Backdrilling vs. Laser Microvias, Part 2: Reliability, DFM and the Case for Hybrid Design

The first part of this discussion compared backdrilling and laser microvias primarily through signal integrity, cost, manufacturability and application fit. Part 2 examines a different question: which technology will remain reliable after the board leaves the laboratory?

PCB DESIGNELECTRONIC SYSTEM DESIGNPCB MANUFACTURING

Atanu

10/1/20267 min read

a table topped with lots of electronics on top of a wooden table

A via is both an electrical interconnect and a mechanical feature. It carries signal or power, connects copper structures across layers and experiences thermal expansion, vibration, shock, plating stress and current-induced heating. A technology that looks excellent in a simulation can still fail if the stackup, fabrication data or qualification process is not controlled.

The most useful conclusion is not that one technology replaces the other. Backdrilled PTH vias and laser microvias occupy different positions in the design space. Backdrilling is generally stronger, cheaper and better suited to current-carrying or harsh-environment applications. Microvias provide greater routing density and lower parasitics, particularly in fine-pitch BGA and very-high-speed regions. A hybrid architecture can allocate each technology where it creates the greatest value.

The automotive radar example

The source material presents a 16-layer automotive radar ECU using a hybrid via strategy. The board had to support 77 GHz radar signals, dense routing beneath a BGA and the reliability expectations associated with automotive electronics.

Microvias were used in the BGA region, where their small geometry enabled denser escape routing. Backdrilled PTH vias were used for longer-distance connections. The reported results were:

· Return loss at 77 GHz: -18 dB for backdrilled PTH routes and -22 dB for microvias in the BGA region.

· Jitter: 7 ps for backdrilled routes and 5 ps for microvia routes.

· Cost: 40% lower than an all-microvia design.

· Board area: 15% smaller than an all-backdrilled design.

This is a strong illustration of system-level optimisation. An all-microvia design might improve density everywhere, but it would apply an expensive HDI process even where a conventional PTH solution was adequate. An all-backdrilled design might reduce cost, but it would require more board area to complete the BGA escape. The hybrid design assigns microvias to the space-constrained region and backdrilling to the longer, mechanically demanding routes.

Reliability under thermal and mechanical stress

Reliability becomes especially important in automotive, aerospace, industrial and medical products. These products may experience large temperature excursions, repeated vibration, mechanical shock or long operating lifetimes.

The comparison reports the following results:

· Thermal cycling from -55 degrees C to +125 degrees C for 1,000 cycles: no failures reported for backdrilled PTH vias, compared with a reported 2–5% failure rate for stacked microvias.

· Drop testing from 1.5 metres across six faces: backdrilled PTH vias passed, while microvias were described as marginal because small structures can shear under stress.

· Vibration from 10 Hz to 2,000 Hz at 20 G: backdrilled PTH vias passed; single-high microvias passed, while stacks above two levels were reported to fail.

The underlying physical reason is that the PTH begins as a full-board mechanical penetration. Backdrilling removes an unused electrical barrel section, but the via remains a substantial plated structure. A microvia is much smaller and shallower. Its performance depends heavily on the quality of laser formation, desmear, plating, resin system, pad capture and the geometry of any stacked structure.

The reported 2–5% stacked-microvia failure rate should not be treated as a universal value. Actual reliability varies by process and qualification method. Nevertheless, it highlights an important design rule: stacked microvias require evidence. The designer should obtain fabricator data and perform product-specific testing rather than assuming that a successful electrical coupon guarantees field reliability.

Current capacity and electromigration

A second major difference is current capacity. Signal routing and power delivery should not be treated as the same via problem. A small microvia may be ideal for a high-speed signal, but its cross-sectional area can be insufficient for a power transition.

The source comparison gives the following illustrative values:

· A 12 mil PTH via: approximately 1–2 A.

· A 4 mil microvia: approximately 0.2–0.5 A.

· Reported electromigration lifetime: more than 10 years at 1 A for the PTH comparison, versus less than 5 years at 0.5 A for stacked microvias.

These values depend on copper thickness, temperature, via geometry, duty cycle, plating quality and the required reliability target. They should therefore be used as design-direction figures rather than universal specifications. The practical message is clear: use PTH vias, parallel via arrays, copper-filled vias or another qualified structure for higher-current power nets. Reserve ordinary microvias primarily for low-current signals unless the power structure has been specifically qualified.

Backdrill DFM checklist

Backdrilling is relatively straightforward compared with an HDI build, but it must be specified precisely. A production-ready design should address the following points:

· Identify the layers to be backdrilled in the fabrication notes.

· Provide a separate backdrill drill file, commonly using Excellon data with depth information.

· Maintain at least 10 mils from the backdrill edge to adjacent copper.

· Maintain at least 6 mils between adjacent backdrills.

· Keep at least 10 mils of dielectric above the target layer to prevent over-drilling into active copper.

· Specify X-ray or cross-section inspection for residual stub verification.

· Consider epoxy filling where necessary to prevent solder wicking or address other assembly concerns.

The most common risk is not the mechanical operation itself; it is an incomplete manufacturing definition. If the CAD data, drill files and fabrication notes do not agree, the supplier may not know which side to drill, which depth to use or which vias are intentionally excluded.

Laser microvia DFM checklist

Laser microvias require a more integrated stackup and process definition. Key checks include:

· Specify a laser-drillable prepreg or dielectric system approved by the fabricator.

· Maintain an aspect ratio of no more than 1:1 unless a different process has been qualified.

· Use a pad at least twice the via diameter as a starting point.

· Limit stacked microvias to two or three levels unless reliability testing supports more.

· Use at least 50 micrometres of staggered offset where required.

· Balance copper distribution to reduce warpage during sequential lamination.

· Specify copper or non-conductive fill for via-in-pad structures.

Microvia DFM is not just a hole-size exercise. The complete sequence includes laser ablation, cleaning, plating, lamination, registration and sometimes filling and planarisation. Each step affects the final reliability of the connection.

Future technology directions

Several emerging technologies may change how designers control via stubs and routing density.

Any-layer HDI

Any-layer HDI with stacked microvias is being adopted for extremely dense designs. It can support 0.4 mm pitch BGA fan-out and 112 Gbps or faster SerDes routing. Its main barrier is cost. The source material estimates three to five times the cost of standard PTH in some implementations, which limits adoption to premium smartphones, medical implants and other high-value products.

Plasma etching

Plasma etching is presented as an alternative to mechanical backdrilling. Its potential benefits include depth control of approximately plus or minus 1 mil, compared with approximately plus or minus 2 mil for mechanical backdrilling. It also avoids mechanical drilling stress and may be suitable for boards thinner than 1.6 mm, where conventional backdrilling becomes difficult.

The source describes plasma etching as early-stage commercialization, with volume production projected around 2027–2028. This is a forward-looking estimate, so designers should confirm the actual production maturity and qualification status before selecting it for a critical product.

Additive and 3D-printed PCBs

Additive PCB manufacturing could embed vias and conductive paths without conventional drilling. This offers potential benefits in placement freedom, material efficiency and stub elimination. Current constraints include limited practical layer counts, immature signal-integrity performance and reliability that is not yet comparable with established subtractive processes.

A decision matrix in plain language

Choose backdrilling when the board is thick, the layer count is moderate to high, cost matters, the environment is harsh, or power vias must carry more than about 1 A. It is also a strong choice for networking, server, aerospace and industrial boards operating at moderate-to-high data rates where equalisation can manage the remaining channel loss.

Choose laser microvias when the board is thin, the BGA pitch is below 0.8 mm, routing density is the dominant constraint, or the signal path operates at 56 Gbps and above with very limited parasitic budget. Microvias are especially valuable in premium compact products where board area is more expensive than fabrication cost.

Choose a hybrid when the board contains both dense package escape regions and long-distance routes. This is often the most practical architecture for automotive radar, 5G modules and complex networking equipment.

Application recommendations

For networking and server boards operating around 25–56 Gbps, backdrilling is generally the cost-effective starting point. For 5G mmWave modules, microvias can be reserved for critical RF paths while backdrilling handles digital interconnects. Automotive radar benefits from a hybrid design that combines BGA density with PTH robustness. Smartphones and wearables with 112 Gbps-class links and severe space constraints are natural microvia applications. Aerospace and defence designs with high reliability and moderate speed often favour backdrilling. Industrial IoT products typically benefit from the lowest-cost solution that meets their signal-integrity requirement.

A practical qualification workflow

Begin by separating electrical, mechanical and power requirements. Identify the channels with the smallest eye margin, the highest frequency content and the most severe EMI constraints. Then classify vias as signal, power, mechanical or mixed-function structures.

Model the proposed stackup and via geometry. Ask the fabricator to review the design before layout freeze, including drill depth, registration, prepreg, plating, microvia stacking, via fill and inspection. Build representative test coupons where the product is safety-critical or reliability-sensitive.

Finally, validate the finished board using measurements appropriate to the application: TDR, insertion and return loss, eye diagrams, jitter, crosstalk, thermal cycling, vibration, drop testing and current stressing. A technology choice becomes robust only when both electrical performance and manufacturing reliability are demonstrated.

Conclusion

Backdrilling remains a highly effective and economical method of via-stub control. It preserves much of the mechanical strength and current capacity associated with PTH construction while reducing the electrical discontinuity that causes high-speed problems. Laser microvias deliver superior routing density and lower parasitics, but their benefits come with higher process complexity and greater dependence on stackup and reliability control.

The strongest design strategy is usually selective. Use backdrilling for long routes, power transitions and harsh environments. Use microvias under dense BGAs and on the most demanding RF or ultra-high-speed paths. Combine the two when the system requires compact geometry without accepting unnecessary cost or reliability risk.

References

1. JLCPCB — https://jlcpcb.com/blog/via-stubs-high-speed-pcb

2. SierraConnect — https://sierraconnect.protoexpress.com/t/backdrilling-vs-blind-vias/3198

3. Aivon — https://www.aivon.com/blog/pcb-manufacturing/pcb-back-drilling-vs-blind-and-buried-vias-which-is-right-for-your-design/

4. ALLPCB — https://www.allpcb.com/blog/pcb-manufacturing/laser-drilling-vs-mechanical-drilling.html

5. ProtoExpress — https://www.protoexpress.com/blog/back-drilling-pcb-design-and-manufacturing/

6. PCBSync — https://pcbsync.com/pcb-back-drilling/

7. ALLPCB — https://www.allpcb.com/blog/pcb-manufacturing/mastering-microvias-a-comprehensive-guide-to-pcb-laser-drilling.html

8. ALLPCB — https://www.allpcb.com/allelectrohub/via-stubs-understanding-and-mitigating-signal-integrity-issues

9. JHDPCB — https://jhdpcb.com/blog/back-drilling-manufacturing/

10. LinkedIn reference — https://www.linkedin.com/posts/amit-bahl-sierra-circuits_highspeedpcb-pcbdesign-pcbmanufacturing-activity-7469780926366031872-Sf73

11. Electronics Stack Exchange — https://electronics.stackexchange.com/questions/670425/drilled-vs-micro-vias