Skip to main content

3.2 Profiling Methods

Thermocouple attachment, soak vs. ramp-to-peak, and when to tune TAL (time above liquidus).

Profiling is how you prove a reflow oven’s heat strategy on the actual board instead of trusting settings alone. By attaching thermocouples directly to key pads, you see how the board warms, how long solder stays molten, and whether different areas reach peak together. This is critical because components and copper areas have very different thermal masses—tiny chips heat up almost instantly, while large BGAs, ground planes, and heat-sinking packages lag far behind. A profile lets you choose between a soak recipe to equalize those differences or a ramp-to-peak recipe for faster runs, and then tune belt speed, zone temperatures, and airflow so every pass produces the same, predictable results.

3.2.1 What a “profile” actually is

A profile is just a temperature vs. time story told by a few well-placed thermocouples (TCs) as your board rides the oven. You’re proving four things at once:

  1. Ramp isn’t too aggressive,
  2. Soak evens temperatures and activates flux,
  3. TAL is long enough to fully wet—but not so long you cook parts,
  4. ΔT (spread) at peak is tight across the board.

Keep that picture in your head; everything below is how to measure and shape it.



3.2.2 Thermocouples that tell the truth (attach like you mean it)

Pick the TC spots first, then the adhesive. You want the copper/pad temperatures, not free-air.

Where to place (6 TCs is a sweet spot)

  • Cold spot: usually under a dense BGA or big ground plane.
  • Hot spot: a chip jungle near open copper.
  • QFN/LFPAK thermal pad: voiding/tilt lives here.
  • Edge vs. center: to see ΔT across the panel.
  • Plastic/connector risk: make sure you’re not exceeding its limit.
  • Second-side risk (on 2-pass builds): under a tall part already soldered.

How to attach (ranked by accuracy)

  • Micro-solder dot to pad (best): tin the pad, set the TC bead in a tiny solder dot. Measures pad metal directly. Keep the dot tiny so you don’t add mass.
  • High-temp epoxy: TC bead pressed into epoxy on the pad. Very good if solder isn’t allowed; cure per datasheet.
  • Kapton assist: tape can hold the wire, but don’t rely on tape alone—the bead will read air.
  • Don’t: blob tape on mask, clip to component plastic, or let the bead float—those read low and late.

Two pro tips

  • Strain-relief the wire so it doesn’t peel mid-run.
  • Trim the solder mask window slightly if you must; better contact beats pretty tape.



3.2.3 Two profile styles: Soak vs Ramp-to-Peak

Both can be “right.” Choose the one that fits your board, paste, and defect risks.

Style

What it does

When it shines

Watch-outs

Soak

Holds in a mid-range (≈150–180 °C) to equalize temps and outgas volatiles, then goes to peak.

Mixed-mass boards, big copper imbalance, QFN voiding risk.

Too long → flux exhaustion, solder balls; too hot → early oxidation.

Ramp-to-Peak

Smooth, near-linear climb to peak; short dwell above liquidus.

Even-mass boards, fine-pitch where tombstoning/bridges hate long soaks; low-temp alloys.

If ΔT is large, weak corners may under-reflow.

How to pick fast

  • If ΔT at peak is already <10–12 °C, a ramp-to-peak is usually cleaner and faster.
  • If voiding or cold corners haunt you, try a gentle soak to equalize and vent before peak.



3.2.4 Time Above Liquidus (TAL): when to lengthen, when to shorten

TAL = time above the alloy’s melt point (≈217 °C for SAC, ≈183 °C for SnPb, lower for Bi-based). It’s your wetting “budget.”

  • Lengthen TAL when you see head-in-pillow (HIP) on BGAs or uneven wetting on massive copper. Keep peak modest; extend time.
  • Shorten TAL when cosmetics degrade, components are heat-sensitive, or flux is aggressive (water-soluble): hit peak cleanly, exit sooner.
  • Second-side reflow: keep TAL just enough to reflow joints without over-cooking first-side plastics—often a slightly lower peak and shorter TAL works if printing is solid.

Rule-of-thumb bands (validate for your paste):

  • SAC: TAL ~40–80 s, peak 235–250 °C.
  • SnPb: TAL ~30–60 s, peak 205–220 °C.
  • Low-temp Bi: TAL tighter, ~30–60 s, peak per datasheet (often 165–185 °C).



3.2.5 A quick, repeatable profiling routine

  1. Seed recipe: start from your paste vendor’s template or your last similar board.
  2. Wire 4–6 TCs to real pads (8.2.2), label them.
  3. First pass: run the board, save the raw plot.
  4. Set TAL with belt speed: too short → slow belt; too long → speed up.
  5. Set peak with late-zone temps: adjust the final 1–2 reflow zones.
  6. Tighten ΔT with soak/blowers: nudge mid-zones and blower speeds to pull corners/center together.
  7. Second-side (if needed): clone recipe, drop peak a bit, re-prove TAL.

Only change one knob at a time, re-run, and annotate the recipe (“+10 mm/min belt; TAL 48→62 s”).



3.2.6 Fast DOE when you’re chasing defects

If you’re stuck between “almost good” profiles, run a pocket DOE over a handful of boards:

  • Factor A: Belt speed (−10% / nominal / +10%)
  • Factor B: Peak zone setpoint (−5 °C / nominal / +5 °C)
  • Optional C: Blower (−1 / 0 / +1 step) or Atmosphere (air vs N₂)
    Measure BGA HIP, QFN voiding, tombstones, and AOI/AOI-related calls. Keep the combo that fixes the Pareto without cooking everything else.

3.2.7 Special cases (where profiles get touchy)

  • Massive thermal pads (QFN/LFPAK): prefer soak + modest peak; windowed apertures help more than heat (see 7.4).
  • Fine-pitch BGA/CSP: HIP hates short, spiky TAL—smooth ramp and steady TAL. N₂ can buy you margin.
  • Tall connectors/plastics: guard maximum part temperature; consider heat shields or a cooler second-side profile.
  • Low-temp alloys: narrower window—avoid long soaks; gentle ramp-to-peak is your friend.
  • Warp-prone panels: fix mechanics (supports/rails) before you chase with heat (9.1.8).


3.2.8 What “good” looks like on the plot

  • Ramp ≤ ~2 °C/s (no saw-teeth).
  • Soak (if used) flat and calm, 60–120 s in range; ΔT shrinking before reflow.
  • TAL inside target, smooth through liquidus (no dip).
  • Peak inside paste + part limits, ΔT at peak ≤10–12 °C typical.
  • Second-side: same shape, slightly gentler.



3.2.9 Lock it and make it findable

  • Save as: Alloy_Side_PanelMass_Air/N2_vX (e.g., SAC_Top_4upHeavy_Air_v3).
  • Attach the golden plot and TC map to the recipe.
  • Note belt speed, zone temps, blower %, and atmosphere.
  • Store with the Golden Recipe (8.5) so night shift loads the same truth.



3.2.10 Release checklist (tape this to the oven PC)

  • TCs soldered/epoxied to pads; spots labeled (cold/hot/QFN/edge/connector/2nd-side).
  • Ramp, soak (if used), TAL, peak inside paste + part limits.
  • ΔT at peak within target (≤10–12 °C typical).
  • Second-side recipe proven (if applicable).
  • Golden plot + recipe saved with clear name/comments.



Bottom line: attach TCs to pads, not air; pick soak when you need equalization/void control and ramp-to-peak when the board is already even; and tune TAL—with belt speed first—to cure HIP without cooking plastics. Do that, and profiles stop being a guessing game; they become a repeatable tool you can trust.

Back to top