Lamination Cycle Parameters for Multilayer Boards

Lamination is the step in which a stack of prepreg and copper becomes a single board, and the press cycle is the recipe that makes it happen. The resin has to flow far enough to fill the gaps around the inner layer patterns, cure far enough to reach the specified glass transition temperature, and stop moving before the layers shift out of registration. Every one of those requirements is a function of temperature, pressure, vacuum and time.

What the Press Has to Achieve

The press has four jobs. It has to heat the stack at a controlled rate, apply pressure so that the resin flows, hold a vacuum so that air and volatiles can escape, and cool the finished panel under pressure so that it comes out flat. A cycle that satisfies three of them and misses the fourth produces a panel that looks acceptable and fails later.

The order and the timing of those steps matter as much as their values. Pressure applied too early squeezes resin out of the edges before it has softened, while pressure applied too late leaves voids that no amount of curing will close.

Prepreg and Resin Flow

Prepreg is glass fabric impregnated with resin that has been partly cured, and the remaining flow is what the press is designed to use. Resin content, gel time and volatile content are the properties that decide how the material behaves, and all three change with storage conditions and with the age of the roll.

Enough resin has to flow to fill the copper pattern on the inner layers, which may be several thousandths of an inch thick. Too little flow leaves voids and resin starvation beside a heavy copper feature, while too much flow leaves the layer thin and the dielectric spacing below the value the design assumed.

The Stages of a Press Cycle

A typical cycle has a ramp, a dwell at a lower temperature while the resin softens and flows, a second ramp to the cure temperature, a cure dwell under full pressure, and a controlled cool down. Vacuum is applied early and released at a point chosen so that the resin has gelled before the pressure comes off.

The dwell stage is where the difference between a good and a marginal panel is decided, because it sets how much resin moves. A short dwell suits a thin, lightly loaded stack, while a thick panel with heavy copper planes needs a longer one to fill the pattern completely.

Temperature, Pressure and Vacuum

Temperature is measured on the panel rather than in the press platens, because a stack of prepreg and copper heats more slowly than the steel around it. The thermocouple is normally placed in a test stack built to represent the production build, and the cycle is written from its reading.

Pressure is applied in stages in most recipes: a low pressure during flow, a higher pressure during cure, and full pressure during cooling. Vacuum removes the air between the layers and the volatiles released by the resin, and its loss during a cycle is one of the few faults that can be seen immediately in the finished panel.

Registration and Movement During Cure

The inner layers are held in position by tooling pins or by a riveted stack, but the resin still moves as it flows and the layers move with it. The amount of movement depends on the flow, on the fit of the pins and on the symmetry of the copper distribution across the panel.

Asymmetric copper is the common cause of a shift, because the resin flows further where there is less copper to fill. The registration allowance for a given build should therefore be set from the measured movement of that build rather than from a general rule; the layer to layer relationships that have to be preserved are set out in the guide to layer assignment rules.

Thickness, Resin Content and Dielectric Spacing

The finished thickness of a dielectric layer is set by the prepreg used and by the amount of resin that flows out of it. This is why the impedance of a controlled impedance layer is a lamination variable as much as a design variable, and why a change of prepreg grade requires a new stack calculation.

Resin content is measured on the finished panel by a burn off test, and the result should sit inside the window agreed with the laminate supplier. A panel that is resin rich will have a lower dielectric constant than the design assumed, and one that is resin starved will be brittle at the edges and hard to drill.

Cooling, De-stressing and Panel Flatness

Cooling under pressure is what keeps a panel flat. If the pressure is released while the panel is still hot the layers are free to move, and the panel bows or twists as it cools, distortion that usually survives the rest of the process.

The rate of the drop matters as well. A fast descent through the glass transition temperature locks in stress that appears later as warpage after the board has been through reflow, and the copper distribution on the two outer layers can pull the panel in one direction if it is not balanced. Where that happens, the usual remedy is a change to the artwork rather than to the press, as described in the guide to copper balance and thieving.

Monitoring and Recording the Cycle

Every press cycle should be recorded with the platen temperature, the panel temperature from the test stack, the pressure at each stage, the vacuum level and the time at each transition. The record is what allows a batch to be reviewed when a downstream problem appears weeks later.

The press itself needs attention too. A worn seal, a blocked vacuum line or a platen with an uneven temperature profile will change the cycle without changing the recipe, and a periodic survey of the platen surface with a thermal map is the usual way to catch that drift.

Faults Traced to the Lamination Cycle

Delamination, voids, resin starvation, thickness variation, misregistration and warpage all have a lamination origin, and several of them look identical to defects caused elsewhere. Voids can come from the press, from the prepreg or from contamination on the inner layers, and the location of the voids is often the clue that separates those causes.

Where a fault is traced to the cycle, the response should be a controlled change and a new record rather than an adjustment made by feel. The material properties that set the boundaries of the cycle are covered in the guide to laminate material properties.

FAQ

How long does a lamination cycle take? It depends on the thickness of the stack and on the resin system, and a heavy panel may take several hours from load to unload. The duration is set by the panel thermocouple rather than by the clock.

Can a cycle be shortened to increase output? It can within limits, by raising the ramp rate or trimming the flow dwell, but both changes affect resin flow and registration. Any shortening should be qualified on a test stack before it is used in production.

Why do panels bow after lamination? Usually because the pressure was released while the stack was still hot, or because the copper distribution on the outer layers is unbalanced. The first is a cycle fault and the second is a design issue, and the two are told apart by whether the bow is consistent across panels.