Reducing approach temperature makes a liquid-to-liquid CDU more demanding to size: at the same cooling duty and coolant flow rates, a smaller temperature difference requires greater heat exchanger conductance. That generally means more heat transfer area, a different plate configuration, or both. Increasing flow can help under some conditions, but it also increases hydraulic demand and cannot compensate indefinitely for a shrinking temperature difference.

For technical evaluation, the first decision is whether the specified approach is necessary at the facility’s warmest water condition. A tighter approach can allow warmer facility water while maintaining the required coolant supply temperature to the servers. Its value must be weighed against exchanger size, pressure drop, pumping power and the margin available when operating conditions depart from the rating point.

Define the approach at the correct connections

For this discussion, approach temperature is the secondary coolant supply temperature leaving the CDU minus the primary facility water temperature entering it:

Approach = T_{secondary supply} − T_{primary inlet}

A CDU delivering secondary coolant at 40°C with facility water entering at 35°C therefore operates at a 5 K approach. These figures are an illustrative design condition, not a recommended operating temperature. The acceptable secondary supply temperature must come from the connected equipment’s thermal and coolant requirements.

Confirm the measurement locations before comparing quotations. A temperature measured at a CDU outlet is not necessarily the temperature delivered to a rack after distribution piping. Sensor tolerances, pump heat and heat gains along the circuit can consume part of a small temperature allowance. Suppliers should also identify whether their quoted approach applies to heat exchanger ports or complete-unit connections.

Approach must be distinguished from the coolant temperature rise through the IT equipment. The latter determines secondary flow for a given heat load. Approach describes the temperature separation between the two circuits at one end of the exchanger. Specifying one does not establish the other.

Why a tighter approach changes exchanger size

The basic thermal relationship is Q = U × A × ΔT_lm, where Q is heat duty, U is the overall heat transfer coefficient, A is effective heat transfer area, and ΔT_lm is the logarithmic mean temperature difference. For a counterflow exchanger, that mean depends on the temperature differences at both ends.

The supply-side approach therefore provides only part of the sizing information. The other terminal difference is the secondary return temperature minus the primary outlet temperature. Both must be included in the thermal calculation, along with any correction required for the exchanger’s actual flow arrangement.

Consider an illustrative 1 MW duty with secondary coolant entering the exchanger at 50°C and leaving at 40°C. Suppose primary water enters at 35°C and leaves at 40°C. The terminal differences are 5 K and 10 K, giving a logarithmic mean temperature difference of approximately 7.2 K.

Now increase the primary inlet to 38°C while holding duty, flow rates and secondary temperatures constant. The primary outlet becomes 43°C. The terminal differences fall to 2 K and 7 K, and the logarithmic mean falls to approximately 4.0 K. Required UA rises by roughly 80%. If U remained unchanged, the required area would increase by the same proportion.

That is a calculation of thermal conductance, not a prediction that every supplier’s unit will become 80% larger. Plate geometry, channel velocity, fluid properties and flow distribution affect U. Adding plates can change channel velocities, so exchanger selection requires an iterative thermal and hydraulic calculation. An approach specification alone cannot establish the necessary plate count or footprint.

The example also shows why a linear sizing allowance is unreliable. Reducing approach from 5 K to 2 K does not imply a proportional change in capacity or area. As a terminal temperature difference approaches zero, the required conductance becomes increasingly demanding. A finite passive exchanger requires a positive local temperature difference to transfer heat.

Flow can help, but it has a different job

Each circuit must satisfy its heat balance: Q = ṁ × c_p × ΔT. Secondary flow follows from the heat load, coolant properties and permitted temperature rise across the IT equipment. Primary flow follows from the heat load and the allowed facility-water temperature rise. A smaller approach does not, by itself, prescribe a higher flow rate.

Increasing primary flow reduces the primary temperature rise at a fixed duty. This lowers the primary outlet temperature and can improve the mean temperature difference. Higher velocity may also improve the heat transfer coefficient. However, primary flow cannot change the approach defined by a fixed primary inlet temperature and a fixed secondary supply target.

Increasing secondary flow reduces the temperature rise through the IT equipment for a fixed load and may improve heat transfer coefficients. It also changes the secondary return temperature and exchanger temperature profile. Its net benefit must be calculated against the complete operating point; higher secondary flow is not an automatic substitute for additional exchanger area.

Both options require a pressure-drop check. Hydraulic pump power is proportional to volumetric flow multiplied by pressure rise, with electrical input also depending on pump and motor efficiency. Evaluate the exchanger together with piping, valves, manifolds and connected equipment. A thermal selection that needs more flow than the facility can supply, or more differential pressure than the secondary pump can deliver, is unusable at that duty.

Fluid selection adds another constraint. Water-glycol mixtures have different heat capacity, viscosity and heat transfer behaviour from water. A water-based rating should not be transferred directly to a glycol circuit. Request calculations using the specified concentration and operating temperatures, including any colder operating condition that could constrain pump performance.

Compare guaranteed operating points, then examine the envelope

A nominal capacity becomes useful only when its boundary conditions are visible. Two CDUs carrying the same kilowatt rating may deliver substantially different capacities at the facility’s available water temperature. Ask for a common selection basis that includes:

• Required heat duty, secondary supply target and secondary return temperature or flow.
• Primary inlet temperature, available primary flow and allowable primary pressure drop.
• Coolant composition, relevant fluid properties and secondary pressure requirements.
• Guaranteed capacity, temperature tolerances and the basis for any fouling allowance.
• Available duty under the specified pump or module redundancy condition.

Use the warmest credible facility-water inlet condition when checking the limiting thermal case. If the primary inlet rises while the required secondary supply remains fixed, the available approach contracts. A unit selected at a favourable nominal temperature may then require more primary flow, allow the secondary supply temperature to rise, or operate below its nominal capacity.

Request performance at several defined operating points rather than relying on one favourable rating. Include full load at the warmest primary inlet, expected normal operation and the relevant redundancy condition. For part-load operation, examine control stability and minimum flow limits. An exchanger with sufficient peak capacity still needs pumps and valves that can regulate temperature across the intended load range.

Reserve should also have a clear purpose. Extra exchanger conductance, additional pump head and a lower normal supply setpoint address different constraints. A stated percentage of “design margin” is difficult to assess unless the supplier explains which operating condition it protects and how performance changes when that margin is consumed.

The useful selection question is therefore specific: can the CDU deliver the required secondary supply temperature and heat duty at the warmest available primary inlet, within both circuits’ flow and pressure limits? Once that point is demonstrated, the evaluator can judge whether paying for a tighter approach provides worthwhile facility-water flexibility—or whether a modestly higher secondary supply temperature, within equipment limits, produces a better system-level result.