A FIT rate (Failure In Time) is a measure of how often a component is expected to fail, expressed as the number of failures per one billion hours of operation. It’s the standard reliability metric used across the electronics and semiconductor industries. A component with a FIT rate of 100, for example, would be expected to produce 100 failures if one billion hours of total operating time were accumulated across all units in the field.
How FIT Rate Works
The “billion hours” unit sounds enormous, but it makes more sense when you consider how reliability engineers actually think about failure. A single chip might run for decades without failing, but a company could ship millions of identical chips. Those millions of devices running simultaneously rack up billions of collective operating hours quickly. FIT gives engineers a way to predict how many failures they’ll see across an entire population of parts.
If you deploy 10 million devices, each running 24 hours a day, they accumulate about 87.6 billion device-hours per year. A component rated at 50 FIT would be expected to produce roughly 4,380 failures across that fleet in a year. For a single device, 50 FIT translates to an astronomically long expected lifespan, but at scale, those rare failures add up.
FIT Rate and MTBF
FIT rate is directly related to Mean Time Between Failures (MTBF), which expresses reliability as the average number of hours a single device will run before failing. The conversion is straightforward:
- MTBF = 1,000,000,000 / FIT rate (result in hours)
- FIT rate = 1,000,000,000 / MTBF (result in FIT)
A component with a FIT rate of 100 has an MTBF of 10 million hours, or roughly 1,140 years. That doesn’t mean every unit will last 1,140 years. It means that across a large population, the average time between individual failures works out to that figure. Both metrics describe the same underlying reliability, just from different angles. FIT is more convenient when comparing components or adding up failure contributions in a system, while MTBF is more intuitive when thinking about a single device.
Where FIT Rates Come From
Manufacturers can’t simply run a chip for a billion hours and count failures. Instead, they use accelerated life testing. The basic idea is to stress components at elevated temperatures (and sometimes higher voltages) to force failures that would normally take years to appear, then mathematically project those results back to normal operating conditions.
For semiconductors, this projection relies on the Arrhenius model, which describes how chemical and physical degradation processes speed up with temperature. In a typical test, batches of parts are aged at several high temperatures. Engineers record when failures occur, fit those failure times to a statistical distribution (usually lognormal for semiconductors), and then calculate the activation energy of the failure mechanism. That activation energy tells them exactly how much the high temperature accelerated the failures compared to real-world conditions, allowing them to extrapolate an expected failure rate at normal operating temperatures.
Industry standards from organizations like JEDEC define exactly how these tests should be conducted and analyzed, which is why FIT rates from different manufacturers are broadly comparable.
The Bathtub Curve and When FIT Applies
FIT rates assume a constant failure rate over time, which is only true during one phase of a component’s life. Reliability engineers describe a component’s full lifespan using what’s called the bathtub curve, named for its shape when you plot failure rate against time.
The curve has three phases. First is the infant mortality period, where failure rates are elevated because of manufacturing defects, weak solder joints, or contamination. These early failures drop off quickly. Next comes the useful life period, a long, flat stretch where failures occur randomly at a low, roughly constant rate. This is the phase where FIT rates apply and where most components spend the vast majority of their operational life. Finally, there’s the wear-out period, where failure rates climb as materials degrade, metal interconnects thin out, or insulating layers break down.
Manufacturers use burn-in testing (running parts at high stress for a short time before shipping) to weed out infant mortality failures. The FIT rate you see on a datasheet represents the flat, constant portion of the curve.
Typical FIT Values for Electronics
FIT rates vary widely depending on the complexity and type of component. Simple passive components like resistors can have FIT rates in the single digits or low tens. Capacitors tend to be somewhat higher, depending on the type. Integrated circuits are more complex and typically carry higher FIT rates, ranging from the low hundreds for simple logic chips to several hundred for complex microprocessors or memory devices.
These numbers matter most when you’re designing a system with many components. A circuit board with 500 parts, each rated at 50 FIT, has a combined board-level FIT rate of 25,000, or an MTBF of about 40,000 hours (roughly 4.5 years). Reliability engineers use FIT rates to identify which components contribute the most failure risk and to predict the overall reliability of a finished product. This system-level calculation is one of the main reasons FIT exists as a standardized unit: it lets you simply add up the FIT rates of individual parts to estimate the total.