Via-in-Pad DFM: Preventing Solder Drain During Reflow
In PCB design, engineers often adopt via-in-pad to raise routing density or improve electrical performance by placing a via directly in the component pad. The structure saves escape space under fine-pitch packages and can shorten the current return path. It also introduces a manufacturing risk that is easy to underestimate at layout time: solder drain.
During reflow, molten solder can leave the pad, enter an open via, and starve the joint. Design-for-manufacturability (DFM) for via-in-pad is therefore not optional decoration. It is the set of process and layout controls that keep solder volume on the pad where the joint needs it.
Survival Guidelines for Via-in-Pad
For vias located within SMD pads, the recommended practices are ordered by reliability.
- Primary (mandatory) approach: via-in-pad must be resin plugged or filled.
- Fallback (design workaround): if via filling cannot be implemented because of cost or process limits, then add a solder-mask dam that physically isolates the via from the pad, or relocate the via to a corner or edge of the pad, away from the primary wetting area.
The core objective is to prevent molten solder from draining through the via to the opposite side during reflow. That drain leaves the pad with insufficient solder and produces a weak, unreliable, or open joint.
Filling closes the barrel. A mask dam attempts to block the path from pad copper to the hole without changing the drill. Moving the via reduces how much of the wettable pad still touches an open hole. Each step down this list accepts more residual risk. The first option removes the capillary path. The later options only constrain it.
What Happens During Reflow
The DFM rules exist because of the way solder paste behaves once it melts.
An unfilled via is a hollow, copper-plated barrel. When the paste becomes liquid, solder can flow along the via wall into the hole by wicking and capillary action. Flow continues until the via is partly or fully filled, or until solder exits on the backside of the PCB. The top-side pad then loses a large fraction of the solder volume that was printed for that joint.
The joint that remains may look acceptable in a quick optical glance and still fail mechanically or electrically. Fine-pitch BGA pads are especially sensitive because the printed paste volume is already small. A via that captures even part of that volume can drop the ball below a reliable fillet. Chip components see a related failure when one pad loses solder and the part tombstones or is left with a starved termination.
Backside solder beads are a second symptom of the same mechanism. Solder that leaves the top pad does not disappear. It can emerge on the opposite side as unwanted balls that interfere with assembly, test, or later soldering. Filling the via stops both the starved top joint and the backside escape.
Three Treatment Options Compared
Industry practice offers three mainstream ways to handle via-in-pad. The choice depends on pitch, reliability target, via size, and whether the fabricator can fill the hole.
| Option | Treatment | Advantages | Disadvantages | Application scenarios |
|---|---|---|---|---|
| Option 1 (recommended) | Resin fill plus plated-over planarized cap | Completely blocks solder drain. Pad surface is flat. Highest reliability. | More complex process, higher cost, slightly longer lead time. | BGA pads, fine-pitch devices, and products with high reliability requirements. |
| Option 2 (compromise) | Solder-mask dam | Uses solder mask for physical isolation. Lower cost. | Mask dam can fail because of alignment or print variation; vias can still trap solder beads. | Highly cost-sensitive designs and components with wider lead spacing. |
| Option 3 (design optimization) | Move via to pad corner or edge | No change to the manufacturing process; no extra cost. | Gives up some electrical advantages of via-in-pad; still needs measures to prevent solder flow. | Projects that cannot afford via filling but have enough layout space. |
Option 1 is the only method that restores a closed, planar pad. Resin fill occupies the barrel. A plated cap then returns a solderable surface that assembly can treat like an ordinary pad. That combination is why filled via-in-pad is the default for BGA and other fine-pitch lands.
Option 2 keeps an open or tented hole and tries to separate it from the paste deposit with a mask bridge. The cost is lower because the fabricator does not fill and plate over the via. The weakness is process variation. If the dam is thin, misaligned, or broken, the capillary path reopens.
Option 3 is a layout change rather than a process change. Shifting the via to the pad corner or edge reduces overlap with the primary wetting area. Some electrical benefit of a centered via is lost, and solder can still move if the remaining overlap is large. The method is useful when fill is unavailable and the pad is large enough that a partial via does not consume the whole land.
Why Solder-Mask Covering Is Not Filling
A common question is why a via that is already covered with solder mask still needs to be filled. Covering and filling are different barriers.
Liquid photoimageable solder mask (LPISM) forms a thin film over the via opening. That film has finite thickness. Under reflow thermal stress it can crack, or it can simply be too thin to stop molten solder from penetrating the opening.
Capillary pathways can also remain at the edge of the film. Even when the opening looks covered, molten solder can creep along small gaps between the via wall and the solder mask and enter the hole. Masking therefore reduces splash and can tent a small via, but it does not replace a filled and sealed barrel.
Only physically filling and sealing the via provides a robust barrier against solder drain. Fabrication notes that say "covered" without "plugged" or "filled" leave this distinction unresolved and should be treated as incomplete for true via-in-pad lands.
Design Execution Checklist
To avoid yield loss or field failures from via-in-pad, apply the following checks during design and documentation.
- Run a global scan. Use the PCB tool's DRC to identify every via that intersects an SMD pad.
- Review attributes for each via-in-pad instance. Confirm that plugged or filled requirements are specified in the design output and fabrication notes. If they are missing, update the design and call out via plugging or filling on the PCB fabrication drawing.
- If via filling is not feasible, often because the via diameter is too large for the fabricator's fill process, add a solder-mask dam between the pad and the via. A dam width of at least 4–5 mil is recommended. Alternatively, move the via to the pad edge so that it only partially overlaps the land and the wetted contact area with molten solder is reduced.
The checklist has to reach the fabricator. A filled via that exists only in the layout database, and not in the notes, may be built as an open plated hole. Confirm the house can fill the actual drill size before the stack-up and via list are frozen.
Closing Guidance
Via-in-pad is a double-edged structure. It can solve routing density and return-path length, and it can also create systemic solder defects if the barrel remains open to the pad.
If the via is in the pad, fill the via. If fill is not available, add a solder-mask dam. If there is no fill and no dam, starved joints and difficult rework are the expected outcome. Treat via-in-pad with explicit design rules, document the required process on the fabrication notes, and confirm capabilities with the PCB manufacturer so assembly results stay consistent.
