Field glossary · measurement
Delta-T
Treat delta-T as a measured interval with named sensor locations, subtraction order and units. NIST's interval conversion makes 9°F equal to 5°C without a 32-degree offset. ONICON specifies better than ±0.18°F differential uncertainty for its matched System-40 sensor pair; 0.18°F is 9% of a 2°F spread. Small operating differences therefore need paired-sensor accuracy and synchronized readings, rather than additional display decimals.
The temperature difference between water entering and leaving a coil; the chilled water value the plant's efficiency rides on
Convert the interval without adding an offset
NIST SP 811 converts a Fahrenheit temperature interval to Celsius by multiplying by 5/9, with no 32-degree offset. A 9°F difference is therefore 5°C, while a 9°F absolute temperature is a different conversion. NIST also defines a 1°C interval as exactly 1 kelvin.
Keep the sign explicit. A cooling-water load warms water, while a cooling air coil lowers air temperature. A positive number labeled 'delta-T 10' leaves the reviewer guessing which stream and subtraction order produced it. Name entering and leaving points before applying any design comparison.
Pair error becomes large when the spread is small
ONICON's System-40 BTU meter specification uses matched 1,000-ohm platinum RTDs with differential uncertainty better than ±0.18°F. At a 10°F spread, 0.18°F is 1.8% of the difference; at a 2°F spread, it is 9%. These are temperature contributions alone, before flow-measurement uncertainty.
Two separate probes each specified at ±0.5°F can produce a worst-case difference error of ±1.0°F when their errors oppose each other. That is why individual sensor accuracy is not the same specification as matched-pair differential accuracy. A stable display can retain a stable offset.
Place both probes in the same stable temperature environment to reveal their relative offset, then verify their installed response. Equal readings in a common bath do not establish equal immersion depth, thermal contact or response time after installation.
Synchronize the temperature and flow readings
During a load transition, a fast probe and a slow probe can create an apparent temperature difference that does not represent one water parcel or one steady operating state. A supply trend sampled now and a return trend carried forward from an earlier interval can create the same artifact.
Pair timestamps and use readings after the measured boundary has stabilized. Store both temperature timestamps beside the hydronic balance reading so the three inputs belong to the same operating interval.
With the common plain-water factor of 500, a measured 60 GPM and 8°F spread gives 240,000 Btu/h. A 1°F bias in that spread changes the calculated output by 30,000 Btu/h, or 12.5%. DOE's data-center metering guide expresses the same water-side cooling balance as tons = GPM x delta-T / 24. That arithmetic cannot correct a biased temperature pair.
Distinctions
- Approach temperature
- A difference between specified streams or a stream and saturation condition; its locations differ from the inlet-to-outlet rise of one stream.
Frequently asked questions
Does a 0.1-degree display establish 0.1-degree accuracy?
No. Resolution describes the display increment. Accuracy, differential uncertainty, installation error and response time determine how well the difference represents the actual temperatures.
Is a low delta-T always excessive flow?
No. First establish the load, the measurement boundary and the accuracy of the pair. Small heat transfer, mixing and sensor error can each produce a small measured difference.
Sources checked
Also defined as
- Electrical: The temperature rise of a hot component over a reference, either a similar component under similar load or the ambient air
- Datacenter: The temperature rise across the unit coil; a strong, stable delta-T means the unit is drawing genuinely hot return air