled deployments.
For data centre owners, effective free cooling should reduce lifecycle cost, improve energy performance and release capacity for IT growth. Look for manufacturer and supply partners that can demonstrate year-round performance and verifiable mixed-mode operation, to ensure that free cooling becomes a dependable contributor to ROI, resilience and net-zero progress.
