Discover why controlled impedance in UHDI beats subtractive etch on tolerance. Explore the loss, material, and stack-up factors to discuss with your fabricator.
Impedance control has always come with an asterisk. Most of the industry works to a tolerance of plus or minus 10 percent, and historically, everyone accepts that as the cost of building with subtractive etch processes. Ultra HDI changes that math enough that it's important for designers to understand the why and how that can apply to traces that are larger than the standard Ultra HDI feature sizes.
As the semi-additive and modified semi-additive processes were being introduced and developed for use in low-volume, high-mix production environments, much of the focus was on processing the trace and space features at levels below 75 microns. What hasn’t had quite as much attention is the impact on controlled impedance.
Today, much work is being done developing thinner dielectric materials, learning to process those new materials and understanding the impact of narrower traces on controlled impedance traces.
Now that the industry is comfortable with the fact that SAP and mSAP processes can consistently and reliably provide plus or minus 3% controlled impedance. Let’s talk about the fact that these processes aren’t only capable of ultra fine line, they can also be used on much larger feature sizes, bringing this benefit to a far-reaching set of designs.
Standard subtractive processing etches copper away from the surface down, and that process doesn't leave a clean rectangular trace behind. It leaves something closer to trapezoidal, wider at the base than at the top, because the etchant works on the sidewalls the whole time it's cutting through the copper. That geometry is baked into every impedance calculation on a subtractively etched board, and it's part of why the plus or minus 10 percent tolerance became the industry default in the first place. It's not that anyone's process is bad. The cross-section itself introduces variability that's hard to design around.
Ultra HDI's additive or semi-additive processes don't have that issue. If you can picture it, copper is built up in a resist “trench” rather than etched down from a solid sheet and the sidewalls come out very close to vertical instead of trapezoidal. Line widths hold within a couple of microns of design intent. The result? Designs that used to land somewhere in a 10 percent tolerance window on the first build are now landing within 3 percent, often on the very first prototype, without any modification for etch factors.
Tighter impedance control isn't only a number that looks better on a spec sheet with Ultra HDI process. It also can mean lower insertion loss, cleaner signal transitions, and better coupling behavior on differential pairs. The copper geometry the simulation assumed is a lot closer to the copper geometry that actually gets built. Pair that with the low-loss materials increasingly available for Ulta HDI builds, liquid crystal polymer and advanced PTFE among them, and you get performance that traditional subtractive boards can't reliably reach, particularly for antennas, filters, and high-speed links where every fraction of a dB matters.
Thin dielectrics factor in here, too. Fine line widths paired with thin, well-controlled dielectrics let designers hit specific impedance targets, 50 ohms, 40 ohms, or differential impedances up around 100 ohms, while still keeping dielectric loss and dissipation factor low. That combination, tight line control plus thin low-loss dielectric, is difficult to achieve reliably with subtractive processing regardless of how careful the design is.
One detail to keep in perspective: short escape routes out of a BGA, the kind of routing that gets you from a fine-pitch pad out to a wider impedance-controlled line, generally aren't long enough to meaningfully affect overall impedance. If your design is mostly short runs off a dense BGA field, the precision Ultra HDI offers may matter less than it would for a design with long, sustained impedance-controlled traces. Knowing which situation you're in changes how much of this conversation applies to a given design.
Getting the benefit of tighter impedance control means having the dielectric constant, material choice, and spacing conversation with your fabricator before the design is finalized. The materials that make this level of control possible aren't universal. Availability, lead time, and cost vary, and a fabricator who works with these processes regularly can help you land on a material and stack-up combination that hits your electrical targets without chasing something exotic and hard to source.