In short: Surface-mount technology (SMT) wins on density, speed and automated assembly cost, while through-hole (THT) wins on mechanical strength and high-current or high-stress connections, so most real boards are mixed-technology: SMD for the bulk, through-hole for connectors and power parts.
Two ways of attaching a part
A through-hole component has leads that pass through drilled, plated holes and are soldered on the opposite face. A surface-mount device (SMD) sits on copper pads and is soldered on the same face it is placed on. That single difference drives almost every trade-off in this article: hole count, routing freedom, assembly process, repair method and mechanical behaviour.
If you are new to the board itself, start with what a PCB is, then come back. For the SMD side of the story, our companion piece on SMD component types and packages covers 0402, SOIC, QFN and BGA in detail.
Density: why SMT took over
A drilled hole blocks every routing layer it passes through, and a through-hole pin needs a pad plus clearance on both faces. Typical through-hole pitch is 2.54 mm (0.1 in). Common SMD pitches are 1.27 mm for SOIC, 0.5 mm for QFP and 0.8 mm down to 0.4 mm for BGA. A 0402 resistor measures 1.0 x 0.5 mm, while an axial resistor with its lead bends occupies roughly 10 mm of board length.
SMD parts can also be placed on both sides of the board, effectively doubling usable area. A 100-pin QFP at 0.5 mm pitch has a footprint of about 14 x 14 mm; a DIP with 100 pins at 2.54 mm would not realistically exist. The routing consequence is large: blind and buried via types and microvias only make sense when component pads are on the surface and not occupying every layer with holes.
Key fact: Moving a mid-complexity design from through-hole to SMT routinely shrinks board area by 50 to 70 percent, and that is before using both sides of the board.
Mechanical strength and thermal behaviour
This is where through-hole earns its place. A lead soldered inside a plated barrel resists pull, shear and torque far better than a joint resting on a pad. A USB-C receptacle plugged and unplugged thousands of times, a screw terminal, or a heavy electrolytic capacitor on a vibrating motor drive all stress the joint mechanically. SMD connectors therefore add through-hole shield tabs or mounting posts for exactly this reason.
Current is the second factor. A plated barrel gives a large copper cross-section and good heat transfer into the planes, which is why power diodes, TO-220 regulators and transformers stay through-hole. Pad size and copper width, covered in trace width and current capacity, define what an SMD pad can carry.
SMD parts have their own failure mode: thermal mismatch. A ceramic BGA (CTE near 6 ppm/°C) soldered to FR-4 (about 14 to 17 ppm/°C in-plane) cycles in temperature and fatigues the solder balls over time. Big MLCC capacitors can crack when the board flexes. Through-hole leads absorb this strain in the lead itself.
Assembly cost: volume decides
SMT assembly is almost fully automated: solder paste is stencil-printed, a pick-and-place machine places 20,000 to 100,000 or more components per hour, and the board goes through a reflow oven with a profile peaking near 235 to 250 °C for lead-free SAC305. Through-hole assembly needs insertion (automatic for axial and radial parts, often manual for odd shapes) and then wave soldering or selective soldering.
- High volume: SMT is far cheaper per placement and needs no drilling for component holes, which cuts drill time and bit wear.
- Prototype and one-off: hand-soldering through-hole parts needs no stencil and no oven, so tiny runs can favour it.
- Tooling: SMT needs a stencil, feeders and reflow profiling; BGAs add X-ray inspection.
Rework and repair
A through-hole part can be removed with a desoldering pump or braid and a soldering iron, and the plated barrel usually survives several cycles. SMD rework is more demanding: small passives need hot tweezers or hot air; QFNs and BGAs need a rework station with controlled preheat, a nozzle sized to the package, and reballing for BGAs. Overheating lifts pads, so a limit of two or three rework cycles per site is a common rule in quality plans. Moisture matters too: a plastic BGA baked per its MSL rating before rework avoids popcorning.
Key fact: Check the moisture sensitivity level on the part label. An MSL 3 device exposed beyond its 168-hour floor life must be baked before reflow or rework.
When mixed-technology boards are the right answer
Pure-SMT boards are common in phones, but industrial and automotive designs usually mix technologies. Typical split: the processor, memory, regulators, passives and RF go SMD; the power connector, relay, transformer, large electrolytics and board-to-board headers stay through-hole. A mixed board adds process steps. Common scheme: reflow the top side, glue or reflow the bottom, then selective-solder the through-hole parts. Pin-in-paste reflow can avoid the wave step by printing paste into the holes.
Tooling helps you manage this. Cadence Allegro, Siemens Xpedition and Altium all let you define separate footprints, hole sizes and assembly side, and DFM checkers such as Siemens Valor flag paste, spacing and wave-shadow problems. Footprint sizing follows IPC-7351, and finished hole size is typically lead diameter plus 0.2 to 0.3 mm.
Practical design rules
- Choose SMD by default; justify every through-hole part with a mechanical or current reason.
- Keep through-hole parts on one side so the wave or selective step touches one face.
- Use thermal relief on barrel pads tied to planes; see thermal reliefs and teardrops for why solder wicking fails without them.
- Check via aspect ratio and plating rules, since thick boards with small holes plate poorly.
- High-speed layouts benefit from SMD's short stubs; see signal integrity in PCB.
Teams that need production-ready layouts can work with Avecas design services, and learners can pair this topic with a stackup study of single vs multilayer boards.
Want to design boards hands-on? Explore our PCB design track in the CourseTron courses and build a mixed-technology board from schematic to Gerber.
FAQ
Is through-hole obsolete?
No. It remains standard for connectors, power components, transformers and anything facing repeated mechanical stress, and many component families are still offered only in through-hole form.
Which is cheaper, SMT or through-hole?
At volume SMT is cheaper because placement and reflow are automated and drilling is reduced. For a handful of hand-built prototypes, through-hole can be quicker and cheaper.
Can you mix SMD and through-hole on one board?
Yes. Reflow the SMD parts first, then use selective soldering, wave soldering or pin-in-paste for the leaded parts.
Why are SMD joints weaker?
The joint is only a fillet on a pad, so it carries mechanical load through solder alone, whereas a through-hole lead is also held by the plated barrel.