Cost & Quoting

Costing and Quoting Inverters, Motors, and Power Modules: A Practical Cost Breakdown

A money-focused breakdown of what actually drives cost per unit when quoting inverter, motor, and power module builds, and where estimates go wrong.

Cost per unit in this category splits into five buckets, and for a mid-power traction or industrial inverter the split usually lands near 55 to 70 percent purchased material (IGBT or SiC modules, DC-link caps, magnetics), 12 to 20 percent direct labor, 8 to 15 percent test and burn-in station time, 3 to 8 percent scrap and rework, and 10 to 18 percent applied overhead. Motors flip the ratio: copper and steel still dominate, but stator winding labor pushes the labor bucket to 20 to 30 percent. Quote the wrong bucket weighting and you can be 15 percent off on the number that decides whether you win the program.

Purchased material is where the dollars sit, and it is also where estimators get lazy. A blended assembly figure hides mix: a batch heavy in 1200V SiC half-bridges can run 3 to 5 times the die cost of a silicon IGBT equivalent, so a flat piece price built on last quarter's mix will underquote a SiC-skewed run badly. The Inverter Assembly Cost calculator keeps the fixed program adder (fixturing, ICT programming, NRE) separate from the per-assembly figure precisely so you do not smear one-time cost into marginal cost. Carry modules and passives as their own BOM lines, not inside a $45 blended touch number.

Direct labor is dominated by winding on motors and by hand-load and lead-dress steps on inverters. Use Motor Winding Labor with an 8 to 15 percent setup-and-handling allowance rather than a raw winding rate: a cell running 12 operations per minute nominal really delivers about 10.9 after threading, indexing, and tie-off, so costing at the nominal rate understates labor by roughly 9 percent. Rotor Balancing Cost adds another labor line only on the share of rotors that actually need dynamic balancing, often 40 to 100 percent depending on speed class, so a flat per-rotor charge across the whole build overstates it.

Test and burn-in time is the bucket most quotes get wrong because it is capital time, not just labor. Drive Test Time converts a batch into station minutes, and at 12 drives per minute with a 10 percent allowance you get roughly 5.5 seconds of functional-test occupancy per drive. Burn-in is different in kind: Inverter Burn-in Capacity models hours-long thermal and power cycling that does not shrink the way functional test does. A 4-hour burn-in at 24 slots per chamber costs a chamber-hour whether one unit or twenty-four ride through, so cost it per chamber-slot-hour and allocate the fixed dwell across the load, not per unit in isolation.

Scrap and rework carry an outsized dollar weight here because a scrapped power module takes expensive die with it. Power Electronics Scrap Cost multiplies scrapped units by a fully burdened per-unit value plus a fixed containment cost, and the per-unit value must include the modules already installed, not just the bare board. At 96 to 98 percent first-pass yield, a line scrapping 2 modules per 100 at $180 loaded each is bleeding $360 per hundred, or roughly $3.60 per good unit, before any rework labor. Thermal Interface Material Usage adds a quiet 5 to 15 percent overage from purge and priming that a clean rate-times-time model misses.

Overhead is where defensible quotes are won or lost, because most estimators apply a single plant-wide rate and call it done. A power electronics line with expensive burn-in chambers and dyno stands carries a heavier equipment burden than a hand-assembly cell, so split your overhead into a labor-driven rate and a machine-hour rate. Motor Test Stand Utilization matters directly to cost: a stand at 75 percent utilization spreads its fixed cost across 25 percent fewer test hours than one at 85 percent, so the per-unit test overhead is roughly 13 percent higher at the lower utilization. Under-absorbed capital quietly inflates every unit that crosses it.

To build the quote, cost from the bottom up and reconcile against a top-down check. Sum module and passive BOM, add winding and assembly labor at allowance-adjusted rates, add functional-test minutes plus burn-in chamber-slot-hours, add expected scrap at your real first-pass yield, then layer the split overhead and margin. Then divide total program cost by units and sanity-check against a known per-unit benchmark: $30 to $60 of assembly-and-test content per mid-power inverter excluding modules is a defensible range. If your bottom-up number lands far outside that, find the bucket that drifted before you send the price.

The estimates that blow up share three failure modes. First, costing gross capacity instead of good capacity: winding and burn-in lines lose 10 to 15 percent to uptime and yield, so a plan built on nameplate throughput underprices the units you actually ship. Second, hiding NRE inside piece price, which makes marginal-volume and second-source decisions wrong. Third, ignoring warranty exposure: thermal and high-voltage failures surface years later, and skipping the reserve makes a low-yield line look cheaper than it is. Quote the good-unit cost, keep fixed and variable separated, and reserve for the field, and your number will hold up under audit.

Published 2026-07-01.