Through-hole on a Mostly-SMT Board — Selective Wave vs Hand-Solder Economics

"How many through-hole parts before we should switch from hand-soldering to selective wave?" is the most common question we get from hardware teams during DFM review. The textbook answer — "around 8 joints per board" — is wrong for our floor, our labour rates, and most of the BOMs we see. Real-world breakeven is between 25 and 60 joints per board, depending on several variables that aren't in the textbook.

Below is the actual cost model we use to make this call on every mixed-tech board.

What goes into hand-solder cost per joint

• Operator-time per joint — 8–18 seconds for a typical 0.8mm pin into 1.6mm board, depending on access, pad geometry, and operator grade.
• Loaded labour rate — ₹420/hour for a Class 3 hand-solder operator on our floor (skill premium, training, certifications, supervision).
• Inspection time — every hand joint AOI-checked plus optional cross-section on first article. Adds ~3 seconds per joint amortised.
• Rework rate — typical 1.5–2.5% on hand-soldered joints, each rework costs ~90 seconds.

What goes into selective wave cost per joint

• Setup and tooling — ~30 minutes per program plus a per-build solder-pot purge.
• Machine-time per joint — 1.5–3 seconds in the nozzle dwell.
• Operator-supervision rate — one operator per two machines, lower skill grade than hand-solder cell.
• Per-build flux and nitrogen consumption — ~₹180/board on our setup.

"Selective wave is fixed-cost-heavy and variable-cost-light. Hand solder is the opposite. The breakeven is whichever side of the curve your build volume × joint count puts you on." — Pioneer Horizon floor supervisor

Take a representative industrial control board: 12 THT joints (a 6-pin connector, a 2-pin terminal block, two 2-pin relays). Build quantity 500.

Hand solder cost

• 12 joints × 12 seconds avg = 144 seconds per board.
• + 36 seconds inspection.
• + 2% rework × 90 seconds = 1.8 seconds amortised.
• Total per board: ~181 seconds = 0.050 hours × ₹420 = ₹21 per board.
• 500 boards × ₹21 = ₹10,500 total labour.

Selective wave cost

• Setup: 30 minutes × ₹600/hr (machine + operator) = ₹300 one-time.
• 12 joints × 2.2 seconds = 26 seconds per board cycle.
• + ₹180/board consumables.
• Total per board: 26 sec × ₹600/3600 + ₹180 = ₹184 per board.
• 500 boards × ₹184 + ₹300 setup = ₹92,300 total.

At 12 joints × 500 boards, hand-solder wins by a wide margin.

Now flip the variables

Same build, 60 THT joints (more typical for legacy industrial designs):

• Hand solder: 60 × 12 = 720 sec + inspection + rework ≈ 800 sec/board × ₹420/3600 = ₹93/board × 500 = ₹46,500.
• Selective wave: 60 × 2.2 = 132 sec/board cycle + ₹180 consumables = ₹202/board × 500 + setup = ₹101,300.

Still hand-solder by a fair margin at 500 units. But push the build to 3,000:

• Hand solder: 3,000 × ₹93 = ₹279,000.
• Selective wave: 3,000 × ₹202 + ₹300 = ₹606,300.

Hmm — still hand-solder wins. So why does anyone selective-wave anything?

The simple model uses Indian labour rates and ignores three factors that flip the calculation in many real builds.

1. Joint quality consistency

Hand solder rework rate is 1.5–2.5% on average — but it's heavily operator-dependent and concentrates on hard-access joints. The standard deviation of joint reliability across operators is wider than across a single calibrated wave nozzle. For Class 3 product on critical applications, the consistency of the wave wins, even when the mean labour cost is similar.

2. Throughput and calendar time

A 6-operator hand-solder cell tops out at roughly 80 boards/hour for a 12-joint board. A selective wave running the same board does 140–180/hour. If your customer needs 5,000 units in two weeks and you have one shift of hand-solder capacity, you don't have a labour-cost question — you have a calendar question.

3. Indirect cost of operator scarcity

Class 3 hand-solder operators with current certifications are scarce. Recruiting and certifying a new operator takes 4–6 months and ₹40,000+. Backfilling for absences eats supervisor time. The wave machine doesn't call in sick.

When hand-solder still wins

• Prototype builds (the setup amortisation kills wave).
• Boards with under 15 joints and build volumes under 2,000.
• Mixed THT types where the wave would need multiple nozzle changes.
• Components that are temperature-sensitive on adjacent parts (already-mounted electrolytics, etc).

A selective-wave-friendly design recovers its cost premium fast. A selective-wave-hostile design throws the economics back to hand solder regardless of joint count. Five rules we enforce in DFM review:

1. Courtyard around every wave joint

Minimum 4mm clearance from any nozzle path to the nearest SMT component. Tall SMT components (electrolytics, tantalums, inductors above 5mm) need 6mm. Without this, you get splash damage on the SMT side and nozzle deflection.

2. Component orientation

For 2-pin connectors and discretes, pin orientation parallel to the wave direction is mandatory. Perpendicular pins drag solder and bridge. We re-orient on every layout we touch.

3. Solder thieves

Trailing pads at the end of a long pin row catch excess solder and prevent bridging on the last pin. Required for any connector with more than 8 pins in a row. Cost: 0.05 cm² of board real estate. Benefit: zero bridging on that connector.

4. Through-hole pad geometry

• Annular ring 0.3mm minimum on outer layers.
• Pad diameter at least pin + 0.6mm for wave; pin + 0.4mm is too tight and floods the topside.
• Tear-drops on traces meeting wave pads — survival of thermal cycling and prevents trace lift.

5. Panelisation

Selective wave nozzles can't reach within 5mm of a panel edge or V-groove. Components in that zone go to hand-solder regardless. We plan panel break-out lines around THT zones during DFM, not after.

For the difference between selective and full wave, our selective-wave-vs-reflow comparison covers the process-level distinction.

For any new mixed-tech board crossing our DFM desk, this is the order of questions we ask, in order, to land on a process recommendation.

1. Build volume? Under 200 boards lifetime — hand-solder by default. Setup amortisation kills any wave argument.
2. Joint count per board? Under 15 — hand-solder unless points 4 or 5 below apply.
3. Class 3 or commercial? Class 3 on a build above 25 joints and volume above 1,000 — selective wave for consistency, period.
4. Calendar pressure? If your delivery curve demands more than 80 boards/hour through the THT process, wave wins regardless of cost math.
5. Operator availability? If we don't have a hand-solder cell free for the build window, wave is the only option that meets your date.
6. DFM-compatible layout? If the design fails wave courtyards, fix the layout or accept hand-solder.

Hybrid approach

For ~30% of mixed builds, the right answer is hybrid: wave 80% of the joints, hand-solder the remaining 20% where wave can't reach or shouldn't go (next to heat-sensitive parts, inside enclosures, under shielding cans, at panel edges). The hybrid is not a compromise; it's the rational allocation. Our MES handles the routing automatically once the DFM call is made.

What this looks like on your quote

You'll see selective-wave costs broken out as a per-board line on the quote, not buried in the assembly NRE. Hand-solder shows as labour minutes. You can see exactly where the line is being drawn, and we'll walk through the trade-off if you want a different mix.

For a quick second opinion on whether your THT layout is wave-friendly, share the Gerbers — we'll come back with a courtyard map and the cost-breakeven for your volume in two working days.