Sensor Costing

What Drives Cost Per Unit for Industrial Sensors and How to Quote It

A money-focused breakdown of what actually moves cost per unit in sensor and instrumentation manufacturing, how to build a quote that holds, and the estimating traps that erode margin.

Sensor cost per unit splits into five buckets: direct material, direct labor, machine and test time, scrap and rework, and allocated overhead. For a mid-volume pressure transmitter quoted at 5,000 units per year, a typical split is material 38 percent, labor 22 percent, test and calibration time 18 percent, scrap 7 percent, and overhead 15 percent. If your fully loaded cost lands at 84 dollars, that is roughly 32 dollars material, 18.50 labor, 15 test, 6 scrap, and 12.60 overhead. The Sensor Manufacturing Cost calculator lets you enter each bucket so a quote is built from parts, not marked up from a competitor's price.

Material cost is dominated by the sensing element and package. A piezoresistive silicon die runs 4 to 9 dollars, a 316L stainless housing 6 to 14 dollars, and hermetic feedthroughs 2 to 5 dollars, before oil fill, o-rings, and connectors that add 3 to 7 dollars. Buy at 5,000 units per year and you will not clear the price breaks that appear near 25,000, so expect to pay 12 to 18 percent more per die than a high-volume competitor. The Temperature Probe Cost calculator separates sheath, element, and lead-wire cost so you see which line item moves when you switch from a thin-film RTD to a thermocouple.

Calibration and test time is the bucket estimators most often understate. A multi-point pressure calibration across five setpoints with settling can take 8 to 14 minutes of bench time plus 2 minutes handling. At a 95 dollar per hour loaded bench rate, 12 minutes is 19 dollars per unit, more than the die on many parts. Flow meters are worse: a wet-flow rig calibration can run 25 to 60 minutes. The Flow Meter Calibration Cost and Instrumentation Test Time calculators convert those minutes and rig rates into per-unit dollars so calibration stops hiding inside a vague overhead percentage.

Scrap is not a flat percentage; it compounds through the line. If a sensor line runs 94.5 percent first-pass yield, you build 5,291 to ship 5,000, and every scrapped unit at the seal-and-trim step has already absorbed most of its material and labor. Value scrap at the stage it occurs: a unit lost after calibration carries 60 to 75 dollars of accumulated cost, not the 32 dollar material figure. A 2 point yield swing from 94.5 to 92.5 percent on 5,000 shipped adds about 115 rebuilt units, near 8,000 dollars a year at 70 dollars sunk each.

Burn-in and reliability screening add fixed cost that is easy to omit. A 48-hour burn-in at 125 degrees C spreads chamber depreciation, energy, and floor space across whatever the chamber holds. A 500,000 dollar chamber over 7 years is roughly 71,000 dollars a year, and at 102 units per day effective throughput that is near 2.80 dollars per unit before energy. The Sensor Burn-in Capacity calculator gives the throughput denominator so you can allocate that fixed cost honestly; underloading the chamber to 40 percent triples the per-unit hit to over 8 dollars.

Overhead allocation decides whether a quote wins or loses money quietly. Spreading plant overhead on labor hours penalizes labor-light automated lines and hides cost on manual ones. Prefer an activity basis: allocate calibration overhead per calibration event, test overhead per test minute, and burn-in overhead per chamber-hour. The Sensor Calibration Workload calculator gives the event count that becomes your allocation denominator. Get this wrong and a high-touch instrument looks cheaper than it is, and you win the order at a loss you only discover at year-end variance.

A defensible quote states volume, yield, and rate assumptions on the face of it. Write down 5,000 units per year, 94.5 percent FPY, 95 dollar per hour bench rate, 12 minute calibration, and a 22 percent margin, then let a reviewer challenge each number. When the customer pushes to 82 dollars, you can show that a move to 25,000 units per year unlocks die price breaks and burn-in loading that legitimately take cost to 74 dollars, rather than cutting margin blind. Firmware, often forgotten, is real: amortize Smart Sensor Firmware Workload hours across the program volume, not the first order.

Common estimating failures cluster in three places. First, quoting material only and treating test, calibration, and burn-in as overhead, which understates true cost by 25 to 40 percent on precision instruments. Second, using shipped volume for yield-sensitive material, ignoring the 5 to 8 percent of units you must build and scrap. Third, applying a single blended labor rate when a calibration technician at 95 dollars and an assembler at 55 dollars do very different work. Build the quote bottom-up in the Sensor Manufacturing Cost calculator, name every rate, and reconcile to last quarter's actual cost per unit before you send it.

Published 2026-07-01.