Backdrilling vs. Laser Microvias, Part 1: Choosing the Right Way to Control Via Stubs
As PCB data rates increase, the via is no longer a passive hole in the board. It becomes part of the transmission channel. A plated through-hole that extends beyond the signal layer can leave a copper barrel segment—known as a via stub—that behaves as a discontinuity. At sufficiently high frequencies, that stub can create resonances, reflections, insertion loss, jitter, eye closure and unwanted radiation. Two widely used ways to control the problem are backdrilling and laser microvias. Both can improve signal integrity, but they solve different design problems. Backdrilling modifies a conventional plated through-hole after plating. Laser microvias create short, blind interconnects between adjacent layers and avoid the full-length barrel altogether.
PCB DESIGNPCB MANUFACTURING
Atanu
9/28/20267 min read
As PCB data rates increase, the via is no longer a passive hole in the board. It becomes part of the transmission channel. A plated through-hole that extends beyond the signal layer can leave a copper barrel segment—known as a via stub—that behaves as a discontinuity. At sufficiently high frequencies, that stub can create resonances, reflections, insertion loss, jitter, eye closure and unwanted radiation.
Two widely used ways to control the problem are backdrilling and laser microvias. Both can improve signal integrity, but they solve different design problems. Backdrilling modifies a conventional plated through-hole after plating. Laser microvias create short, blind interconnects between adjacent layers and avoid the full-length barrel altogether.
What is backdrilling?
Backdrilling, also called controlled-depth drilling, is a secondary mechanical drilling process. A standard through-hole via is first drilled and plated. A larger drill is then introduced from the opposite side of the board and removes the unused portion of the copper barrel. The goal is not to remove the entire via, but to leave only a very short residual stub.
A typical backdrill may use a drill diameter 8–10 mils larger than the original via. The drilling depth is controlled by CNC equipment, and the remaining stub is checked by inspection or cross-sectioning. A common design target is less than 10 mils, or 0.25 mm, with about 6 mils, or 0.15 mm, used as a more aggressive target.
Backdrilling is attractive because it works with standard PTH construction and can often be added to an existing multilayer design with relatively limited layout changes. The designer must, however, provide suitable clearance around the enlarged backdrill and ensure that the fabricator can control the depth accurately.
What are laser microvias?
A laser microvia is a small blind via formed by laser ablation. A CO2 or UV laser removes dielectric material, the hole is metallised, and the resulting via connects adjacent layers. Typical diameters range from about 50 to 150 micrometres. When the aspect ratio is held to 1:1 or less, the depth is commonly limited to 100 micrometres or less.
Microvias may be staggered or stacked. Staggered structures offset adjacent vias, while stacked structures place one microvia directly above another. Stacking can support compact vertical interconnection, but reliability must be managed carefully. The source report recommends treating two or three stacked levels as a practical limit unless additional qualification demonstrates that a more complex structure is reliable.
Microvias are a natural fit for HDI boards, fine-pitch BGAs and compact products. They reduce pad sizes and allow routing channels that would be difficult or impossible with conventional through-hole vias.
Signal-integrity trade-offs
The electrical reason for removing a stub is straightforward. A stub creates an additional resonant structure attached to the main transmission line. Depending on its length and the frequency content of the signal, it can produce a notch in the channel response and increase reflections. These effects can degrade the eye diagram and increase bit-error risk.
In the comparison covered by the white paper, backdrilling reduced the residual stub below 10 mils and produced approximately -22 dB return loss at 5 GHz, compared with about -25 dB or better for laser microvias. Insertion loss was approximately -1.8 dB for backdrilling and -1.5 dB or better for microvias.
The reported peak-to-peak jitter was approximately 9 ps for backdrilling and 6–8 ps for microvias. Eye height was about 0.78 UI with backdrilling and 0.80–0.85 UI with microvias. These figures indicate a real advantage for microvias, but also show that a properly executed backdrill can deliver a channel that is very close in performance for many applications.
The same pattern appears in parasitics. Backdrilled vias are reported at about 0.2–0.3 pF capacitance and 0.5–0.8 nH inductance. Microvias are lower, at about 0.1–0.2 pF and 0.3–0.5 nH. The difference matters most when the channel budget is extremely tight or when the design operates at very high frequencies.
EMI and crosstalk
A stub can also radiate. Backdrilling removes the stub-end discontinuity and, according to the source comparison, can reduce radiated emissions by more than 20 dB at resonance frequencies. Microvias provide a further improvement because their physical dimensions and loop inductance are smaller.
The reported crosstalk reduction is approximately 10–15 dB for backdrilling and 15–20 dB for microvias. In both cases, the actual result depends on via spacing, reference-plane continuity, return-current paths, stackup geometry and the surrounding routing. No via technology should be evaluated in isolation from the complete interconnect.
Cost and manufacturing
The largest difference between the technologies is often economic rather than electrical. Backdrilling starts with a standard PTH via and adds a controlled-depth drilling step. The source report estimates an additional 5–10% for high-layer-count boards.
Laser microvias often require an HDI construction with laser-drillable prepregs and sequential lamination. Multiple lamination cycles increase process time, fabrication complexity and inspection requirements. For stacked microvias, the report indicates a possible two- to four-times cost increase compared with standard PTH construction with backdrilling.
An illustrative 10-layer prototype run of 100 boards is estimated at approximately $500–$800 with backdrilling and approximately $1,500–$3,000 with laser microvias. These figures are not universal quotations; they are useful as an indication of relative cost direction. Actual pricing depends on board dimensions, copper weights, material system, drill count, yield, volumes and fabricator capability.
Engineering cost also matters. Backdrilling generally requires a backdrill rule in the CAD system, defined target layers, clearance checks and a controlled-depth manufacturing specification. Microvias require a full HDI design approach, including stackup planning, laser-drillable dielectric selection, aspect-ratio limits, sequential-lamination rules and reliability review.
Practical design rules
For backdrilling, useful starting points include:
· Keep the residual stub below 10 mils and target about 6 mils where practical.
· Specify an oversize drill about 8–10 mils larger than the original via.
· Allow at least 10 mils from the backdrill edge to adjacent copper.
· Allow at least 6 mils between neighbouring backdrills.
· Maintain at least 10 mils of dielectric above the target layer to avoid over-drilling.
· Confirm that the fabricator can hold approximately ±2 mils depth tolerance.
For laser microvias, useful starting points include:
· Use diameters in the 50–150 micrometre range, subject to fabricator capability.
· Maintain an aspect ratio of no more than 1:1 unless the process is specifically qualified.
· Use pads of approximately 200–250 micrometres where appropriate.
· Select laser-drillable prepreg and verify compatibility with the resin system.
· Treat stacked microvias beyond two or three levels as a reliability topic requiring qualification.
· Use staggered structures and an offset of at least 50 micrometres when required by the process.
These are design starting points, not substitutes for a fabrication drawing and supplier-specific design rules. The final values should be agreed with the selected PCB manufacturer before layout release.
When should each be used?
Backdrilling is generally preferred for thick multilayer boards, backplanes, midplanes, server motherboards and designs that already use through-hole vias. It is also compelling when the board must support PCIe, high-speed Ethernet or other differential links but the project cannot justify a complete HDI redesign.
Laser microvias are preferred where routing density is the overriding constraint. Fine-pitch BGA escape, smartphone-class products, compact medical electronics and thin HDI boards are typical examples. Microvias can also be justified on critical RF paths where every fraction of a picofarad and nanohenry matters.
A hybrid approach often provides the best result. Microvias can be used beneath a fine-pitch BGA for the first two or three layers, while backdrilled PTH vias support longer-distance routing to connectors. This confines the most expensive process to the area where it creates the greatest layout or electrical benefit.
What the case studies show
In a 12-layer, 25 Gbps networking board, approximately 0.6 mm average stubs caused compliance problems. The initial design had 18 ps peak-to-peak jitter and 0.55 UI eye height. After backdrilling reduced residual stubs below 0.15 mm, return loss improved from -12 dB to -22 dB at 5 GHz, insertion loss improved from -3.5 dB to -1.8 dB, jitter fell to 9 ps and eye height increased to 0.78 UI. The reported cost increase was 8%, and the board passed PCIe compliance testing.
A 10-layer 5G mmWave module provides a different lesson. Unmitigated stubs produced return-loss notches greater than 8 dB at 24 GHz and required additional equalisation. Laser microvias eliminated the notches on the critical RF routing and reduced equalisation-related power consumption by 12%, but cost increased by approximately 150% compared with the backdrilled design.
The conclusion is not that microvias are too expensive to use. It is that their use should be targeted. If only a few RF paths require the lowest possible parasitics, microvias can be reserved for those paths while backdrilling handles the majority of the board.
A disciplined selection process
Start by identifying the channels with the highest frequency content and the smallest eye or loss budget. Model the existing via structure, including the stub length, pad geometry, antipads, reference planes and return-current path. If a controlled-depth backdrill can reduce the stub below the required electrical limit with adequate clearance, it should be evaluated first.
Move to microvias when the board cannot accommodate the required PTH geometry, when BGA escape density is the primary constraint, or when the RF and signal-integrity budget justifies the additional process cost. In either case, obtain fabricator confirmation before freezing the stackup. Depth tolerance, drill registration, dielectric thickness, plating quality, sequential-lamination capability and inspection method can determine whether the theoretical design will work in production.
Finally, validate the finished board. Use TDR where appropriate, insertion- and return-loss measurements, eye diagrams, jitter analysis, crosstalk measurements and the relevant compliance test. A sound design rule is valuable, but measured hardware remains the final arbiter.
Conclusion
Backdrilling and laser microvias are complementary tools. Backdrilling is often the most economical way to remove the dominant electrical problem in a conventional multilayer board. Laser microvias offer greater routing density and slightly lower parasitics, but at higher manufacturing and engineering cost.
For many high-speed products, the strongest strategy is selective rather than absolute: use backdrilling where standard PTH construction is efficient, use microvias where density or RF performance demands them, and combine both when the board contains distinct routing environments.
References
1. Proto Express — https://www.protoexpress.com/blog/back-drilling-pcb-design-and-manufacturing/
2. JLCPCB — https://jlcpcb.com/blog/via-stubs-high-speed-pcb
3. ALLPCB — https://www.allpcb.com/blog/pcb-manufacturing/laser-drilling-vs-mechanical-drilling.html
4. PCB Sync — https://pcbsync.com/pcb-back-drilling/
5. Sierra Circuits — https://sierraconnect.protoexpress.com/t/backdrilling-vs-blind-vias/3198
6. JHDPCB — https://jhdpcb.com/blog/back-drilling-manufacturing/
7. PCBCool — https://pcbcool.com/?p=51495
8. LinkedIn post — https://www.linkedin.com/posts/amit-bahl-sierra-circuits_highspeedpcb-pcbdesign-pcbmanufacturing-activity-7469780926366031872-Sf73
9. Electronics Stack Exchange — https://electronics.stackexchange.com/questions/670425/drilled-vs-micro-vias
10. Advanced PCB — https://www.advancedpcb.com/en-us/resources/blog/understanding-back-drilling-in-pcb-manufacturing/
