Tombstoning And The Forces That Lift A Chip
A two terminal chip component can lift onto one end during reflow and stand on its termination like a headstone. The defect is called tombstoning, it happens within a second or two at the moment the solder becomes liquid, and it is one of the few assembly faults that the layout and the process can each cause on their own. A board that tombstoned once will usually do it again, because the cause is a repeatable imbalance rather than a random event.
This article explains the mechanism, the forces that decide which way the part falls, how the layout and the paste contribute, and how the problem is removed from a running line.
What Happens At The Moment Of Melting
The part sits on two deposits of paste. As the oven brings the assembly to the melting point, both deposits become liquid at very nearly the same time, and each of them pulls on the termination above it because molten solder wets the metal and the surface tension draws the part down onto the pad. If the two pulls are equal and simultaneous the part settles flat. If one end becomes liquid first, or pulls harder, the component rotates about that end and the other termination lifts clear of its paste.
Once one end is off the pad there is nothing to pull it back. The surface tension at the wetted end holds the part upright, and the component stays there through cooling. The joint that does form is often a sound fillet, which is why the defect is found by inspection rather than by electrical test — the part is standing, but it is still connected.
The Balance Of Forces
Four things decide the balance. The first is the timing of the melt, which is set by the thermal mass of each pad and the amount of copper attached to it. The second is the volume of paste on each side, which is set by the aperture. The third is the position of the part relative to the pads, which is set by the placement machine. The fourth is the wetting force at each termination, which is set by the finish on the part and on the pad and by the flux.
Any asymmetry in those four tips the part. A pad connected to a plane heats more slowly than one connected to a thin trace, so the two ends melt at different times, and that difference alone is enough. The same applies to a pad with more copper on one side, to a via in one pad and not the other, and to a component placed slightly off centre by a machine whose vision was not taught correctly.
Pad Geometry, Mask And Copper Balance
The most common layout cause is an asymmetric connection. A designer who connects one end of a resistor to a ground plane for a good reason and leaves the other end on a normal trace has created a thermal difference that no process change will remove. Where the connection is necessary, the usual remedy is to add copper to the other pad so that the two thermal masses are closer, or to route the plane connection through a narrow neck rather than a wide one.
Mask and pad dimensions matter in the same way. A mask defined pad with a different opening on each side, an annular ring that is larger on one pad, or a pad that is wider than the other changes the way the solder pulls. The dimensions themselves are covered under PCB pad design standards, and the relationship between the pad and the part during placement is described under placement order and pad positioning.
Paste, Placement And Profile
A deposit that is short on one side lifts the part at that end before the solder melts, because the component is supported by the paste on the other side and tipped by its own weight. The same happens when one aperture is partly clogged, when the paste on one pad has dried more than the other, or when the stencil aperture is reduced on one side only to satisfy an area ratio calculation.
Placement contributes offset and force. A part placed off centre has more paste under one termination, and a placement force that is too high collapses the paste under the first end to touch and leaves the second end standing high. The profile then decides how much time the part has to settle: a slow ramp lets both ends reach the liquid state together, while a fast ramp magnifies any difference in thermal mass between the pads.
Removing It From The Line
The first check is whether the defect appears on one component, on one position or across the board. A single position points to the layout of that position and to the copper attached to its pads. A single component type points to the part terminations or to the paste volume. A whole board points to the profile or to the paste.
The process remedies are then applied in order of cost. The placement offset is corrected, the aperture is checked against the drawing, the profile is lengthened so that both ends melt together, and the soak is adjusted so that the smaller pads are not overheated while the large ones catch up. Nitrogen is sometimes used to improve wetting and reduce the oxide on the terminations, which reduces the size of the imbalance but does not create one where the layout has none. Related shifting mechanisms are described under SMT component shift causes.
Inspection And Verification
Inspection is visual or automated optical, and the defect is easy to find because a standing part is obvious from above. The difficulty is not detection but classification: a part that is standing at 45 degrees on one pad and touching the other is a different case from one that is completely vertical, and the second is a genuine open circuit.
Verification of a fix is a run of a few panels with the corrected setting and a full inspection of the affected position, followed by a check that the change has not introduced a different defect elsewhere. The measurement that supports this is the paste volume on the two pads and the placement offset, both of which are available from the printer and the placement machine without any additional equipment.
Small Parts And Fine Pitch
The defect is most common on the smallest bodies, because a 0402 or 0201 chip has very little mass to resist the pull of a molten fillet and very little paste under each termination to hold it. The same part also has the tightest placement tolerance, so an offset that would be invisible on an 0805 is enough to unbalance an 0201. A line that runs fine pitch passives on the same panel as large ones should expect the problem on the small parts and look for its cause in the placement accuracy rather than in the profile.
The pad spacing sets an additional limit. Where two pads are close together, the paste on one side can wick towards the other before the solder melts, and the resulting shift is a separate mechanism from the imbalance described above. Both produce a standing part, and separating them requires looking at the deposits after printing rather than at the finished board.
FAQ
Does tombstoning mean the component is faulty? Usually not. The terminations and the body are normally within specification, and the defect is a consequence of the assembly conditions unless the wetting of one termination is visibly different.
Does more paste help? More paste on both pads gives the part more to settle into, but paste on one pad only makes the imbalance worse. The volume has to be correct and symmetric rather than generous.
Can a tombstoned part be reworked? It can, with hot air and a fresh deposit of paste, but the operation is slow and risks damaging the adjacent parts, which is why the defect is treated as a process problem rather than a rework problem.
