OSAT Final Test Cost Per Good Unit Calculator
Test is a large share of the cost of many packaged ICs, and it is driven by seconds. This calculator shows how test time, index time, parallel sites, utilization and yield combine into cost per good unit.
Your numbers
Tester depreciation, handler, interface hardware, floor space, labor and utilities, per available hour.
Time to unload, load and settle devices between insertions.
Loss from serial test sections and site-to-site overhead.
Share of paid hours spent testing, after jams, changeovers and waiting for lots.
Your results
This model ignores setup, retest policy and probe-card or contactor consumables beyond what you load into the hourly rate. Add retest and consumables to the hourly cost for a fuller view.
Cost per good unit vs parallel sites
Sweeps Parallel test sites from 1 to 32, holding your other inputs fixed.
Cost of test by product from your ERP
Optional: ask about reporting test cost per lot and customer from your ERP and tester logs.
Optional. The tool and every export work without this. No spam.
How to use the calculator
Start with the fully loaded hourly cost of the test cell, including the tester, handler, interface hardware, labor and floor space. Enter the test time and handler index time per insertion, the number of sites and the efficiency of running them in parallel. Add the cell utilization and the final test yield. The result is units per hour, cost per unit tested, cost per good unit and the cell hours needed for your monthly volume. The chart shows what adding sites does to cost per good unit.
- Take cell cost from your finance or fixed asset records.
- Use measured utilization from the MES or tester logs.
- Check retest rates, which effectively lower yield and raise cost.
The cost model
Units per hour equals 3600 divided by the sum of test time and index time, times sites, times parallel efficiency, times utilization. Cost per unit tested equals the hourly cell cost divided by units per hour. Cost per good unit divides that by yield. Because the index time is paid once per insertion, parallel sites spread it across more devices, which is why multi-site test cuts cost with diminishing returns as efficiency falls.
Worked example
At 120 dollars per hour, 6 seconds test, 0.7 second index, 8 sites, 90% efficiency and 80% utilization, the cell runs about 3,870 units per hour. Cost per unit tested is about 3.1 cents, and at 96% yield cost per good unit is about 3.2 cents. A volume of 2 million units needs about 517 cell hours a month. Reducing test time by one second gives a visible saving, which is why test engineers chase test time reduction first.
Common mistakes
Using the tester purchase price instead of a fully loaded hourly cost, or ignoring utilization, hides the real cost. Another common mistake is assuming perfect multi-site scaling. Serial sections such as contact checks and shared instrument time mean 16 sites rarely deliver double the throughput of 8. Finally, remember that retest and handler jams change the effective yield and time per good unit, so use measured data where you can.
Frequently Asked Questions
What should be included in the hourly cost?
Include depreciation or lease of the tester and handler, interface boards and sockets, maintenance contracts, operator labor, floor space, power and cleanroom allocation if applicable. Excluding any of these makes cost per unit look lower than it is.
Why does index time matter so much?
It is dead time between insertions during which the tester cannot test. For short test times, index time can be a large share of the cycle, so faster handlers or more sites can have more impact than small test time cuts.
Does yield belong in test cost?
Yes for cost per good unit. Failed units still consume test time, so the cost of testing them is borne by the good units. A yield drop therefore raises the cost per shipped part directly.
How is this different from the die cost model?
This covers only the final test step. Wafer cost, probe, assembly and packaging are separate steps in the total cost per shipped unit, and each can be modeled in the same way.
If your OSAT or fabless team needs test cost reported by customer, product and lot from ERP and tester data, Netray can help design that reporting on your own infrastructure.
Related Tools
Burn-In Arrhenius Acceleration Factor and FIT Calculator
Calculate the Arrhenius temperature acceleration factor, equivalent use time and a point-estimate FIT rate for burn-in or HTOL results.
Electronics ManufacturingSemiconductor Cycle Time X-Factor Queue Calculator
Estimate queue wait, X-factor and total flow time per layer from tool utilization and variability. Includes a utilization sweep chart.
Quality & InspectionSemiconductor OCAP Out-of-Control Action Plan Template
Fill-in OCAP record for an SPC out-of-control event: rule violated, tool and lot hold, containment, checks, disposition and sign-off.
Electronics ManufacturingEtch Rate Uniformity Log Worksheet
Editable log of etch monitor wafers with automatic within-wafer non-uniformity and etch depth per row, plus average and worst-case totals.
More Costing & Finance tools
See all- Wafer Cost Per Good Die Calculator
- SAP QM Inspection Effort and Cost Calculator
- SAP Standard Cost Estimate BOM Rollup Calculator
- Legacy ERP Retirement Cost and Payback Calculator
- JD Edwards World Month-End Close Procedure Template
- Distributor Customer Credit Application Template
- Distributor Customer Profitability Worksheet
- Distributor GMROI and Inventory Turns Calculator
- Vendor Early Pay Discount Worksheet