Key collaborator: Tony Chetta
The challenge
The end customer handed us a very thorough test profile for the D product family. It was solid—but if we ran it “as-is,” the timeline stretched into years, which obviously didn’t match go-to-market reality.
A needed a way to collapse the schedule without hand-waving the stats. The goal was simple: use highly accelerated testing to hit the customer’s reliability and confidence targets in a fraction of the time—while keeping the results defensible.
The solution
I built a two-part toolset that covers both sides of reliability work: planning the accelerated tests and making sense of the results. Under the hood, it leans on physics-of-failure (Arrhenius, Coffin–Manson) and reliability statistics (Weibull, Success Run) to translate accelerated stress back into real-world life.
1) Reliability Growth & Analysis Tool
This is the “after the test starts” engine—data analysis + live tracking.
- Binomial sample size calculator
Quickly spits out required sample sizes for success-run targets (e.g., n = 22 for R90 / C90). - Arrhenius acceleration model
Computes acceleration factors using activation energy (Ea) for plating systems (Gold / Silver / Tin) and temperature rise from current loading (up to 70A). - Weibull life data analysis
Median Rank Regression to estimate β (shape) and η (characteristic life) from failure data. - Field equivalence
Converts test cycles into estimated years of service, so the customer can see what the stress really “buys” us.
Link: Analysis Dashboard
2) DOE & HASC Planner
This is the “before we touch hardware” planner—DOE setup + stress-cycle definition.
- HASC stress profiling
Builds interspersed stress cycles (mechanical durability + electrical/environmental ALT). - Duty cycle modeling
Converts field duty cycles into equivalent test time, so we’re not guessing test duration. - Constraint management
Hard-locks protocol rules like “no hot plugging” and current–temperature guardrails. - DOE matrix generation
Auto-generates DOE matrices (e.g., 2×2×2 Temp/Humidity/Cycle) to maximize coverage without bloating test time.
Link: DOE Planner