Commerce: Specification Economics, TUR and Calibration Decisions
GATE IN does not carry a separate commerce section. The commercially relevant reasoning appears in instrument selection: accuracy versus cost, calibration intervals, test uncertainty ratio, and lifecycle cost. Treat this as design context, not as a paper section.
Core rules
- Total measurement uncertainty includes calibration uncertainty, drift, resolution, installation effects and repeatability.
- Test uncertainty ratio (TUR) = specification tolerance / measurement uncertainty. A 4:1 ratio is a common calibration target; a TUR near 1:1 leaves little margin for accepting good parts or rejecting bad ones.
- Guard band: accept only when the measured value is inside the tolerance by at least the measurement uncertainty.
- Calibration interval should be short enough that drift stays within an allocated fraction of tolerance; if drift is 0.3°C per year and the drift budget is 0.5°C, the interval is about 1.6 years, so 12 months with guard band is prudent.
- Cost of ownership = purchase + installation + calibration + downtime + energy. Over-specifying accuracy raises all of these.
- Simple payback = capital cost / annual savings; NPV discounts future savings.
Worked example
A process tolerance is ±1.0°C and calibration uncertainty is ±0.25°C. TUR = 1.0/0.25 = 4:1, which is acceptable. The transmitter drifts 0.3°C/year. Allocate 0.5°C of the tolerance to drift: 0.5/0.3 = 1.67 years. A 12-month interval leaves margin for drift and calibration uncertainty. Choosing a ±0.1°C transmitter instead would cut the uncertainty but may not change the pass/fail decision, so the extra cost buys little.
Traps that cost marks
- Comparing purchase price only and ignoring calibration and downtime.
- Treating accuracy as one number when it is quoted as % of reading, % of span, or ± counts.
- Accepting a TUR of 1:1 as adequate for a tight tolerance.
- Ignoring installation effects such as thermowell immersion or impulse-line temperature.