Surface-Mount Soldering

Chip sizes, tack-and-reflow, drag soldering fine-pitch, hot air and paste, tombstones and bridges.

Package sizes and what can be hand-soldered

Surface-mount (SMD) parts are soldered to pads on the surface of the board with no holes. They are smaller, lighter and cheaper to assemble in volume; by hand they need a steady hand, magnification and the right tools. Chip resistors and capacitors are named by their size in hundredths of an inch (0603 is 0.06 × 0.03 in) or their metric size in tenths of a millimeter (1608 is 1.6 × 0.8 mm).

Size (inch)MetricDimensionsTypical resistor powerMass
040210051.00 × 0.50 mm62.5 mW0.8 mg
060316081.60 × 0.80 mm100.0 mW2.0 mg
080520122.00 × 1.25 mm125.0 mW5.0 mg
120632163.20 × 1.60 mm250.0 mW10.0 mg
PackageHand-solderable?Method
1206, 0805, 0603YesTack and solder with a fine tip
0402With care, magnificationTweezers, fine tip, low temperature
0201, 01005Practically noPaste, stencil and reflow
SOIC, SOT, TSSOP (pitch 0.65 to 1.27 mm)YesDrag solder or tack and flux
QFP, LQFP (0.4 to 0.8 mm)Yes, with practiceDrag solder with flux, or hot air
QFN, DFN (leadless)DifficultHot air with paste; pad under the chip
BGA, WLCSPNoReflow oven or hot-plate with paste; X-ray to inspect

The power ratings are typical for thick-film resistors; follow the data sheet. The pads themselves are designed per IPC-7351 (see PCB Design and Layout).

Chip parts: tack and solder

The simplest hand method for a two-terminal chip part:

  1. Tin one pad: melt a little solder onto one pad with a fine tip.
  2. Place the part: hold it with tweezers on its pads, reheat the tinned pad and slide the part into place, then remove the tip; the part is now tacked on one end.
  3. Check the alignment, under magnification, and reheat the tack to adjust if necessary.
  4. Solder the other end: touch the tip to the pad and the part terminal, feed a small amount of fine wire (0.5 mm) and remove.
  5. Reflow the first end briefly, to relieve any tension and make a proper fillet.

Worked example: a joint on an 0805 chip

  • Pad 1.0 × 1.3 mm. The joint needs about 0.08 mm³ of solder (see Solder Alloys, Flux and Paste): with 0.5 mm wire (area 0.20 mm²) that is a length of 0.4 mm: a tiny touch, about the thickness of the wire.
  • Heat energy is tiny: the 0805 end and its pad are about 0.05 g of copper, 4.3 J to reach 250 °C. At 30 W that is under 0.2 s. The problem is not the heat but the control, and that is why a small tip and short contact give the best result.

Fine-pitch ICs: drag soldering and paste

Fine-pitch leaded packages (SOIC, TSSOP, QFP) can be soldered by dragging a loaded tip along the pins with plenty of flux. The flux makes the solder pull onto the pad and lead rather than bridge, and the bridges that form are removed by the drag itself. The gap between pads is the limit:

PitchTypical pad widthGap between padsGap as share of pitch
1.27 mm0.60 mm0.67 mm52.8 %
0.80 mm0.45 mm0.35 mm43.8 %
0.65 mm0.40 mm0.25 mm38.5 %
0.50 mm0.30 mm0.20 mm40.0 %
0.40 mm0.22 mm0.18 mm45.0 %
0.30 mm0.17 mm0.13 mm43.3 %
  • Align and tack: tack two diagonal corner pins with a little solder and flux, check the alignment of all pins under magnification, then re-tack if needed.
  • Flux first: run flux gel or a pen along the pins on all sides.
  • Drag: load a hoof or chisel tip with a small bead of solder and pull it slowly along the pins, in one pass, at about 2 to 3 mm per second.
  • Remove bridges by adding flux and dragging again, or with wick (see Rework, Desoldering and Repair).
  • Inspect all pins under magnification and check for bridges, dry joints and bent pins (see Joint Quality, Inspection and Failure).

Worked example: drag soldering a 100-pin QFP

  • 100 pins at 0.5 mm pitch is 25 pins per side. Pin row length = 24 × 0.5 + 0.3 = 12.3 mm.
  • At 2.5 mm/s a drag takes 4.9 s. With fluxing and checking, about 8 s per side, the whole job is 52 s ≈ 0.9 min plus cleaning (illustrative).
  • The gap between pads is 0.20 mm (40 % of the pitch): easy to bridge with a loaded tip, which is why plenty of flux and a clean tip matter more than speed.

QFN, thermal pads and BGAs

Leadless packages (QFN, DFN, LGA) and BGAs hide their joints under the body. They need solder paste and heat from above (hot air) or below (a hot plate or oven), and the board is inspected by X-ray for BGAs. QFN pins on the sides can be inspected with a magnifier, but the large central thermal pad cannot, and it carries most of the heat of the chip.

Worked example: paste on a QFN thermal pad

  • A 3.2 × 3.2 mm thermal pad has an area of 10.2 mm². Stencil designers use a “window pane” aperture covering about 50 %, to limit voids and floating (the chip can ride on a thick layer of molten solder).
  • Paste volume = 10.2 × 0.50 × 0.12 = 0.61 mm³, metal 0.31 mm³ (51 % by volume), which spreads over the pad to a layer 31 µm thick.
  • A thin layer like this is deliberate: it keeps the chip from floating. The usual aim is a void area under 25 to 50 % of the pad (the specification decides), and via-in-pad designs fill or cap the vias so that the solder does not drain away.

Paste, stencils and hot air

For more than a few joints, solder paste is faster and gives neater joints (see Solder Alloys, Flux and Paste for paste and stencils). The process:

  1. Apply paste: print through a stencil (taped to the board for a prototype), or dispense by syringe.
  2. Place the parts with tweezers or a vacuum pen; the paste holds them.
  3. Reflow the board with hot air, a hot plate or an oven (see Reflow Profiles, Ovens and Hot Plates).
  4. Inspect and rework any joints that need it.

With hot air, keep the airflow low so that parts do not blow away, use a nozzle slightly larger than the part, heat in circles and watch for the paste to turn bright and the part to self-align as the surface tension of all the pads pulls it into the center. Small parts can correct a placement error of up to about half the terminal width.

Tombstones

A tombstone (or “Manhattan”) defect is a small chip part standing on one end. It happens when one pad’s solder melts first: its surface tension pulls that end down and the other end up, before the second pad has melted. The cause is an imbalance of heat or paste between the pads.

Worked example: why 0805 parts tombstone

  • Surface tension pulls on the terminal with a force of about F = γ × w = 0.46 × 1.25 mm = 0.58 mN. The 0805 weighs 5.0 mg, i.e. 0.049 mN: the pull is 12 times the weight.
  • The torque of the pull about the far end is F × (part height 0.6 mm) = 0.35 µN·m, against 0.049 µN·m for the weight: 7 times greater.
  • So if the second pad is not yet molten the part tilts. A temperature difference of 5 K between the pads, at a ramp of 3 K/s, gives a delay of 1.7 s, ample for the lifting.
SizeWetting pull per terminalWeightPull ÷ weightTorque ratio
04020.23 mN0.008 mN29×29×
06030.37 mN0.020 mN19×19×
08050.58 mN0.049 mN12×15×
12060.74 mN0.098 mN8×8×

Smaller parts have a larger ratio and tombstone more. The cures follow from the cause (the figures are illustrative):

  • Even heating: preheat the board and keep the ramp slow, so that the temperature difference between the pads is small.
  • Equal paste on both pads and symmetrical pad sizes and copper connections (a pad on a plane heats slower than a pad on a thin track; add thermal reliefs).
  • Avoid oversized pads (more solder, more pull) and use a nitrogen atmosphere or a lower-surface-tension paste in production.
  • Place accurately, because a part off-center has an unequal wetting area on its two pads.

Bridges and how to remove them

A bridge is solder joining two adjacent pins. It is caused by too much solder or paste, a poor mask between the pads, a dirty tip or insufficient flux. To remove it, add flux and touch with a clean tip so that the excess is drawn back onto the pads, or use wick (see Rework, Desoldering and Repair). Bridges between pads separated by solder mask (a “mask dam”) are far less likely, so the board design matters (see PCB Design and Layout).

Troubleshooting

SymptomLikely causeWhat to do
Tombstoned chipUneven heat, paste or pad sizeReheat both ends; use even heat; check pad symmetry
Bridged pinsToo much solder, no fluxAdd flux and drag or wick
Dry or open pinBent lead, too little solder, unwetted padAdd flux, press the lead, reflow
Part skewed or rotatedPoor placement, too much airflowReheat gently and align with tweezers
Solder balls near the chipPaste slump, moisture, fast rampDry the board; check the profile
Chip blown off by hot airFlow too highLower flow, bigger nozzle
QFN does not workThermal pad not wet or voidedReflow with more flux; use X-ray where possible

Related

FAQ

What is the smallest part I can hand-solder?

With a good microscope and steady hands, 0402 and 0.4 mm-pitch QFPs; 0201 and 01005 parts really need paste and reflow.

Drag soldering or paste: which is better?

Drag soldering is quick for a single leaded IC. Paste with a stencil is better for many parts or for any leadless package.

Why do my resistors stand up on one end?

It is the tombstone effect: one pad melts first. Heat both ends evenly, use the same amount of paste and keep the pads symmetrical.

Do I need flux on SMD work?

Almost always: extra flux (gel or pen) makes the solder flow onto the pads and stops bridges, even with flux-core wire.

Can I solder a BGA by hand?

Not with an iron. Use a hot-air station, a pre-heater and paste or a reballed part, and inspect it (X-ray or electrical test).

References

  • IPC-7351, Generic Requirements for Surface Mount Design and Land Pattern Standard; IPC-A-610; IPC J-STD-001.
  • Lee, N.-C., Reflow Soldering Processes and Troubleshooting, Newnes (tombstoning and self-alignment).
  • Lau, J. H., Ball Grid Array Technology, McGraw-Hill.

Last reviewed: October 2026. Package sizes, powers, masses and defect forces are typical or illustrative.