Selective Wave Soldering vs Reflow for Through-Hole — Cost, Reliability, Lead Time

The all-SMT board is the design ideal. The mixed-technology board is the production reality. Connectors, through-hole electrolytics, transformer pins, header strips, mechanical mounting posts — every one is a through-hole component that has to be soldered after the SMT line has run, or alongside it, and the decision between selective wave and intrusive reflow drives unit cost, reliability, and lead time on the rest of the build.

Three paths to soldering through-hole components on a modern SMT-heavy board:

• Intrusive reflow (pin-in-paste). The through-hole pin is inserted into a paste-filled hole and reflowed on the main SMT pass. No second soldering step.
• Selective wave soldering. A second pass through a small molten-solder fountain that targets only the through-hole pins on the bottom side.
• Hand-soldering or selective laser. Operator-level rework for low-quantity boards or specialised pin patterns.

Each has a cost profile, a reliability profile, and a design constraint set. The right choice depends on board complexity, volume, the specific through-hole components on the BOM, and your tolerance for setup-time amortisation across the run.

"The biggest mistake we see is teams choosing the soldering method after the board is laid out. The choice should drive the layout, not follow it." — Pioneer Horizon DFM engineer

This article walks through the three trade-offs that decide which method fits which programme.

Intrusive reflow is the cheapest option when it works. There's no second soldering step, no separate setup, no incremental fluxing operation. The through-hole pin is treated as just another SMT joint on the same reflow profile. For boards with a small number of through-hole components, this is unbeatable on cycle time.

What pin-in-paste requires

• Compatible components. The through-hole part has to survive the SMT reflow profile (typically 245°C peak for several seconds). Many electrolytic capacitors are not rated for this — confirm the J-STD-020 rating before committing.
• Sufficient hole-to-pin clearance. The paste needs room to deposit and the pin needs room to displace it. Typical rule: hole diameter = pin diameter + 0.3mm to 0.4mm. Too tight, and the pin pushes paste out the bottom; too loose, and there's not enough solder to form a complete fillet.
• Adequate paste volume. The hole has to be filled with enough paste to form the joint. We use a step stencil with thicker apertures over through-hole positions (typically 200µm vs 100µm for surface mount) and overprinted aperture footprint (typically 1.5× the pad area).
• Bottom-side support. A pallet or fixture that supports the bottom-side pins during paste deposition and component insertion. Without it, paste smears and pins shift.

What pin-in-paste cannot do

Some components fundamentally don't survive an SMT reflow profile or have geometries that can't accept the required paste volume. Notable cases:

• Large electrolytic cans. The thermal mass shadows surrounding SMT joints — components within 5mm of a 16mm-diameter electrolytic see a 15–20°C cooler profile and risk cold joints.
• RJ45/USB connectors with plastic housings. Most modern connectors are reflow-rated, but legacy parts often aren't. Confirm before designing.
• High-current pins. A 4A pin needs a 1.2mm hole. Filling that with paste reliably requires a 250µm stencil aperture and tight process control — borderline economic.

Yield profile on intrusive reflow

Well-engineered pin-in-paste on the right components hits 99.6–99.8% first-pass yield on the through-hole joints. Marginally engineered (paste-starved holes) drops to 96–98%, which sounds high but on a board with 40 through-hole pins translates to roughly one rework event per board.

When pin-in-paste isn't viable, selective wave is the next answer. A selective wave machine has a small (typically 3mm–8mm wide) solder fountain that's programmed to target individual through-hole pin clusters on the bottom side of the board. The board passes over the fountain on a fixture, the fountain runs through a programmed path, and only the pins that need solder get it.

What selective wave gets you

• Component flexibility. No constraint on through-hole component thermal rating beyond the wave-pot temperature exposure (typically 260–265°C for 2–4 seconds on the leads only, not the body).
• Joint reliability. Wave-soldered through-hole joints have more solder volume than pin-in-paste joints — typical fillet height 50–70% of barrel depth vs 30–50% for paste. For mechanical components carrying current or strain, this is a real reliability difference.
• Mixed pin-count flexibility. A board with 5 connectors and 80 pins handles as easily as a board with 50 connectors and 800 pins. Setup is largely fixed-cost.

What selective wave costs you

• Setup time. Each new board design requires programming the wave path, validating coverage, and confirming no SMT components are within the splash zone. Typically 2–4 hours of engineer time per new board.
• Pallet/fixture cost. A custom support fixture per board family — $300–1,200 depending on complexity.
• Cycle time. A typical selective wave pass adds 30–90 seconds per board to the build cycle. On low-volume builds, this rolls off into pallet cost; on high volume, it limits throughput.
• Design constraints. Bottom-side SMT components within 2–3mm of any wave target are at risk. The DFM rules here are stricter than they look — see our selective-wave economics article for the full constraint set.

Yield profile on selective wave

Well-engineered selective wave hits 99.7–99.9% first-pass yield on through-hole joints. Failures are dominated by bridging on tight-pitch connectors (the 2.54mm header right next to a power connector is the usual suspect) and by joint starvation on pins shadowed by tall adjacent SMT components.

The choice between intrusive reflow and selective wave breaks down into three quantitative comparisons, plus a few qualitative gates.

Number 1: Cost per board, including amortisation

1. Intrusive reflow: incremental cost is the step stencil (~$400 one-time) plus a slightly larger paste deposit. Per-board cost on a 10,000-unit run: ~$0.04 stencil amortisation, ~$0.08 incremental paste = $0.12.
2. Selective wave: fixture ($800) + setup ($300 of engineer time) + 60s of incremental cycle time. Per-board cost on a 10,000-unit run: ~$0.08 fixture amortisation, ~$0.03 setup, ~$0.35 cycle time at typical line rates = $0.46.
3. Hand-solder: ~$0.80–1.50 per through-hole pin at typical operator rates. For a board with 10 pins, that's $8–15 per board — a non-starter at volume.

At 10,000 units, intrusive reflow saves $0.34 per board over selective wave — $3,400 on the run. At 100,000 units that's $34,000. The economics push hard toward pin-in-paste whenever it's viable.

Number 2: Yield differential cost

If intrusive reflow on your specific board gives 98.5% first-pass and selective wave gives 99.7%, the yield differential is 1.2 percentage points. On a board with $40 of total assembly value, that's $0.48 of rework cost per board avoided by selective wave. The cost-vs-yield comparison now favours selective wave above ~$0.46 - $0.48 break-even point — for premium boards, lean toward selective wave.

Number 3: Field-reliability cost

For mechanical components carrying load — connectors that mate and unmate, header strips that flex during enclosure assembly — the higher fillet volume of wave-soldered joints translates to a measurable field-reliability advantage. Hard to put a number on, but on automotive-grade builds we lean toward selective wave for any connector that will be plugged more than 10 times in field use.

Lead-time impact

Intrusive reflow shortens lead time by 1–2 days on a typical run because there's no second soldering step. On a 4-week NPI cadence, that's 5% of the schedule. Don't undervalue it.

The right call usually emerges from running these three numbers. The wrong call is making the choice late — after the layout is locked, after the BOM is set, after the pallet would have been quoted differently.

The soldering-method choice doesn't sit downstream — it reaches back into BOM and layout. Get the choice right at the start and the design is clean. Get it wrong (or postpone it) and the design has to be revised, the BOM has to be re-quoted, and the cost saving evaporates.

Cascade into BOM selection

• If targeting intrusive reflow: every through-hole component on the BOM must be reflow-rated. Engineering checks every datasheet for J-STD-020 compliance. Substitutes for non-compliant parts add $0.50–3.00 per part typically.
• If targeting selective wave: components can be selected for cost or performance without thermal-rating concerns. Standard electrolytics, standard connectors — typically $0.10–2.00 cheaper per part than their reflow-rated equivalents.
• BOM hybrid possibilities: some boards mix both methods — reflow-rated through-holes on the SMT pass for the high-count low-thermal-mass parts (connectors), then selective wave for the few thermally fragile parts. Setup overhead doubles but BOM cost is optimised. Justifies itself on programmes > 25,000 units.

Cascade into layout

• Pin-in-paste layouts: through-hole pads sized to accept enlarged paste deposit (typically 1.5× pad area). Vias near pins kept tented to prevent paste suction.
• Selective wave layouts: bottom-side SMT components kept at least 2.5mm from any wave-target pin. The wave fountain has a splash radius and adjacent SMT joints will be reworked into damage if you're tight.
• Fixture-ready layouts: for selective wave, the bottom side needs flat clamping zones for the pallet. Components in those zones cause the pallet to lose seal.

Cascade into testing

Wave-soldered joints have larger fillets and more solder mass — they show up differently on AOI. The AOI algorithm needs to be tuned to expect either signature. Mixed-tech boards we ship from a customer that switched mid-design from pin-in-paste to selective wave required AOI re-tuning that cost $1,200 in engineering and 2 days of line idle.

"Pick the soldering method at DFM, not at first build. The cost of postponing the decision is paid in BOM rework, layout rework, and AOI rework — all of it avoidable." — Pioneer Horizon process engineer

If you're spinning a board with a non-trivial through-hole population and the soldering method isn't settled, walk us through the BOM — a thirty-minute DFM session will tell you which method fits and what BOM adjustments unlock the cheaper path.