Reflow Conveyor Vibration: Causes, Symptoms and Fixes

Conveyor vibration is one of the few reflow faults that leaves no trace in the profile data. The thermocouples read correctly, the zones hold their setpoints, and the oven passes its daily check, yet boards come out with shifted chip components, tombstoned passives and occasional solder balls next to fine-pitch pads. The cause is mechanical rather than thermal: the board is moving relative to the paste while the alloy is still liquid. Because the symptom looks like a paste or a placement problem, the conveyor is often the last place anyone looks.

Where the Vibration Comes From

The oven itself is the first source. Blower bearings, drive chains and gearboxes wear, and a worn bearing transmits a periodic impulse into the frame that the rails then carry to the board. Fans that are out of balance, or that have picked up flux condensate on the blades, produce a steady hum that is easy to ignore and hard to see. Measuring the frame with a simple accelerometer at several points along the tunnel will show whether the amplitude grows towards one end, which points at the blower nearest that zone.

Loose hardware is the second source. Rail clamps, tunnel section joints and the brackets that carry the return rails all work loose over months of thermal cycling. A single loose bolt on a rail support can turn a rigid structure into a spring, and the board then oscillates at the natural frequency of that spring as it passes, which is how a mechanical fault is mistaken for a solder joint defect. Tightening hardware to the specified torque as part of the preventive maintenance schedule costs very little and removes an entire class of defect.

How Vibration Shows Up in the Product

A component that shifts during reflow is pulled by the surface tension of the molten alloy, and it moves in the direction that minimises the energy of the joint. In practice that means a chip capacitor slides along its pads, rotates, or lifts at one end. The severity depends on the mass of the part and on how long the paste stays liquid. A 0402 chip needs very little energy to move; a large electrolytic or a connector needs much more, which is why a vibration problem often shows up only on the smallest parts.

Tombstoning is the extreme case. One termination wets first, the other is held above the paste by the pull of the first joint and never makes contact. This is usually attributed to an unbalanced thermal profile or to uneven pad heating, and those are real causes, but a board that oscillates while the second termination is coming up to temperature will tombstone even when the profile is textbook. Comparing defect maps against the position of the board in the tunnel helps separate a thermal cause from a mechanical one.

Rail Width and Edge Support

How the board is carried decides how much of the vibration reaches it. A 1.6 mm board supported only at two edges with a long unsupported span behaves like a diaphragm, and any impulse from the rails sets it moving in the middle. Narrowing the rail spacing, or adding a centre support rail for panels above about 250 mm, raises the natural frequency and reduces the amplitude at the centre. For thin boards the effect is pronounced, and a 0.8 mm panel may need support even at modest widths.

Rail width matters as well. If the rails are set wider than the panel, the board sits on the inner edges of the rail and rocks. Setting the rails so that the panel is captured over its full edge, and adjusting the rail height so both sides contact evenly, removes the rocking. The setting should be checked after every product change, because a rail position that suits one panel size will not suit another, and the difference is invisible from the operator side of the machine.

Board Support Fixtures and Their Limits

Where rails alone are not enough, a support fixture or a pallet with pins underneath the board is the practical remedy. Pins should be placed under heavy parts and under the centre of large panels, and they should be positioned so that they do not sit under a pad on the opposite side. A pin that touches the underside of a reflowing joint creates a cold spot and a defect of its own, so the fixture layout has to be reviewed against both sides of the board.

Fixtures bring their own risk. A pallet that is not flat, or that has warped after repeated passes, will hold the board in a curve and change the thermal contact. The pallet should be checked for flatness on the same schedule as the preventive maintenance plan and replaced when it moves outside tolerance. Magnetic or mechanical clamps that hold the board down are useful, but they must not apply enough force to bow the panel, because a bowed panel changes both the profile and the position of the paste.

Speed, Acceleration and Profile Interaction

Conveyor speed sets how long the board spends in each zone, and it also sets the rate of change of direction at the entry and exit. A speed that is too high for the load can cause the chain or belt to surge as the board transfers between sections, producing a jolt exactly at the point where the paste begins to melt. Slowing the transfer, or fitting a transition section that supports the board across the joint, removes the jolt without disturbing the reflow profile.

The profile and the vibration interact through the liquidus time. The longer the alloy stays molten, the longer any disturbance can move a part. Where a defect is marginal, shortening the time above liquidus by a few seconds, while keeping the peak and the soak within specification, can be enough to bring it under control. Any such change should be verified against a full profile measurement rather than assumed, since the soak and the peak also depend on the thermal mass of the assembly.

Measurement and Maintenance

Vibration is measured, not judged by feel. A handheld accelerometer placed on the rail at the entry, the middle and the exit will show the amplitude at each point, and repeating the reading after any maintenance gives a baseline for comparison. Readings taken with a board in the tunnel are more representative than readings taken empty, because the board adds mass and changes the response. The measurement is quick, and it turns a subjective argument into a number.

On the maintenance side, the items worth checking on a fixed interval are the blower bearings, the drive chain tension, the lubrication of the return rails and the tightness of the tunnel section joints. Chain tension that is too loose produces a periodic surge; too tight it loads the bearings and shortens their life. Both extremes show up as vibration, so the tension should be set to the manufacturer’s figure rather than to a general impression of tightness. Recording the readings alongside the profile data keeps the profile record complete and makes a gradual change visible before it becomes a defect.

FAQ

Can vibration cause solder voids? It can contribute, but voids in a reflowed joint are more often the result of flux volatiles or of paste that was printed with an inconsistent volume. Vibration matters most for component shift and for solder balls. Where voids and shifted parts appear together, treat them as two problems and check the paste deposit and the conveyor separately.

Should the oven be run with a support fixture for every product? No. Fixtures add handling, add cleaning and add a source of flatness error. Use them where the panel is large, thin, or carries heavy parts, and where the defect history shows position or tombstone problems that the rails alone cannot control.

How often should vibration be measured? A baseline reading at installation and after every major service, plus a reading whenever a position defect appears that the paste and placement data cannot explain. The measurement takes minutes and gives the maintenance team a target instead of a guess.