Impedance Boards and How to Choose a Fabricator

A controlled impedance board is a normal board with one additional requirement: the characteristic impedance of specific traces must fall within a stated band. That requirement propagates backwards through the whole fabrication process, because impedance depends on trace width, dielectric thickness and dielectric constant, and every one of those has a tolerance of its own.

What Impedance Control Really Requires

Impedance is not a property the fabricator adds at the end. It is a consequence of the geometry and the material, so controlling it means controlling the width of the trace after etching, the thickness of the dielectric after lamination and the dielectric constant of the material. Each of those three has a distribution, and the impedance distribution is the combination.

That is why an impedance requirement is a request for process control rather than for a measurement. A fabricator who can hold the three contributors within tight limits will produce boards within the impedance band as a matter of course.

The Stackup Is the Contract

The stackup is where the impedance requirement is expressed. It specifies the number of layers, the dielectric thickness between each pair and the copper weight, and those figures determine the trace width needed for each target impedance.

A stackup supplied by the designer is a proposal; a stackup agreed with the fabricator is a commitment. Fabricators have preferred constructions using materials they stock and processes they have characterised, and using one of those constructions is both cheaper and more likely to hold the impedance. Our layer assignment notes describe how the arrangement follows from the routing requirements.

<img src="https://www.gopcba.com/wp-content/uploads/2026/05/IOT-Electronics.jpg" alt="impedance board stackup with controlled impedance test coupons” />

Why Test Coupons Matter

A coupon is a structure on the panel border that mirrors the traces on the production boards. Its purpose is to be measured, and if necessary destroyed, so that the boards themselves do not have to be.

Impedance coupons are usually designed as a set of lines of known length, sometimes with different widths so that several impedances can be verified at once. A time domain reflectometer measures their impedance, and the result is recorded against the panel. Without coupons, the only way to verify the impedance is to cut up a customer board, which nobody wants to do. The verification approach is described further in our article on PCB electrical test coverage.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/pcb15.jpg" alt="fabricator measuring a controlled impedance test coupon” />

Questions to Ask a Fabricator

The first question is what impedance tolerance they can hold as standard, and the second is how they measure it. A supplier who answers both with a number and a method is operating a controlled process. One who answers with a general assurance is not.

The next questions concern the stackup: which laminate grades are stocked, which dielectric thicknesses are standard, and what the thickness tolerance is on each. Dielectric thickness tolerance is often the largest single contributor to impedance variation, and a fabricator who does not know their figure cannot control the impedance.

Material and Its Effect on Impedance

The dielectric constant of the laminate determines the impedance for a given geometry, and it varies between materials and even between batches of the same material. Low loss laminates are usually specified with a tighter tolerance on dielectric constant, which is one reason they are used in high speed work.

Where the design uses a material the fabricator does not normally stock, the impedance may end up outside the band even though the geometry is correct. Confirming the material and its dielectric constant tolerance before the layout is frozen avoids that outcome. Our high Tg material notes cover the related thermal and mechanical properties.

Cost and What Drives It

Controlled impedance adds cost in three ways: tighter process control, coupons that must be designed and measured, and the possibility of scrap when a panel falls outside the band. Of those, the coupons are the smallest and the scrap risk is the largest.

Impedance is held more easily on a thicker dielectric and a wider trace, because the same proportional error in each dimension produces a smaller effect on the result. A design that demands both high impedance and very fine traces should expect to pay more and to have fewer suppliers able to quote.

Documenting the Requirement

The fabrication drawing should state the target impedance, the tolerance, the layers and the coupons on which it will be measured. Where several impedances are required, each should be listed separately rather than summarised.

Stating the requirement this way makes the board easier to quote, because the fabricator can see immediately whether the stackup supports it. It also protects the buyer, since a requirement that was stated in writing is one that can be verified against the delivered result. Our design tolerances article describes how the rest of the drawing should be framed.

Coupon Design and What to Include

A generic coupon verifies that the process is in control, which is useful. A coupon designed for the product verifies the specific impedances the product depends on, which is more useful. The difference is a few minutes of layout time and a much clearer answer when the panels arrive.

Where several impedances are required, the coupon should include a line for each, at the width and on the layer the design uses. Coupons are also where the dielectric thickness and the plating thickness are measured, so they serve more than one purpose on the same piece of panel border. Treating them as a design deliverable rather than as a fabricator detail is what makes the impedance requirement verifiable at all.

When Impedance Control Is Not Needed

Not every design benefits from impedance control, and specifying it unnecessarily adds cost without improving the product. A board whose signals are electrically short, whose edges are slow and whose interfaces are forgiving will work without it.

The decision follows from the rise time and the trace length rather than from the clock frequency. A low frequency design with fast logic edges and long traces can need control, while a higher frequency design with short traces and slow edges may not. Our high speed routing requirements article sets out how that judgement is made.

Working With the Fabricator Through a Revision

A stackup change between revisions invalidates the impedance characterisation that the fabricator built up, so the first panels of the new revision are less certain than the last panels of the old one. Knowing this makes it worth batching revisions rather than issuing them one at a time.

Where the stackup must change, the fabricator should be told why, because they may be able to propose an alternative that preserves the construction while meeting the new requirement. The alternative is often cheaper than a new characterisation, and it is only available if the reason for the change is visible. Our copper thickness notes describe one common reason for such a change.

FAQ

Can impedance be held by adjusting trace width alone? No. Width is the easiest variable to adjust, but the dielectric thickness and the dielectric constant set the impedance as much as the width does.

How tight an impedance tolerance is normal? Ten percent is a common standard. Tighter figures are available and cost more, and only interfaces that genuinely need them should ask for them.

Should the designer supply the stackup or the fabricator? The designer states the requirement and the routing constraints. The fabricator is usually the better party to propose the construction that meets them.