PHASE 07ASSEMBLY
Hand-build the board in the right order. Sequence is everything — the wrong order lifts parts you've already placed.
Assembly rewards patience and a plan. The parts go down in a deliberate order, every joint gets flux, and you inspect the board before you ever apply power. Rush the order and you'll spend longer reworking than you saved.
The build order
Read this before you melt anything — the order is the whole game.
Order of operations
Hot-air the hard parts first, iron the rest after. Do it the other way and you knock off what you already placed.
Do the fine-pitch, heat-hungry parts first on the bare board — U1, the module, and J1, the connector — with hot air or paste-and-. Then iron-solder the passives and small discretes. Finally, fit the through-hole parts (switches, headers, test points). Work the other way and the hot-air rework for U1/J1 blows your freshly-placed 0805s right off the board.
CheckWhy solder the WROOM module before the 0805 resistors?Show
Hot-air rework near already-placed passives blows them off — do the hot-air parts first, irons after.
▸Deep dive· Why the heavy parts go down first
U1 and J1 are the heat-hungry parts: the module is a big slab with many pads (several hidden underneath), and the USB-C connector has chunky retention tabs that drain heat away. To solder them you flood the whole area with hot air or run the board through — heat that radiates several millimeters in every direction. If the little 0805 passives are already sitting nearby, that same heat remelts their joints, and the airflow can tumble them off (or stand one up on end — ). So you place the hard, heat-hungry parts onto the bare board first, then iron the passives one at a time afterward, where the heat stays local and nothing you've already placed gets cooked twice.
Solder it, heavy parts first
Bench time — hot-air the module and connector first, then iron the rest.
The hot-air gun is the same hazard moving — 300 °C+ air burns skin and scorches anything in the blast path, so mind where it points whenever it leaves the work. A soldering iron sits at ~340 °C and never looks hot — it burns instantly. Work somewhere ventilated (flux fumes irritate), wear eye protection (hot flux can spit), and put the iron back in its stand the moment it leaves your hand. Tin the tip on a damp sponge, and treat the board, tweezers, and parts as hot for a while after. Lead-free or not, wash your hands when you're done.
The big, many-pad parts go down first — before any passives crowd them.
- Lay a thin line of solder paste along U1's and J1's pads — a small dab per pad is plenty (it spreads as it melts). A stencil makes this neat, but a steady hand works.
- Set each part in place, pads roughly aligned — it only has to be close; molten paste pulls it true by surface tension. Tweezers, gentle.
- Reflow with hot air: ~300–350 °C, medium airflow, small circles over the part until the paste flashes from grey to shiny and the part settles flat. Keep the nozzle moving so you never cook one spot.
- Let it cool untouched (nudging a molten joint ruins it), then inspect: every pad shiny and slightly concave, no bridges between pins. Re-touch any dull or bridged joint with a little flux and a quick pass.
Iron the passives, then the through-hole parts
U1 and J1 are down — now the iron does the rest. Passives one at a time, through-hole last.
- Tin one pad: melt a small bead of solder onto ONE pad of the footprint. A low mound is plenty.
- Grip the part in tweezers, reheat the tinned pad, and slide the part's end into the molten solder. Hold still a second while it freezes — the part should sit flat, not perched on a blob.
- Solder the other pad: touch the iron to pad and part-end together, feed in a little solder, lift. Re-touch the first joint with a dab of flux if it looks dull.
- Repeat, one passive at a time — the heat stays local, which is the whole reason these go down after the hot-air parts.
- Through-hole last (switches, headers, test points): push the part fully through, solder ONE pin, check it sits flat against the board, then do the rest. For header rows: one pin at each end first, recheck flatness, then fill in.
CheckA resistor sits tilted up on a solder blob instead of flat. Fix?Show
Reheat the tinned pad while pressing the part gently flat with tweezers — don't add more solder.
Flux and drag-soldering
Flux is the difference between a bridge and a clean joint.
If a reflowed row came out bridged — or you have no hot-air station and are iron-soldering U1's and J1's rows directly — this is the technique. Flood the pads with flux, then the fine-pitch rows: load the iron tip with fresh solder and drag it steadily along the row, letting surface tension and flux pull just the right amount onto each lead while clearing the bridges. On a lead-free board you're working in , which wants a slightly hotter tip and gives a more matte joint.
- Set the iron around 340 °C (SAC305 is lead-free — it wants a little more heat) and flood the pad row with liquid flux.
- Load the tip with a small bead of fresh solder.
- Rest the bead at one end of the row and drag steadily along it at about 3 mm/sec — let surface tension pull just enough onto each lead.
- Lift at the end. If a bridge stays behind, add more flux and drag once more — don't add solder.
- Inspect: each lead should show a shiny, slightly concave fillet, and no bridges between pins.
CheckYour drag pass leaves a bridge between two pins. First move?Show
More flux and a clean dragged pass — flux lets surface tension pull the excess off; you rarely need wick for a small bridge.
▸Deep dive· Why dragging molten solder doesn't bridge every pin
It feels like dragging a bead of molten metal across a row of pins should short them all together — flux is what makes it not. Liquid flux strips the oxide off the copper and lowers the solder's surface tension, so molten solder wets clean metal eagerly but beads up and refuses to stick to the between pads. Drag a loaded tip along the row and surface tension pulls just enough solder onto each lead while the excess rides along; any bridge that forms gets reflowed and pulled apart by that same tension. Run out of flux and the magic stops — the oxide creeps back and solder clumps wherever it lands. On this lead-free board you're dragging , which melts hotter and dries to a more matte finish than old leaded solder, so set the iron a touch higher.
Inspect, then continuity
Hunt for bridges under magnification, then prove no VBUS–GND short before any power.
Screen, then continuity
Check your work before you ever apply power.
Under magnification, hunt for solder bridges and (a passive stood up on one end). Then run a sweep with your meter: confirm the grounds are connected and — the one that matters most — confirm there is NO continuity between VBUS and GND. A short there would destroy the board the instant USB is plugged in. This is the POST_ASSEMBLY_CONTINUITY gate.
CheckYour meter beeps continuity between VBUS and GND before power-on. Power it anyway?Show
Never — that's a dead short; find and clear it before any power reaches the board.
Checkpoint
Quick check — assembly
Pass the build's POST_ASSEMBLY_CONTINUITY checklist — every item checked or marked N/A.