Finished Thickness: 6 Rules for PCB Tolerance

Finished thickness is the dimension a board has after every fabrication step is complete, and it is the value that a connector, a card guide or an enclosure will actually see. It is not the laminate thickness ordered, and it is not the copper weight on the drawing. It is a stack of contributions that add up.

Each contribution comes from a different process with its own tolerance, and the total is what the finished thickness tolerance has to cover. Understanding that stack is what allows a sensible callout on the drawing, rather than a number that forces every process into its tightest window at once.

Finished thickness measured with a micrometer across a PCB stack

What Finished Thickness Includes

The stack begins with the laminate and the prepreg that bond the layers, continues with the copper foil and the plated copper added in the holes and on the surface, and finishes with the solder mask and the surface finish. Every one of those contributes to the final dimension.

On a multilayer board the dielectric is pressed, so its thickness depends on resin flow and on the copper pattern it flows around. That is why the same stackup can produce a different thickness on a panel with heavy copper than on one with sparse copper, even when the material lot is identical.

Copper Plating and Its Contribution

Plated copper adds to the surface thickness as well as to the hole wall, and the amount added is controlled by the plating time and current density. A heavy copper board plated for a longer period gains more on the surface, which then moves the finished thickness by tens of microns.

Plating distribution is not perfectly uniform across a panel, so the edges of the panel can be thicker than the centre. Where the finished thickness is tight, the measurement position should be stated, and the acceptance decision should be made on the same positions from lot to lot, as described in copper thickness measurement.

Dielectric, Prepreg and Laminate Build Up

Prepreg is supplied with a resin content and a flow figure, and the pressed thickness depends on how much resin leaves the bond line during the press cycle. More flow means thinner dielectric, which is a benefit for impedance and a risk for resin starvation at the edges.

Laminate cores have their own thickness tolerance, and stacking several cores multiplies the variation. A four layer board made from two cores has half the accumulated error of an eight layer board made from four, which is a reason why the finished thickness tolerance broadens as the layer count rises.

Solder Mask and Surface Finish on Top

Solder mask thickness over the finished copper adds a small but measurable amount, typically in the range of 15 to 30 microns across the masked area. Its own thickness varies with the coating method and with the copper topography beneath it.

Surface finish adds very little on most processes, though hot air levelling and thick immersion deposits can be measurable. The fabrication notes used by the shop should state which layers are included in the callout, because a drawing that names a finished thickness without that detail cannot be checked unambiguously.

Measurement Methods and Their Limits

A micrometer or a contact gauge measures the total and is fast, but it cannot say which layer moved. A microsection shows the individual layers and the plating thickness, and it is the reference method when a result has to be explained rather than only recorded.

Non contact methods based on eddy current or X-ray fluorescence can report copper plating and solder mask thickness separately and are useful in production, but they need calibration against the section before the numbers can be trusted. The trade offs between methods are the same ones that apply when a shop measures board quality on a routine basis.

Microsection image showing copper plating and solder mask thickness

Tolerance Stack and Impedance Control

The tolerance stack sets how much of the finished thickness variation must be absorbed by the connector, the card guide or the enclosure that receives the board. Where the stack is large, the mechanical interface has to be designed with clearance, because a board that is at the top of its tolerance will not fit a slot cut to the nominal value.

On controlled impedance boards the same stack is measured for a different reason. Dielectric thickness sets the impedance, so the variation that moves the finished thickness also moves the impedance, which is why the data from an impedance coupon and the thickness record belong together.

Press Cycle, Plating and Drift Over a Lot

Drift across a production lot is the practical enemy of finished thickness control. A press that runs slightly hotter than the qualification build will produce thinner dielectric, and a plating bath that has aged will deposit more slowly. Neither shows up on a single board, and both appear as a slow trend across a lot.

That is the argument for charting the measurement rather than only comparing it to limits. A value that is inside the tolerance but has moved consistently in one direction over ten panels is already a signal, and catching it then costs far less than scrapping a batch at final inspection.

Choosing the Right Callout for the Drawing

The drawing should state the nominal finished thickness, the tolerance and the layers included. Where a heavy copper build is involved, the design rules for heavy copper work should be consulted, because the plating time that makes the copper current carrying also adds to the finished dimension.

A callout that cannot be met by the process will be met by negotiation or by sorting, and both cost time. Agreeing the stack with the fabricator before the artwork is released is the cheaper route, and it is the same principle as any other tolerance agreed at design review.

Records, Coupons and Final Inspection

Keep the measurement position, the method and the instrument with the result, and retain a coupon from the lot where the product requires it. The acceptance limits and the test methods published by the IPC give a common language for that record, and quoting the revision prevents two sites from measuring against different editions of the same document.

Final inspection should check the finished thickness on the panels that carry the most copper and on those that carry the least, because those two extremes bracket the lot. A result from the middle of a panel describes the middle of the panel and nothing else.

FAQ

Is finished thickness the same as laminate thickness? No. Laminate is only one contribution, and plating, mask and finish all add to it. A drawing that quotes laminate thickness and a drawing that quotes finished thickness describe two different dimensions.

How much does solder mask change the finished thickness? Typically around 15 to 30 microns over the copper, depending on the coating method and the copper topography. On a tight tolerance it is a measurable share of the budget and should be named in the callout.

Why does the finished thickness vary across one panel? Because plating distribution, resin flow and mask thickness all vary with position, and copper density changes from area to area. The fix is to define the measurement position rather than to expect a perfectly uniform panel.

Leave A Comment