Board Thickness Tolerance and How It Affects Assembly

Board thickness looks like a simple dimension, but it sits at the intersection of the stack-up, the impedance target, the connector and the assembly line. A panel that finishes a tenth of a millimetre outside its nominal value can still pass every electrical test and then fail to seat in a card guide or feed through a wave solder machine. Controlling it starts with understanding where the tolerance comes from.

Why Board Thickness Is Specified

Thickness is specified because other things depend on it. Connectors, card guides, edge connectors, wave solder pallets and depanelisation fixtures are all designed around a nominal board thickness with a tolerance band. Impedance controlled traces are designed around the dielectric spacing that thickness represents, and a change in thickness changes that spacing directly.

It also affects manufacturability. A thin board is more flexible and harder to handle through automated lines, while a thick board needs longer drilling, more router passes and different lamination pressure. The specified thickness is therefore a negotiation between electrical requirements, mechanical requirements and what the process can hold repeatably.

Where the Tolerance Comes From

Finished thickness is the sum of the copper, the prepreg, the core and the plating, each with its own variation, and then the whole stack is pressed to a target that has its own distribution. The dominant contributors are the prepreg resin content and the press pressure, because both change how far the stack compresses during lamination.

Plating adds another variable. Copper plated onto the surface and into the holes increases the finished thickness, and the amount depends on the plating time, the current density and the pattern density on the panel. Board thickness tolerance and copper thickness tolerance are therefore linked, and specifying both tightly without recognising that link leads to unnecessary cost.

Standard Thickness Versus Finished Thickness

Laminate is supplied in standard sheet thicknesses, and designers often specify a finished thickness that does not correspond to any combination of standard core and prepreg. The fabricator then has to build to the finished dimension using whatever combination comes closest, which may need non standard cores or a build up from two thinner sheets.

Specifying a finished thickness instead of a laminate thickness removes the ambiguity, but it also transfers responsibility for achieving it to the fabricator. The better practice is to agree the construction with the fabricator during design review, so that the stack-up and the achievable tolerance are known before the artwork is released rather than after the first panels are measured.

Thickness and Impedance Control

Controlled impedance depends on the trace geometry, the dielectric constant and the distance to the reference plane. That distance is set inside the stack-up during lamination, so thickness variation between panels translates directly into impedance variation. On a controlled impedance board the tolerance on dielectric thickness is often the largest single term in the impedance budget.

This is why impedance controlled boards are usually specified with a tighter thickness window and with test coupons in the panel border. A coupon measures the actual dielectric spacing and the resulting impedance, which is far more informative than measuring the finished thickness of the board alone. The relationship between stack-up and tolerance is explored further in this guide to high speed design rules.

Thickness and Connectors, Edge Fingers and Card Guides

Edge connectors and card guides impose hard limits. A board that is too thick will not enter the slot, and one that is too thin will not make reliable contact with the fingers or will sit loosely in the guide. The tolerance on the connector is usually tighter than the tolerance the laminate process can hold on its own.

Where the board is at the edge of a connector’s range, the fabricator may need to select a construction that targets the middle of the window rather than the nominal value, or to measure and sort panels. Both options add cost, and both are much easier to arrange if the requirement is known at quotation rather than discovered at final inspection.

Thickness and Routing, Scoring and Depaneling

Mechanical routing removes material in several passes, and a thicker board needs more passes or a slower feed. The router also leaves a rougher edge on a thick stack because the bit deflects more, which matters where the edge is a controlled dimension or where the board will be inserted into a guide.

V-scoring behaves differently, because the depth of the score is a fixed fraction of the board thickness. A variation in thickness therefore becomes a variation in the depth of the remaining web, which changes how the board separates and how much stress the components see at depanelisation. The methods and constraints are described in this guide to board outline tolerance.

Thickness and Assembly

Assembly processes are designed around a thickness. A wave solder pallet is machined to accept a specific board, stencil printing assumes a flat and consistent surface, and a reflow profile is tuned to the thermal mass that the thickness represents. A heavier board takes longer to heat, so a thickness change can shift the profile enough to affect the joint quality.

Thickness also interacts with warpage. A thick, asymmetric stack holds more stress and moves more when it is heated, and the result can be a board that will not sit flat on a conveyor or under a stencil. Balanced construction and a controlled press cycle reduce that risk, and the design side of it is covered in this guide to choosing the layer count for a board.

Measuring Thickness Correctly

Finished thickness should be measured with a micrometer or a suitable gauge at several defined points on the panel, away from the edges and from plated areas that add local thickness. Measuring one point in the middle of a panel and reporting that as the board thickness hides the distribution, which is the part that actually causes assembly problems.

Coupons that include the full stack, including the surface finish and the plating, give a more accurate figure than a calculation from nominal values. Where the board will be measured by a customer with a different method, agreeing the method and the measurement points in advance avoids a disagreement about a result that is outside tolerance by a few micrometres.

Specifying and Communicating Requirements

The specification should state the finished thickness, the tolerance, the measurement method and whether the requirement is driven by a connector, a card guide or impedance. Stating the reason lets the fabricator make sensible choices about the construction rather than simply targeting a number that may not be achievable at that tolerance.

For most designs a tolerance of plus or minus ten percent of the nominal finished thickness is achievable without special measures, and tighter windows should be justified by a real mechanical or electrical need. The requirement belongs in the fabrication notes together with the stack-up, so that it is reviewed at the same time as the rest of the construction.

FAQ

Why is my board thicker than the design value? Plating, solder mask and the surface finish all add material to the finished board, and the laminate construction may have been built to a different target. Finished thickness is a process result rather than the sum of nominal values, so it should be measured on a coupon rather than assumed from the stack-up.

Does thickness tolerance affect impedance? Yes, because the dielectric spacing is set during lamination and the impedance depends on the distance to the reference plane. Panel to panel thickness variation therefore produces panel to panel impedance variation, which is why impedance controlled boards are usually given a tighter thickness window and a test coupon.

What tolerance can a fabricator normally hold? Around ten percent of the nominal finished thickness is routine for most constructions. Tighter windows are possible with selected materials and closer process control, but they raise cost and lead time, so the requirement should be justified by a connector, a card guide or an impedance target rather than chosen by default.