EVT, DVT and PVT sound like ceremony borrowed from large consumer-electronics companies. They are more useful than that: three builds, each answering one question, each with an exit criterion you can write down before you start. Skip a question and you don’t save the build. You just answer it later, at a worse moment and a higher price.
Three builds, three questions
| Build | The question | Parts made how | Typical tests | You exit when |
|---|---|---|---|---|
| EVT: engineering validation | Does this design meet its requirements? | Soft tools are fine; intended materials and processes | Function, power, signal quality, thermal, EMI pre-scan | One production-worthy configuration meets every requirement |
| DVT: design validation | Does it hold up when made the production way? | Hard tools or production-capable processes | Reliability, environmental, mechanical, EMC and safety certification | Yield-limiting issues have corrective actions you trust |
| PVT: production validation | Can the line build it at rate and yield? | Production tooling, fixtures and line | Line yield, line speed, operator training | Production yields at production speed on at least one line |
The exit criteria and tooling expectations follow Instrumental’s build handbook; the test lists follow the standard EVT and DVT definitions. In a mass-market program, the same handbook puts these builds at roughly 100 to 1,000 units for EVT, 300 to 2,000 for DVT and 1,000 to 20,000 for PVT. A startup’s builds may be far smaller. The questions don’t shrink with them.
Before EVT: make the prototype repeatable
A prototype that works on your bench is not an EVT unit until someone else could build another one. When we converted an off-the-shelf drill into a precision soil core sampler, we documented the teardown step by step and delivered ten prototype units with an assembly manual, fabrication illustrations and wiring diagrams included. That documentation is what turns “the prototype” into “a build”.
Have ready for EVT:
- A requirements list with a number and a measurement method for each line
- A BOM with a sourcing status for every part, and alternates for single-source parts
- A test plan covering function, power, thermal behavior and an EMI pre-scan
DVT: freeze the design, then test it hard
DVT is where the design stops changing. Parts come off hard tools or production-capable processes, and the first hard tool for each part is qualified here. It is also where certification belongs. EMC testing and safety certification are DVT activities, and under FCC rules, changing a radio’s frequency-determining circuitry or maximum power after certification requires a new grant. Certify the design you will ship, not the one you are still debugging.
Have ready for DVT:
- Frozen schematics, layout and CAD under revision control
- Tooling ordered, with time in the plan for at least one round of corrections
- A certification lab booked, and your own pre-scan results in hand
- Reliability and environmental test plans with written pass/fail limits
PVT: prove the line, not the product
PVT units are built on the production line at production speed, and in the handbook’s words, “all units are intended to be sold to customers.” What is under test now is the process: fixtures, work instructions, operator training and yield.
Process equipment is part of the product. On a modular smart-home program, prefabricated mechanical, electrical and plumbing panels were assembled and quality-checked in custom factory jigs with purpose-built test equipment. For a pandemic-era respirator helmet, we commissioned a custom vacuum-forming machine, with molds and tooling, to support a small-batch assembly line.
Have ready for PVT:
- Test fixtures and a factory-test firmware image
- Work instructions an operator can follow without an engineer standing by
- Yield tracking by station, with a short list of named defect types
- A golden sample, and written limits for what it represents
The shortcut that isn’t
The tempting move is to merge builds: take an EVT-maturity design, give it DVT goals, and hope. The handbook calls this “XVT” and is blunt about it: investing DVT-scale resources into an EVT-maturity design “does not get your product out faster.” Merge stages only when a question is genuinely already answered, for example a module and enclosure carried over unchanged from a product you already ship.
What to do this week
- Write the exit criterion for your next build in one sentence, before you order parts for it.
- Mark every BOM line as off-the-shelf, soft-tooled or hard-tooled, so you know which build each part is actually ready for.
- Call a certification lab now, even if testing is months away. Their schedule is part of your DVT schedule.
See where you stand with the production readiness score. If you want a second pair of eyes on your build plan, send us a brief.