Tombstoning Causes and Prevention in Reflow Soldering

Tombstoning is the defect in which a two terminal chip component stands up on one end after reflow instead of lying flat on its pads. The joint on one side forms and pulls the part upright, while the other termination is left in the air or resting on flux residue. It is one of the most visible defects on a board and one of the most annoying, because a single lifted part can stop a line and because the cause is often a small imbalance that is hard to see before the fact.

What Tombstoning Looks Like

The lifted component stands at an angle that can range from a slight tilt to a fully vertical position. In the mild cases the part is still connected on one end and may even pass an electrical test, which makes the defect worse than an open because it can survive into the field and fail later under vibration or thermal cycling.

The direction of the lift is a clue. Where components on a board lift in the same orientation, the cause is usually in the process, such as a thermal gradient across the board or a print direction. Where the lifts are random, the cause is more likely to be in the deposit, the placement or the component itself.

Why One End Lifts

During reflow the solder on both terminations melts and wets, and the surface tension of the molten solder pulls each end towards the pad. If one end melts, wets and pulls before the other, the torque it generates is no longer balanced, and the component rotates up around the joint that formed first. The part is a lever, and the paste on the two ends has to behave the same way at the same time.

Anything that makes the two ends differ can start that rotation. A difference in paste volume, a difference in pad size, a difference in heating rate, a difference in termination finish or a component that is not centred on its pads all change the balance of forces, and the defect appears as soon as the imbalance is large enough to overcome the weight of the part.

Pad Geometry and Thermal Balance

Pad geometry is the first thing to examine because it is fixed in the design and cannot be corrected on the line. The two pads of a chip component should be identical in size and shape, and both should be connected to copper that heats at a similar rate. A pad that is tied to a plane on one side and to a thin trace on the other will heat at a different speed from its partner.

Thermal relief connections, plane clearance and the direction of the traces all influence how quickly each pad reaches the melting point of the paste. Where a design cannot avoid an asymmetry, the imbalance can sometimes be compensated in the stencil, but the better answer is to keep the two pads symmetrical in the first place.

Solder Paste Volume and Deposit Shape

The volume of paste on each pad sets how much solder is available and how much flux has to be driven off before the alloy melts. A deposit that is larger on one side takes longer to heat and applies a different force, so the two ends of the part do not wet at the same moment. Print defects such as a partially filled aperture are therefore a direct cause of tombstoning.

Deposit shape matters as well as volume. Paste that has slumped or been smeared bridges unevenly to the termination, and the wetting front then advances at a different rate on the two ends. Measuring the deposit before reflow is the only way to separate a printing problem from a reflow problem, and the technique is described in the guide to solder paste inspection.

Reflow Profile and Heating Rate

The profile controls how the two ends of the component heat. A slow, even ramp gives the paste on both pads time to reach the same temperature, while a fast ramp exaggerates any difference in thermal mass between them. The soak zone, where the assembly is held at an intermediate temperature, exists partly to even out that difference before the alloy melts.

The peak temperature also matters. Where the paste is melting at the very edge of its window, small differences in heating rate decide which end melts first, and the defect rate rises sharply. Profiles should be measured on the actual assembly rather than copied from a similar product, because the thermal mass and the copper distribution of the board change the result.

Placement Accuracy and Nozzle Effects

A component that is placed off centre has more paste under one termination than the other, and the imbalance appears as soon as the solder melts. Placement accuracy should be verified with the component data and with a placement force that is enough to embed the termination in the paste without squeezing it out to the side.

Nozzle condition and vacuum release timing affect the same variables. A worn nozzle or a late vacuum release can rotate the part or drop it from a height, and both change the position and the shape of the paste under the terminations. These causes are simple to check and are often overlooked when the investigation starts with the oven.

Component and Termination Factors

The component itself contributes to the problem. Termination finishes vary in how quickly they wet, and a batch with a poor or oxidised finish will lift more often than a batch with a good one. Components at the small end of the size range have less mass to resist the torque, which is why the defect is most common on the smallest chip sizes.

The pad finish on the board matters for the same reason. A finish that wets unevenly, or one that has aged in storage, changes the point at which each end of the part begins to pull. The available finishes and their behaviour are compared in the guide to surface finishes.

Detection and Rework Limits

Detection is visual for a lifted part and automatic for a machine that has been taught the defect. An automated optical system will find a tombstoned component easily because the height difference is large, but it will not find the marginal cases where the part is tilted by only a few degrees and still connected at both ends.

Rework is possible on a small component with hot air or a soldering iron, but it becomes impractical as the size falls and as the surrounding density rises. A board with several lifted parts is usually better reflowed again or scrapped, because each rework operation carries its own risk of damaging the pad or the neighbouring parts.

Design and Process Prevention

Prevention starts with symmetrical pads and a stencil that delivers the same volume to both. It continues with a profile that heats the assembly evenly and a placement programme that centres each part accurately. Where the defect persists, the analysis should compare the two ends of a lifted part rather than the board as a whole, because the answer is almost always a difference between them.

The wider lesson is that small chip components are sensitive to balances that larger parts ignore. The same discipline of controlling tolerance that protects other parts of the assembly, as described in the guide to component tolerance and reliability, is what removes the conditions that tombstoning needs.

FAQ

Why do the smallest chip components tombstone most often? They have the least mass to resist the torque produced by one joint forming before the other, so the same imbalance that a larger part absorbs without moving is enough to stand a small part on its end. The smaller the part, the more precisely the paste volume and the placement have to be controlled.

Can tombstoning be fixed by changing the profile alone? Sometimes, if the root cause is a heating imbalance between the two pads. Where the cause is a difference in paste volume, pad geometry or placement, a profile change will reduce the symptom without removing it, and the defect will return when the material or the product changes.

Is a slightly tilted component acceptable? Only if the joint meets the acceptance criteria on both ends, since a tilted part may have one joint that is well formed and another that is marginal. Where the criteria are in doubt, the safer decision is to treat the part as a defect and to investigate the cause rather than to pass it.