Key takeaways
- PCB fabrication and PCB assembly are two different industries. A board house makes the bare board; an assembly house populates it.
- A factory cannot quote from a 3D render. It needs Gerbers, a pick-and-place file, a bill of materials and assembly drawings.
- The seam between fabrication and assembly is where problems concentrate, because two companies work from the same data with different assumptions.
- Building five boards and five thousand boards are different jobs on an identical design, so prototype and production assembly are quoted differently.
PCB fabrication and PCB assembly are two separate manufacturing steps, usually done by two separate companies: a board house etches and builds the bare circuit board, and an assembly house (often called an EMS, for electronics manufacturing services) places and solders the components onto it. Confusing the two, or assuming one vendor automatically does both well, is where a lot of new hardware founders lose time and money on their first order.
This guide explains what each side actually does, the files a factory needs before it can quote your job, where the handoff between fabrication and assembly tends to go wrong, and what changes between building a handful of prototypes and running a production batch. It’s a sub-topic of the broader question of where to manufacture your product, which covers material and process choice, volume, and geography beyond electronics alone.
Fabrication and assembly are two different businesses
A PCB fabricator (a “board house”) manufactures the bare printed circuit board: layers of copper and insulating material etched into circuit patterns, drilled, plated, and finished. It ships you an empty board with no components on it.
An assembly house, often called an EMS company (electronics manufacturing services), takes that bare board and populates it: placing resistors, capacitors, connectors, and integrated circuits, then soldering everything into a working assembly. The finished product, board plus components, is called a PCBA (printed circuit board assembly).
| Business | Starts with | Delivers | Core equipment |
|---|---|---|---|
| Board house (fabrication) | Copper-clad laminate | Bare, unpopulated PCB | Etching lines, drills, plating baths |
| Assembly house (EMS) | Bare PCB + components | Populated, soldered PCBA | Stencil printer, pick-and-place, reflow oven |
Some companies do both under one roof and call themselves “turnkey” providers, buying components and handling fabrication and assembly as a single order. Others specialize in one side only. Either way, the two processes use different equipment, different skills, and different failure modes, so it helps to think of them as two industries that happen to sit next to each other in your supply chain rather than one continuous process.
What happens at a board house
Fabrication starts with a stack of copper-clad laminate, most commonly a glass-fiber-reinforced epoxy material graded FR-4 (a NEMA performance class, not a single branded material). The board house images your circuit pattern onto the copper using a photoresist process, etches away the unwanted copper, drills the holes that will carry connections between layers, and plates those holes with copper so signals can pass from one layer to another.
Layers, materials, and finishes in plain language
- Layers. A simple product might use a two-layer board (copper on top and bottom only). A denser product with many connections often needs four, six, or more internal copper layers sandwiched between insulating material.
- Copper weight. Copper thickness is specified per layer and affects how much current a trace can carry. Your design files, not a conversation with the factory, should specify this.
- Surface finish. The exposed copper pads need a finish to stay solderable and resist corrosion before assembly. Common options include ENIG (electroless nickel immersion gold, which per the industry specification IPC-4552 lays down a thin nickel layer with an even thinner gold layer over it, good for fine-pitch parts and long shelf life), immersion silver (a thinner, lower-cost finish defined in IPC-4553), OSP (organic solderability preservative, a clear organic coating that protects copper until it’s soldered, defined in IPC-4555), and HASL (hot air solder leveling, where the board is dipped in molten solder and leveled with hot air knives, an older and cheaper option that isn’t practical for very fine-pitch parts).
The formal workmanship and design standards for all of this are published by the Global Electronics Association, the trade body that renamed itself from IPC in June 2025 (the individual standard numbers, like IPC-2221 for generic printed board design, kept their IPC prefix even after the rename). If a factory or a spec sheet references an IPC standard number, that is what it is pointing to.
What happens at an assembly house
Assembly takes your bare board and turns it into a working circuit. For surface-mount technology (SMT), the dominant approach for anything built since the 1990s, components sit on pads on the board surface rather than through drilled holes. Through-hole assembly, where a component’s leads pass through the board and are soldered on the far side, is still used for parts under real mechanical stress, like connectors and larger capacitors, because the solder joint has to hold up to repeated plugging and unplugging.
Paste, placement, reflow, and inspection
- Stencil printing. A thin metal stencil, designed to the aperture guidance in IPC-7525, sits over the board and solder paste is squeegeed through openings that align with each pad.
- Placement. A pick-and-place machine reads a placement file and sets each component onto its wet paste deposit. High-speed production machines are rated for tens of thousands of placements per hour; that is a manufacturer’s theoretical maximum, not what a typical prototype run achieves.
- Reflow. The populated board travels through a reflow oven on a controlled temperature profile: a preheat and soak stage, a ramp to above the solder’s melting point, a peak dwell that actually forms the joint, then a controlled cooldown. The industry reference profile for this, tied to component moisture sensitivity, comes from the joint IPC/JEDEC standard J-STD-020.
- Through-hole and inspection. Any through-hole parts are typically added and hand- or wave-soldered afterward, and the finished assembly is checked against IPC-A-610, the standard that defines what an acceptable solder joint and component placement actually look like, at one of three product classes depending on how demanding the end use is.
The files and documents a factory needs from you
A factory can’t quote or build from a 3D render. It needs a specific set of manufacturing data files.
Gerbers, pick and place, BOM, and assembly drawings
- Gerber files. The open format almost every board house expects for describing each copper layer, solder mask, and silkscreen. The format is maintained by Ucamco, and the current version in wide use, Extended Gerber (RS-274X), replaced the original Gerber format in 1998. Newer revisions of the format can now carry some placement and component metadata directly, though most shops still ask for a separate placement file as well.
- Pick-and-place (centroid) file. A spreadsheet-style file listing every component’s reference designator, X/Y location on the board, rotation, and which side it belongs on. This is how the placement machine, or the technician programming it, knows exactly where each part goes.
- Bill of materials (BOM). A parts list: reference designator, manufacturer, manufacturer part number, quantity, and a description, matched line for line against your schematic and layout.
- Assembly drawings. Diagrams that show anything a file format can’t, like polarity markings, mechanical notes, or components that need hand assembly.
Worth separating from all of this: if your product needs to declare material content for regulatory reasons (for example REACH or China RoHS compliance), that uses a different data format, IPC-1752, for material declaration reporting. It has nothing to do with getting your board built and shouldn’t be confused with the BOM you send for manufacturing.
Where the two industries meet, and where handoffs fail
The seam between fabrication and assembly is where problems concentrate, because it’s the one point where two different companies, working from the same files, have to agree on what those files mean. A few recurring failure patterns:
- Finish mismatch. The board house builds to whatever surface finish is called out in the fabrication notes. If the assembly house expected something else, for example a finish that doesn’t hold up to a second reflow cycle, that mismatch shows up as solder defects that look like an assembly problem but actually originated at fabrication.
- Silent file drift. A last-minute layout change sent to the board house but not reflected in the pick-and-place file sent to the assembly house creates a board that is physically correct but assembled against outdated placement data.
- Undocumented assumptions. Anything not captured in a file, like which of two similar-looking connectors to use, becomes a guess made by whoever is holding the board at that stage.
None of this is exotic. It’s the ordinary cost of a design’s data passing between two organizations that didn’t design it, which is exactly why complete, unambiguous files matter more here than almost anywhere else in the process.
Prototype assembly versus production assembly
Building five boards and building five thousand boards are different jobs, even on an identical design.
| Prototype assembly | Production assembly | |
|---|---|---|
| Quantity | Single units to low tens | Hundreds to thousands or more |
| Process | Often partly manual, hybrid with automated placement for the fine-pitch parts | Fully automated lines, optimized for one job running continuously |
| Component sourcing | Frequently sourced individually per build | Procured against a production forecast, often with better per-unit pricing |
| Priority | Speed and flexibility to support design changes | Repeatability and cost per unit |
A prototype build proves the design can be assembled and function. It does not, on its own, prove the design can be produced consistently at volume. Whether a “production-ready” claim actually means design files finalized for manufacturing, or something closer to a working sample, is worth asking directly. The deliverables behind the label matter more than the label itself.
What to look for in an assembly partner
- Ask what happens when a component isn’t available. A partner who avoids relying on a single vendor for critical parts, and checks availability across multiple suppliers before locking a design, is managing a real and common risk rather than hoping it doesn’t happen.
- Ask how first-sample review works. A factory that reviews initial samples against your files and flags discrepancies before scaling up a run is catching problems while they’re still cheap to fix.
- Verify certifications independently rather than taking a claimed credential at face value; our guide to verifying a manufacturer’s certifications covers how.
- Check references the right way. A portfolio of finished-looking boards proves less than a reference who can describe what a working relationship with that vendor is actually like, which is why how to check a development partner’s references is worth reading before you commit to one.
- Ask the questions that reveal how a firm actually operates before signing anything; see our list of questions to ask before hiring a hardware development company.
How this connects to your enclosure and final assembly
PCB and enclosure design are usually developed together rather than in sequence, because the board’s dimensions, connector placement, and mounting points constrain what the enclosure can look like, and the enclosure’s internal space constrains how the board can be laid out. Getting the board stable before finalizing the enclosure, rather than the reverse, tends to avoid late redesign of both. During prototyping, that enclosure is commonly 3D printed for speed; our guide to how 3D printing changed product development covers where that fits and where it stops being the right process. Once the design is validated, moving to injection molding or another production process is a separate decision, covered in our guide to plastic part manufacturing methods. Final assembly, where the populated board goes into its enclosure alongside a battery, display, or other mechanical parts, is a separate step again, with its own tolerance and access requirements that are worth planning for before the PCB layout is frozen, not after.
Frequently asked questions
Is a PCB fabricator the same company as a PCB assembler?
Sometimes, but not by default. Many shops specialize in one side only. Turnkey providers handle both fabrication and assembly (and often component sourcing) as a single order, but that’s a business choice a vendor makes, not something implied by the words “PCB manufacturer.”
What is the difference between SMT and through-hole assembly?
SMT mounts components directly onto pads on the board surface, which allows higher density and faster automated placement. Through-hole passes component leads through drilled holes for a mechanically stronger joint, and is still used for connectors and other parts that take physical stress in use.
What files does a PCB assembly house need?
At minimum: Gerber files for the bare board, a pick-and-place (centroid) file, a bill of materials matched to your schematic, and assembly drawings covering anything the files alone don’t make clear, like polarity or hand-assembly notes.
What does IPC-A-610 actually cover?
It’s the industry’s visual acceptance standard for finished electronic assemblies: what counts as an acceptable solder joint, component placement, and termination, organized into three product classes depending on how demanding the end application is.
Can the same design go straight from a prototype assembly run to full production?
Not automatically. A prototype run proves the design can be built and can function. Moving to production usually still needs its own manufacturing checks, since component sourcing, process automation, and consistency requirements are different at volume.
What is the difference between a Gerber file and a BOM?
A Gerber file describes the physical board itself, copper layers, solder mask, and silkscreen. A bill of materials lists the components that go onto it. A factory needs both, along with a pick-and-place file, to build a working assembly.
Where Inventornest fits
We handle product design and development in-house, which includes designing the electronic schematics, firmware, and PCB before moving into enclosure and mechanical design so the two inform each other rather than being finalized separately. When production is handled by our manufacturing partners, we retain responsibility for engineering coordination and design clarification: we clarify questions with the factory, review initial samples, and address discrepancies before a design scales into a production run. In our experience, avoiding reliance on a single vendor for a given component, and checking availability across multiple suppliers, is a sourcing discipline worth building in early rather than discovering the need for it mid-build. External laboratory certification and specialist patent-attorney work remain separate professional activities outside this process. If you’re weighing whether your product is ready to move from a working prototype into a manufacturable design, our OEM services page covers how we structure that engagement, or you can get a quote directly.
