Gate readiness checklist (EVT, DVT, PVT)
For every gate
- The decision is based on evidence (test results, measurements), not opinion.
- Every open issue has an owner, a plan, a date and an estimate of what it costs if deferred to the next phase.
- No critical issue is open.
- "We feel behind" and "we are genuinely ready" have been discussed as separate questions.
- The cost and schedule of holding the gate (for example a two-week extension) is written next to the cost of passing it.
- The decision, who made it and why is recorded.
EVT gate · does the design work?
- Every issue found has been triaged to the team that owns it.
- Interface issues (problems between two teams) are identified and have both owners.
- A cross-team review has been held to surface the "that is not my problem" issues.
- Critical issues are fixed and re-tested on hardware.
- High issues have a plan that an owner stands behind.
- Requirements are verified or have a dated plan to be verified.
DVT gate · does it pass?
- Tooling was released on evidence and the evidence is on file.
- Certification and reliability tests are complete, or their remaining slots are booked and the risk is stated.
- Action items are closed on hardware, not just in the tracker.
- The sample run was built with production-intent parts and process.
- Any change since certification has been checked for a re-test.
- Sourcing is confirmed for volume, with single-source parts listed and mitigated.
PVT gate · can we build it?
- The readiness check (line, fixtures, parts, trained operators) was passed before the pilot started.
- The pilot run met target yield, with defects and rework recorded.
- Stress testing (such as HALT) has shown where the weak points are, and the limits are understood.
- Process validation is done for every critical step, not just the ones that were easy to measure.
- Open issues are quantified in euros and weeks at production prices.
- The factory has signed off its own process, in writing.
Why: the book's worked example is a team that passes the DVT gate with two open issues that are "probably fine", declines a two-week extension, and pays five times as much in PVT, where there is no buffer left. The extension is the economically rational choice.
From The Hardest Hardware Lessons · thehardesthardwarelessons.com · Free to use and share with attribution.
One hardware lesson a week, free
Short, blunt lessons from the book: the mistake, what it costs, what to do instead. No spam, unsubscribe any time.
Want the reasoning behind every line?
The book explains each phase, each gate and the mistakes these checklists are built to catch.
Get the book →Play the games