Consumer Electronics Manufacturing: EVT, DVT, PVT and What Units Really Cost

A working prototype is roughly a third of the way to a shippable consumer electronic product. The rest is a disciplined sequence of pilot builds, certification and process control that turns one hand-assembled unit into ten thousand identical ones. This guide walks the stages, the money and the four failure modes that push launch dates.

Per-unit cost stacks up from PCBA, enclosure with amortized tooling, assembly labour, test time, packaging, freight and duty. A useful rule: if your landed cost is not roughly a quarter of intended retail, the model will not survive retail margins and returns.

Any product with a radio needs FCC in the United States and CE or UKCA in Europe, and labs book out weeks ahead. Lithium batteries add UN38.3 transport testing before any air freight.

The trap is sequencing: teams change an antenna, a plastic wall thickness or a supplier after testing and unknowingly invalidate the report. Freeze the RF-relevant design at DVT and book the lab slot then.

Pair this with design for manufacturing work so the frozen design is also the cheap one.

Every consumer electronics program needs three test stations defined before DVT: in-circuit or flying-probe test on the bare board, functional test after assembly, and a short burn-in or final quality gate before packing. The reason is economics rather than diligence.

A fault caught at board level costs a few dollars to rework; the same fault caught after the enclosure is ultrasonically welded costs the whole unit. Budget $15k-$60k per fixture, ask the contract manufacturer to quote test time per unit in seconds, and insist that every station writes a serial-linked record. Without that data the first field failure turns into a guess about which lot to recall.

Radio and safety certification has to be scheduled against a design freeze, because the samples tested must match what ships. For a wireless consumer product sold in the United States and Europe, expect FCC and CE radio testing, EMC, and any battery transport testing such as UN 38.3 for lithium cells.

Typical cost lands between $15k and $60k depending on the number of radios and markets, with four to ten weeks in the lab and a retest fee whenever the antenna, enclosure or firmware transmit behaviour changes.

The failure pattern is always the same: a last-minute mechanical change that alters the antenna environment, discovered after the certificate was issued.

The factory decision drives more of your outcome than any single engineering choice. A CM that is too large will not staff your program; one that is too small cannot absorb a component shortage or hold your line during a peak season. The practical target is a factory where your annual volume is meaningful but not overwhelming — roughly one to ten percent of a line's capacity.

Most disputes we see trace back to a missing document rather than a bad factory. Write the quality agreement while the relationship is friendly — before the first shipment — and it becomes the reference that keeps it friendly.

Most consumer electronics programs slip because a team treats PVT as a scheduling milestone rather than an evidence gate. PVT is the build that proves the factory - not the engineering team - can produce the product with its own operators, its own fixtures and its own documentation. If any of the items below are still open when PVT starts, the run measures your paperwork rather than your process.

Teams tend to negotiate assembly labour, which is usually the smallest lever on the sheet. The savings that matter come from part count, tooling amortisation and yield. Cutting a two-piece enclosure to one moulded part with a snap-fit removes a tool, an assembly station and a screw line item at the same time. Lifting first-pass yield from 88 to 96 percent removes rework labour and scrap from every unit shipped for the life of the product.

Two quotes for the same product are rarely comparable until you normalise them. Ask every contract manufacturer to break the number into PCBA, mechanical parts, tooling, assembly labour, test time, packaging, freight and duty, and to state the volume assumption behind each line.

Then ask what is not included - certification management, engineering change orders, NRE for fixtures, and the cost of a failed lot - because those are the lines that turn a $38 quote into a $46 landed cost.

Programs that hold these three disciplines - a real readiness gate, cost levers picked before tooling is cut, and quotes normalised to the same assumptions - land inside ten percent of their target unit cost. Programs that skip them usually discover the gap after the first container has already shipped.