Tombstoning And Reflow Thermal Imbalance

Tombstoning is the defect in which a small chip component, usually a resistor or a capacitor, stands up on one end during reflow instead of lying flat. One termination is still attached to its pad and the other has been lifted clear, held by the surface tension of the molten solder as it solidified.

This article explains the forces that produce the defect, how the pad geometry and the placement feed them, and how the process is corrected.

The Forces In A Reflowing Joint

Molten solder wets a pad and pulls the component towards the pad centre, because surface tension always acts to reduce the area of the free surface. That pull is balanced by the same force at the other termination. If both ends melt at the same time and the solder wets both equally, the two pulls cancel and the part stays flat.

The defect appears when the balance is broken. If one termination wets before the other, the force at that end acts alone for a short interval, and the part pivots about the molten joint and stands up. The same happens if the volume of paste at one end is much greater than at the other, or if the wetting force at one end is larger because the pad is wider or the surface is cleaner. The window in which the imbalance can act is the time between the first end melting and the second, so anything that widens that window makes the defect more likely.

Thermal Imbalance In The Board

The two pads of a chip component are supposed to be thermally identical, and they rarely are. One pad may be connected directly to a copper plane and the other to a thin trace, so the plane acts as a heat sink and the pad takes longer to reach the liquidus. One pad may be at the edge of a large copper area and the other at the end of a routed signal. The difference in thermal mass is what produces the difference in melting time.

The same effect appears across the assembly. A component next to a connector or a large package sees a different thermal environment from one in open board area, and the position of the part on the panel relative to the conveyor affects the heat it receives. Where tombstoning appears repeatedly in the same location, the cause is almost always a thermal asymmetry in the copper rather than a problem with the placement machine or the paste. The related class of defects that share this mechanism is described under SMT component shift causes.

Pad Geometry And Paste Volume

The pad geometry sets how much force each end can generate. A pad that is much wider than the termination gives the solder a large area to pull against, and one that is narrower gives less. Pad spacing matters as well: pads that are too far apart leave the component sitting on a bridge of solder with a gap under it, and pads that are too close allow the solder to wick up the side of the termination and lift it.

Paste volume contributes in the same way. A deposit that is larger on one pad produces a stronger pull at that end during melting, and a deposit that is placed off centre pushes the component before the solder melts. The relationship between the paste deposit and the pad it sits on is the subject of PCB pad design standards, and it is the first place to look when a chip component will not stay flat.

Placement, Adhesive And Atmosphere

Placement offset usually matters less than thermal balance, but it can still tip the balance. A part placed off centre sits nearer one pad, so the solder at that end begins to act earlier and with a longer lever. A part placed with one termination partly off its pad has a reduced contact area at that end and wets more slowly. Both effects are small on their own and both can be decisive when the thermal balance is already marginal.

Where the mechanism is the interval between the two ends melting, holding the component in place with adhesive during reflow removes the opportunity for it to pivot, and some processes use a small dot of glue under the body for exactly this reason. A nitrogen atmosphere changes the wetting behaviour of the solder and can either help or hurt, depending on the paste and the surface finish, so it is treated as a process variable rather than as a cure. The sequence in which parts are placed and the way the deposit and the pad line up are covered under placement order and pad positioning.

Diagnosing The Cause

The diagnosis starts with the location of the failures. If tombstoning appears at random positions, the cause is likely to be in the paste or the profile. If it appears at the same component on every board, the cause is in the copper connected to that component, and the fix is in the layout rather than in the process. If it appears at every position of the same component in a panel, the cause may be in the stencil aperture for that footprint.

The measurement that supports the diagnosis is the profile at the two pads of an affected component, taken with two thermocouples on the same board. The difference between the two curves shows how long the imbalance lasts, and it converts an argument about causes into a number. Where the difference is large, the layout is the answer; where it is small and the defect still appears, the paste volume and the placement should be examined.

Correcting The Layout

The layout correction is to make the two pads thermally similar. Where one pad is connected to a plane, that connection is made through a thermal relief with narrower spokes, or the plane is relieved around the pad, so that the heat has to pass through a defined path. Where the trace to one pad is much wider than the trace to the other, the wider trace may be narrowed near the pad. The aim is not to make both pads slow, but to make them reach the liquidus at the same time.

Where the layout cannot be changed, the process can compensate. A paste with a lower melting range is not usually an option, but a longer soak brings the whole assembly closer to a uniform temperature before the spike, and a slower ramp to the peak reduces the difference between the two ends. Reducing the paste volume on the pad with the larger deposit narrows the force imbalance as well. These changes are second best to a layout fix, and they should be verified with the same two thermocouples that identified the problem.

FAQ

Does tombstoning affect large components? It is essentially a defect of small chip parts, because a small mass can be pivoted by the surface tension of a joint. A large component has enough inertia and enough paste under it that the same imbalance does not lift it.

Is a nitrogen atmosphere a cure? It changes wetting behaviour and can reduce the defect in some processes, but it does not remove a thermal imbalance between the two pads. Where the underlying difference in copper remains, the defect returns when the atmosphere changes.

Why does a part stand up only on some boards in a run? Small variations in paste volume, placement and board temperature combine, and the defect appears when the combination passes a threshold. That is why it clusters at the positions where the thermal balance is already marginal.