Stress test acceleration factor calculator
Work out how much faster a stress test ages a part than use does: Arrhenius, Eyring, Peck, Coffin-Manson, Norris-Landzberg, voltage and Black's equation.
ReliabilityFree to useHow it worksSourcesWhere it helps
Fills the inputs below with typical figures. Replace any of them with your own.
- Acceleration factor
- 8.6079
- Hours of use per hour of test
- 1,000 hours of test stand for
- 8,608 h
- 0.9826 years
- Test time that covers the use life
- 1,394 h
- 58.09 days
- stress at 105 C (solid line, circles)
- stress at 125 C (dashed line, squares)
- stress at 145 C (dotted line, triangles)
Show the data as a table
| Activation energy (eV) | stress at 105 C | stress at 125 C | stress at 145 C |
|---|---|---|---|
| 0.2 | 1.359 | 1.85 | 2.445 |
| 0.3 | 1.584 | 2.516 | 3.822 |
| 0.4 | 1.847 | 3.422 | 5.976 |
| 0.5 | 2.153 | 4.654 | 9.343 |
| 0.6 | 2.51 | 6.329 | 14.61 |
| 0.7 | 2.926 | 8.608 | 22.84 |
| 0.8 | 3.411 | 11.71 | 35.71 |
| 0.9 | 3.976 | 15.92 | 55.83 |
| 1 | 4.635 | 21.66 | 87.3 |
| 1.1 | 5.404 | 29.45 | 136.5 |
| 1.2 | 6.3 | 40.06 | 213.4 |
- 0.50 eV (solid line, circles)
- 0.70 eV (dashed line, squares)
- 0.90 eV (dotted line, triangles)
Show the data as a table
| Stress temperature (C) | 0.50 eV | 0.70 eV | 0.90 eV |
|---|---|---|---|
| 92 | 1.247 | 1.362 | 1.487 |
| 98.64 | 1.656 | 2.026 | 2.478 |
| 105.3 | 2.177 | 2.972 | 4.056 |
| 111.9 | 2.835 | 4.302 | 6.527 |
| 118.5 | 3.66 | 6.15 | 10.33 |
| 125.2 | 4.685 | 8.688 | 16.11 |
| 131.8 | 5.948 | 12.14 | 24.76 |
| 138.5 | 7.493 | 16.77 | 37.53 |
| 145.1 | 9.371 | 22.94 | 56.14 |
| 151.7 | 11.64 | 31.07 | 82.92 |
| 158.4 | 14.36 | 41.68 | 121 |
| 165 | 17.6 | 55.44 | 174.6 |
Sign in to save your inputs and come back to them later.
| Factor | Formula | Value |
|---|---|---|
| Arrhenius temperature factor | exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] | 8.6079 |
| Mechanism | Model | Constants | Temperature factor | Cycling factor |
|---|---|---|---|---|
| Generic high temperature operating life | Arrhenius | Ea = 0.7 eV | 8.608 | not applicable |
| Gate oxide breakdown (TDDB) | Combined | Ea = 0.3 eV | 2.516 | not applicable |
| Electromigration, aluminium-copper | Black's equation | Ea = 0.8 eV | 11.71 | not applicable |
| Electromigration, deep sub-micron | Black's equation | Ea = 0.8 eV | 11.71 | not applicable |
| Stress migration | Stress migration | Ea = 0.55 eV | 5.427 | not applicable |
| Hot carrier injection | Black's equation | Ea = -0.15 eV | 0.6305 | not applicable |
| Negative bias temperature instability (placeholder constants) | Combined | Ea = 0.1 eV | 1.36 | not applicable |
| Aluminium corrosion, 85/85 style | Peck | Ea = 0.75 eV | 10.04 | not applicable |
| Mobile ion drift | Arrhenius | Ea = 1 eV | 21.66 | not applicable |
| Solder fatigue, tin-lead | Norris-Landzberg | n = 1.9, m = 0.3333, Ea/k = 1414 K | not applicable | 9.615 |
| Solder fatigue, lead-free SAC | Norris-Landzberg | n = 2.65, m = 0.136, Ea/k = 2185 K | not applicable | 73.46 |
| Die attach and solder joints, ductile | Coffin-Manson | q = 2 | not applicable | 17.02 |
| Wire bond intermetallics | Coffin-Manson | q = 4 | not applicable | 289.5 |
| Brittle cracking of die and dielectrics | Coffin-Manson | q = 7 | not applicable | 20,320 |
The temperature column applies each mechanism's activation energy alone, and the cycling column applies its exponents to the cycling ranges above. Pick a mechanism from the presets for the full model with its other terms. 14 mechanisms are listed.
How it works
An acceleration factor says how many hours (or cycles) of use one hour (or cycle) of a stress test stands for. This tool covers the models reliability engineers use most: Arrhenius, Eyring, Peck, Coffin-Manson, Norris-Landzberg, voltage, Black's equation and stress migration, and a product of factors for tests that stress more than one thing. It converts test time to use time, and works out the test time a use life needs.
- Arrhenius temperature factor. For a thermally activated mechanism, the rate follows exp(−Ea / kT). The factor is the ratio of the rate at the stress temperature to the rate at the use temperature. Temperatures are in kelvin and k = 8.617 x 10^−5 eV/K.
AF = exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] - Eyring: a temperature prefactor and a second stress. The Eyring model adds a prefactor that grows with temperature and a term for a second stress S such as humidity or voltage. With p, B and C at zero it reduces to Arrhenius. The forms of the stress term differ between authors; this tool uses one common form.
AF = (Tstress / Tuse)^p x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] x exp[(B + C / Tstress) x Sstress − (B + C / Tuse) x Suse] - Peck: temperature and humidity. Peck's model multiplies the Arrhenius factor by a power of the humidity ratio, with an exponent near 2.7 for aluminium corrosion. An optional power of the voltage ratio covers biased tests.
AF = (RHstress / RHuse)^n x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] x (Vstress / Vuse)^m - Coffin-Manson and Norris-Landzberg: cycling. Fatigue is driven by the temperature swing. Coffin-Manson raises the ratio of the swings to an exponent q. Norris-Landzberg, for solder, adds a cycle-frequency term (a slower cycle gives more creep damage per cycle) and an Arrhenius term on the peak temperature. These factors compare cycles to failure, so time follows from the cycles per day.
Coffin-Manson: AF = (dTstress / dTuse)^q; Norris-Landzberg: AF = (dTstress / dTuse)^n x (fuse / fstress)^m x exp[(Ea / k) x (1 / Tmax,use − 1 / Tmax,stress)] - Voltage, current density and stress migration. Voltage acceleration is exponential (a constant gamma per volt, or per MV/cm for oxide field) or a power law. Black's equation gives electromigration life as a power of current density times an Arrhenius term; the same form with a negative Ea covers hot carrier injection. Stress migration adds a thermomechanical term that falls as the stress temperature nears the stress-free temperature T0 of the metal.
voltage: AF = exp[gamma x (Vstress − Vuse)] or (Vstress / Vuse)^n; Black: AF = (Jstress / Juse)^n x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)]; stress migration: AF = ((T0 − Tuse) / (T0 − Tstress))^(−n) x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] - Combined factors and time. When a test stresses several things at once and they act independently, the factors multiply. Equivalent use time is the factor times the test time; the test time a use life needs is the life divided by the factor. For cycling models the same ratios apply to cycles, and cycles per day convert them to time.
AF total = AF1 x AF2 x ...; equivalent use time = AF x test time; test time needed = use life / AF
Assumptions
- One failure mechanism is being accelerated, and the stress does not change it. If the mechanism changes, the factor does not apply.
- The constants (Ea, exponents, gamma) are properties of the mechanism, technology and package. The presets are typical published values; use your own data where you have it.
- Temperatures are junction or metal temperatures, not oven settings, unless you have added the self-heating.
- Combined factors assume the stresses act independently on one mechanism.
Limits and warnings
- A small error in Ea or an exponent grows with the size of the step: a 0.1 eV error changes a 125 C to 55 C factor by about a factor of 2. The sensitivity charts show this.
- Eyring stress terms, Peck constants and Norris-Landzberg constants are published in several forms. Check that the form and units match the source of your constants.
- Coffin-Manson and Norris-Landzberg cover the cycle range and rate, not dwell time at the extremes, which matters for creep in solder.
- The mechanism table shows the temperature factor for each preset mechanism. It leaves out voltage, humidity or current terms that some of them also need.
- Real mission profiles vary in time. Use an effective temperature, or split the profile into phases, rather than one average that hides the hot hours.
Sources
The methods and default figures rest on these public sources. The wording here is our own; read the source for its full text.
- JEP122, Failure Mechanisms and Models for Semiconductor Devices (JEP122H) (opens in a new tab), JEDEC Solid State Technology Association, September 2016. The standard list of mechanisms, model forms and activation energies. The forms here are written in our own words.
- Semiconductor Device Reliability Failure Models (Technology Transfer 00053955A-XFR), Richard Blish and Noel Durrant (opens in a new tab), International SEMATECH, 31 May 2000. Open report with the model forms, typical constants and worked examples for hot carrier, stress migration and temperature cycling.
- Automotive Electronics Reliability Testing Starts and Ends with the Mission Profile, Thomas Eikenberg (opens in a new tab), Monolithic Power Systems, published by Power Electronics News, 14 May 2021. The worked example of 12,000 hours at 87 C against a 125 C life test.
- An Acceleration Model for Lead-Free (SAC) Solder Joint Reliability under Thermal Cycling, Vasu Vasudevan and Xuejun Fan, IEEE, 2008 Electronic Components and Technology Conference, 2008. The Norris-Landzberg form and its constants for tin-lead solder, and a comparison with lead-free fits.
- Reliability of Controlled Collapse Interconnections, K. C. Norris and A. H. Landzberg, IBM Journal of Research and Development, volume 13, 1969. The original thermal cycling model for solder joints.
Where it helps
The lifecycle stages this tool serves: 3 stages in 2 lifecycles.
More reliability tools
- Failure rate (FIT) calculator
- Qualification test planner
- Defect level (DPPM) estimator
- Error rate confidence calculator
- Lot sampling plan calculator
- Stack and multi-die yield calculator
- Radiation dose margin calculator
- Single-event upset rate estimator
- Component derating checker
- Mission profile builder
- Warpage and thermal mismatch estimator
Method notev1Formula, ranges, sources and limits
Works out the acceleration factor of a stress test for a chosen failure model, converts test time to equivalent use time, and finds the test time a use life needs.
Formula
AF = exp[(Ea / k) x (1 / Tuse − 1 / Tstress)]; AF = (Tstress / Tuse)^p x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] x exp[(B + C / Tstress) x Sstress − (B + C / Tuse) x Suse]; AF = (RHstress / RHuse)^n x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)] x (Vstress / Vuse)^m; Coffin-Manson: AF = (dTstress / dTuse)^q; Norris-Landzberg: AF = (dTstress / dTuse)^n x (fuse / fstress)^m x exp[(Ea / k) x (1 / Tmax,use − 1 / Tmax,stress)]; voltage: AF = exp[gamma x (Vstress − Vuse)] or (Vstress / Vuse)^n; Black: AF = (Jstress / Juse)^n x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)]; stress migration: AF = ((T0 − Tuse) / (T0 − Tstress))^(−n) x exp[(Ea / k) x (1 / Tuse − 1 / Tstress)]; AF total = AF1 x AF2 x ...; equivalent use time = AF x test time; test time needed = use life / AF
| Input | Unit | Range it holds for | Comes from |
|---|---|---|---|
| Model | choice | arrhenius, eyring, peck, coffin-manson, norris-landzberg, voltage-exp, voltage-power, black, stress-migration, combined | Your figure |
| Temperature unit | choice | C, K | Your figure |
| Use temperature | C or K | -273.15 to 1000 | Your figure |
| Stress temperature | C or K | -273.15 to 1000 | Your figure |
| Activation energy | eV | -0.5 to 1.8 | Your figure |
| Eyring prefactor exponent p | number | -2 to 4 | Your figure |
| Eyring stress term B | per stress unit | any | Your figure |
| Eyring stress term C | K per stress unit | any | Your figure |
| Second stress at use | stress units | any | Your figure |
| Second stress in test | stress units | any | Your figure |
| Use relative humidity | % | 0 to 100 | Your figure |
| Stress relative humidity | % | 0 to 100 | Your figure |
| Humidity exponent | number | 0 to 6 | Your figure |
| Peck voltage exponent | number | 0 to 4 | Your figure |
| Use temperature range | C or K | from 0 | Your figure |
| Stress temperature range | C or K | from 0 | Your figure |
| Coffin-Manson exponent | number | 0.5 to 10 | Your figure |
| Use peak temperature | C or K | -273.15 to 1000 | Your figure |
| Stress peak temperature | C or K | -273.15 to 1000 | Your figure |
| Use cycles | per day | from 0 | Your figure |
| Test cycles | per day | from 0 | Your figure |
| Norris-Landzberg range exponent | number | 0.5 to 4 | Your figure |
| Norris-Landzberg frequency exponent | number | 0 to 1 | Your figure |
| Norris-Landzberg Ea over k | K | 0 to 5000 | Your figure |
| Use voltage | V or MV/cm | from 0 | Your figure |
| Stress voltage | V or MV/cm | from 0 | Your figure |
| Voltage gamma | per V or per MV/cm | 0 to 20 | Your figure |
| Voltage exponent | number | 0 to 10 | Your figure |
| Use current density | any unit | from 0 | Your figure |
| Stress current density | same unit | from 0 | Your figure |
| Current density exponent | number | 0.5 to 4 | Your figure |
| Stress-free temperature | C or K | -273.15 to 1500 | Your figure |
| Stress migration exponent | number | 0.5 to 6 | Your figure |
| Test time | hours | from 0 | Your figure |
| Use life to cover | hours, days or years | from 0 | Your figure |
| Use life unit | choice | hours, days, years | Your figure |
Outputs: Acceleration factor (ratio); Equivalent use time for the test time (hours); Test time that covers the use life (hours).
Assumptions
- One failure mechanism is being accelerated, and the stress does not change it. If the mechanism changes, the factor does not apply.
- The constants (Ea, exponents, gamma) are properties of the mechanism, technology and package. The presets are typical published values; use your own data where you have it.
- Temperatures are junction or metal temperatures, not oven settings, unless you have added the self-heating.
- Combined factors assume the stresses act independently on one mechanism.
Uncertainty: The arithmetic is exact. The factor is only as good as the constants: for a 70 K step, an activation energy error of 0.1 eV changes it by about a factor of two.
Known limits
- A small error in Ea or an exponent grows with the size of the step: a 0.1 eV error changes a 125 C to 55 C factor by about a factor of 2. The sensitivity charts show this.
- Eyring stress terms, Peck constants and Norris-Landzberg constants are published in several forms. Check that the form and units match the source of your constants.
- Coffin-Manson and Norris-Landzberg cover the cycle range and rate, not dwell time at the extremes, which matters for creep in solder.
- The mechanism table shows the temperature factor for each preset mechanism. It leaves out voltage, humidity or current terms that some of them also need.
- Real mission profiles vary in time. Use an effective temperature, or split the profile into phases, rather than one average that hides the hot hours.
Where the method comes from
Follows JEDEC JEP122, Failure Mechanisms and Models for Semiconductor Devices (opens in a new tab). The arithmetic here is the same; the platform claims no improvement on it.
Reference example, checked by test: Monolithic Power Systems, as reported by Power Electronics News (14 May 2021): 12,000 hours at an average junction temperature of 87 C is simulated by about 1,393 hours at 125 C, Arrhenius with 0.7 eV; hand-worked in the tool's unit test. Acceleration factor 8.6079; Equivalent use time for the test time 8,607.86; Test time that covers the use life 1,394.07.
Versions
- v1 (2026-10-25): First version, built from the tool idea.
Terms on this page
- Black's equation
- The empirical model for electromigration lifetime, MTTF equal to A times current density to the power minus n times exp(Ea/kT).
- Acceleration factor
- The ratio of time to failure at use conditions to time to failure at stress conditions, derived from temperature, voltage, humidity or cycling range models.
- Stress migration
- Void formation in metal lines and under vias driven by tensile stress gradients from the mismatch in thermal expansion between metal and dielectric, with no current flowing.
- Test time
- The tester-occupied time per device, including pattern application, parametric measurements, settling and index time.
- Activation energy
- The energy barrier of a failure mechanism, in electron volts, that sets how strongly its rate depends on temperature in the Arrhenius model.
- Corrosion
- Chemical or electrochemical attack on bond pads, wires, leads or die metallisation, driven by moisture combined with ionic contaminants such as chloride.
- Electromigration
- Gradual transport of metal atoms along an interconnect caused by momentum transfer from the electron current, producing voids that open the line and hillocks that short to neighbours.
- HCI
- Wear-out in which carriers accelerated by the high lateral field near the drain gain enough energy to damage the gate dielectric interface, degrading transconductance and threshold voltage.
- Failure mechanism
- The physical, chemical or electrical process that causes a failure, such as electromigration, time-dependent dielectric breakdown, corrosion or solder fatigue.
Tell us what you're after and leave your email: we'll connect you with the right person on the Hub.