Why Impedance Drops After Solder Mask

A controlled impedance board arrives from the fabricator with a test report showing the coupon measured exactly on target. Then the same structure is measured again after solder mask is applied, and the impedance reads lower than the design value. Nothing about the copper changed. What changed is the dielectric environment above the trace, and understanding why that happens is the difference between chasing a phantom error and compensating for a real one.

The Measurement That Surprises Everyone

Impedance is not a property of the copper trace alone. It is a property of the transmission line structure, which includes the trace geometry, the dielectric beneath it and everything above it. On a microstrip, the top surface of the trace sees air, and the electric field distributes itself partly in air and partly in the laminate below. Adding a coating replaces part of that air with a solid material of higher dielectric constant.

That substitution increases the effective dielectric constant of the structure, which increases capacitance per unit length, which lowers the characteristic impedance. For a typical solder mask on a typical microstrip, the drop is usually a few percent, and it is repeatable once the coating thickness and material are fixed. It becomes a problem when the mask effect is not included in the design or in the coupon measurement.

How a Coating Changes the Effective Geometry

Physically, the coating adds a thin layer of dielectric directly over the conductor. Because the field is concentrated most strongly at the trace edges and near the trace surface, a coating of only twenty or thirty micrometres still removes a meaningful proportion of the air that was carrying part of the field. The effect is proportionally larger on narrow traces, where a greater fraction of the total field is close to the surface.

This is why the same mask thickness affects a fine pitch trace more than a wide power trace. It also explains why the change is smaller on a stripline, where the conductor is already surrounded by laminate on both sides and the field never sees air. Our discussion of microstrip and stripline routing covers how the two structures differ in their sensitivity to the environment above and below the trace.

Dielectric Constant of Solder Mask Materials

Solder mask is an epoxy based coating, and its dielectric constant is typically higher than that of the laminate beneath it. Values of around 3.5 to 4 are common, against roughly 4.2 to 4.5 for standard FR-4 at low frequency, but the comparison at high frequency is less favourable, because both materials show rising loss and changing permittivity as frequency increases.

The material is also not uniform. Different mask colours, different formulations and different cure schedules all change the permittivity slightly, and a fabricator may substitute one approved material for another without treating it as a design change. If the impedance margin is thin, the mask material belongs in the specification rather than in the fabricator’s discretion. Our notes on PCB dielectric constant explain how permittivity feeds into impedance and how it varies with frequency.

Thickness Control and Coverage Variation

Mask thickness over a trace is not a single number. It depends on the trace width, the spacing to adjacent copper, the screen or spray process and the position on the panel. Where a large copper area sits next to a narrow trace, the mask flows differently and the resulting thickness over each is different. Designers who assume a uniform coating are working with an average that may not apply to the traces they care about.

This variation is one reason a measurement taken at one point on a coupon may not represent the worst case on the panel. Where impedance tolerance is tight, the useful approach is to measure more than one structure, on more than one panel, and to record the spread rather than only the mean.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/Surface-Finish.jpg" alt="TDR test coupon used to measure controlled impedance on a PCB panel” />

Single Ended and Differential Effects Compared

Single ended microstrips see the mask effect directly, because the field above the trace changes. Differential pairs experience it slightly differently, because the odd mode field is concentrated between the two conductors and is therefore less exposed to the air above. The even mode, by contrast, sees more of the surrounding environment, so the two modes shift by different amounts, which changes the differential impedance and the modal conversion behaviour.

In practice the effect on a well designed differential pair is smaller than on a single ended trace, but it is not zero. Where a link has a tight differential impedance budget, the mask effect should be included in the stackup calculation rather than discovered during measurement.

Compensating in the Stackup and the Artwork

Compensation is straightforward once the effect is quantified: the fabricator adjusts the trace width or the dielectric spacing so that the finished, masked structure reaches the target impedance. This is normally done by measuring coupons in both states, bare and coated, and building the difference into the model. The artwork then carries the slightly narrower trace that the model requires.

The important point is that the coupon and the product must be measured in the same condition. A coupon measured before coating and a product measured after coating will disagree, and the disagreement says nothing about manufacturing quality. Our notes on impedance tolerance describe how to define the measurement condition so that the tolerance has a consistent meaning for designer and fabricator alike.

There is a practical limit to how much compensation is wise. If the model requires a trace width close to the fabricator’s minimum, the design has no room left for process variation, and the tolerance will be missed in production even though the nominal value is correct. When that happens, the better answer is usually to increase the dielectric spacing or to move the structure to an inner layer, where the mask effect largely disappears and the tolerance becomes easier to hold.

Verifying With Test Coupons

A coupon is only useful if it represents the product. It should use the same trace geometry, the same stackup and the same copper weights as the board, and it should include both single ended and differential structures if the product uses both. Time domain reflectometry then measures the finished impedance and shows not only the value but also any variation along the length.

Recording the coupon results against the panel and the date builds a history that shows whether the process is stable. When a later batch measures differently, that history is what allows a fabricator and a designer to agree quickly on whether the difference comes from the material, the process or the measurement setup itself.

FAQ

How much does solder mask lower impedance? For a typical single ended microstrip with a standard coating thickness, expect a few percent, often in the range of two to five percent depending on the trace width and the mask material. The exact figure comes from the coupon measurement rather than from a rule of thumb.

Should I design the trace narrower to compensate? The compensation is normally applied by the fabricator after measuring coated coupons, because they control the process and know the coating thickness they will apply. The designer’s job is to specify the target impedance and the measurement condition, not to pre-compensate by guesswork.

Does the mask affect striplines as well? Much less, because the conductors are already embedded in laminate and the field never sees air. The effect is not zero because the coating still changes the boundary conditions at the surface, but it is small enough that many fabricators ignore it for inner layer structures.