Secondary Supply Temperature Is the Entire Rejection Choice

What secondary supply temperature does the IT equipment actually allow? No document on this project states it. Lacking one, the heat-rejection basis enumerates three industry classes as an axis rather than as candidates — 45, 40 and 32 °C, permitting primary supply maxima of 42, 37 and 29 °C at an assumed 3 K approach across the distribution unit [MODELED] — and says plainly that none of the three is the project's value. The outer two branches are 13 K apart; the capability spread between the three rejection routes at the summer design condition is 10 to 15 K. The two spans are the same size.

Until Supply Temperature Is Known, Route Verdicts Are Void

Because one unknown covers as much ground as the whole choice, until the supply temperature is answered any assertion that a route is feasible or infeasible is a void assertion. What the chapter may publish is each route's capability and its intersections with the three branches. That is a constraint on conclusions, not a claim that the routes are indistinguishable.

The hydraulic basis already named this the hardest interface on the project. If primary summer supply sits at 35 to 40 °C and the OEM requires secondary supply at or below 32 °C, the approach across the distribution unit is negative. Coolant-distribution capacity is then not derated. It simply does not exist.

A Route Must Hold on Every Condition It Must Cover

The intersections are printed as nine cells of estimated primary supply temperature, margin and verdict. At 42 °C permitted, the dry route clears the annual-average condition with 15 K in hand, misses summer by 6 K and the extreme condition by 12 K; adiabatic assistance clears annual by 18 K, holds summer by 2 K and the extreme condition by exactly 0 K [MODELED]. At 37 °C permitted, adiabatic assistance fails summer by 3 K and extreme by 5 K. At 29 °C permitted, every route but mechanical refrigeration fails everywhere.

The reading rule sits beside the matrix and is why the columns matter more than the rows: a route must hold on every condition it is required to cover, and may not be declared a winner on a single condition. One qualifier travels with all nine cells — an open item allows derated or staged operation to be adopted, which would relax the extreme-condition row, and the chapter does not presume the answer.

Dry Fails; Adiabatic Is Conditional; 32 °C Needs Chillers

The robust one is that dry cooling cannot carry the year alone, and that holds across all three branches and the whole assumed approach range of 8 to 12 K, which fixes its role in advance of the interface: an economiser stage inside a hybrid system, never the main route.

The fragile one is that adiabatic viability depends entirely on the unknown, survives only in the 45 °C branch, and even there sits in a critical band. Taking the adiabatic effectiveness at 0.70 instead of 0.80, or the unit approach at 4 K instead of 3 K, flips the summer cell on its own [SENSITIVITY].

The third ends an argument. In the 32 °C column every route but mechanical refrigeration fails, so a 32 °C answer makes mechanical refrigeration mandatory rather than supplementary. That is the project's largest single mechanical fork. The conclusion carries its own pending marker: the verdict has no validity for this site and must be reissued after the weather recomputation.

A hybrid of adiabatic assist with mechanical trim is carried only as an interface-reserve study branch. It is not a default to advance. It is not a freeze of any equipment.

Dry Summer Failure Holds Across the Full Approach Range

Sweeping the dry-cooler approach across its full 8 to 12 K range puts the summer primary supply temperature between 46 and 50 °C [SENSITIVITY]. The whole interval sits above 42 °C, the most permissive of the three ceilings, so the dry route's summer failure does not depend on which end of the assumption is taken. The author notes this is the section's only assumption-insensitive result, which is also an admission about the rest of it.

Every cell temperature was computed on a placeholder weather chain a later erratum declared to carry no site validity. The structure of the argument survives recomputation. The values do not.

Dry-Priority Setpoints Would Trigger Wet Cooling Anyway

An independent review of the vendor packet shows the interface producing a control conflict downstream. A new long-term heat-exchanger primary inlet of 35.6108 °C is recorded, with 35.3498 °C at peak and 37.2021 °C at 25 % load, against legacy logic calling for three wet-cooling units whenever the cold-source outlet reaches 34 °C — and no mixing or tempering section between the two measurement points that would explain the difference [MODELED]. On the current record, dry-priority operation would trigger wet cooling while meeting its own target. The reviewer separates this from the site question about wet towers and requires the setpoints, measurement points, state machine, anti-chatter and timeout to be redone once the cold source is chosen, prohibiting the substitution of equipment names under the old thresholds.

The Enquiry Must Ask Four Interface Items Together

The close is a procurement instruction rather than an engineering one: the enquiry must take four items at once — maximum permitted supply temperature, maximum permitted return temperature, the nameplate fluid and rating condition, and the permitted cold-plate pressure drop — because asking only the first and receiving a single number would induce a route freeze on an incomplete basis.

Limits and open items

Modelled and not frozen: the 45 / 40 / 32 °C axis, the 42 / 37 / 29 °C ceilings, the nine-cell margins, and the vendor-packet inlet figures. The 46 to 50 °C dry-route sweep and the 0.70 / 4 K flips are sensitivities on that same axis. None of the three branches is the project's value. [MODELED]

Open: the OEM supply temperature itself; whether derated or staged operation is allowed at the extreme condition; the four approach parameters behind the matrix; and the weather recomputation that has to reissue every cell. [HOLD]

Nothing in the matrix has been built, operated or ordered against. No route is selected. No Engineer of Record has reviewed these quantities. The item to assign is the enquiry itself, with all four limbs in one document.

© 2026 K&K Data Service Inc. All rights reserved. Reproduction or republication is permitted only with clear attribution to K&K Data Service Inc. and a working hyperlink to the canonical URL of this article. Excerpts must preserve the technical context, maturity labels, assumptions, and limitations. No excerpt may imply project approval, field validation, certification, or endorsement that the original article does not state.

Source: K&K Data Service Inc., “Secondary Supply Temperature Is the Entire Rejection Choice,” https://www.kkdatasvc.com/lab/heat-rejection-and-pue/secondary-supply-temperature-is-the-entire-rejection-choice/.