Sixty-Eight Cables in a Phone: What Robot Terminals Do to PCB Design

On 12 August 2026, Honor introduced the Robot Phone, priced from 9,999 yuan, with pre-orders reported above 400,000 units. The device runs an agent-oriented operating system with a system-level agent architecture, and integrates a four-degree-of-freedom titanium gimbal into the phone body. The gimbal contains more than one hundred precision parts and four micro motors, the smallest measuring six millimetres in diameter, machined to an accuracy of plus or minus 0.005 millimetres. Inside the chassis there are sixty-eight cables, nineteen of them coaxial, alongside a 7,060 milliampere-hour battery, with total weight held at 248 grams.

A conventional smartphone solves computing, display, communication, and imaging problems. Once a mechanical gimbal is added, the same device must also handle environmental sensing, motion control, visual feedback, and agent-driven decision making. The hardware architecture starts to resemble a robot more than a handset, and the board inside it inherits the consequences.

From Computing Terminal to Embodied Device

The significance of a product like this is not that it adds a movable camera. It is that the device now contains an actuation system, and an actuation system brings with it a set of requirements that consumer electronics have historically avoided.

Motion control requires feedback from position and torque sensing, which must be processed in a closed loop. That loop demands predictable timing, low-noise measurement, and stable power, because a control system that jitters or misreads its sensors cannot position anything accurately. At the same time, the device continues to run radio frequency communication, high-speed display and camera interfaces, and high-current charging, all inside a sealed enclosure the size of a handset.

The pattern is familiar from other product categories. Humanoid robots, AR glasses, and intelligent vehicles are converging on the same architecture in which sensing, computation, and actuation coexist in a confined space. AI servers provide the training capacity and optical interconnect moves the data; handsets, robots, and vehicles are where the resulting intelligence reaches the physical world. As that continues, the printed circuit board shifts from being a carrier for electronic components to being the interconnect platform that joins computation, sensing, and motion.

The Space Paradox: More Mechanics, Less Board Area

Sixty-eight cables, nineteen coaxial runs, four micro motors, and a complex camera module inside a phone create an obvious space problem. Mechanical assemblies consume volume, but the battery, the main processor, and the communication system cannot give up much ground either. What remains for the board is less room than in a device with no moving parts.

The technical answer is to increase interconnect density rather than to reduce functionality, and that has three consequences for board design.

Higher-order HDI and any-layer structures. As board area shrinks, routing has to move into the vertical dimension. Laser microvias and any-layer interconnection shorten signal paths and free routing channels around dense packages, allowing more connections to occupy the same footprint. In the densest regions, fine-line processes producing traces at 0.075 mm and below provide additional capacity.

Stricter impedance control. Camera interfaces, display links, storage buses, and radio frequency paths all impose signal integrity requirements, and they now run alongside a motor drive stage that generates broadband noise. Differential impedance control tightens toward the plus or minus five percent range on demanding designs, because at high data rates a small deviation in line width, dielectric thickness, or copper thickness translates into insertion loss, reflection, or crosstalk that the link cannot absorb.

Flexible interconnection in three dimensions. A rigid board cannot solve every connection problem in a device with moving parts. Flexible circuits must carry signals between the gimbal, the cameras, the motors, and the sensors, and they must survive repeated motion. Rigid-flex construction allows a single assembly to span fixed and moving regions, reducing connector count and eliminating the wire routing labour that would otherwise be required. In a device with sixty-eight cables, the ability to replace part of that bundle with a flex circuit has direct consequences for weight, volume, and assembly reliability.

This is where the electronics industry’s approach is visibly changing. Distributed products like robot phones, AR glasses, robot joints, and automotive cabins are moving from planar board layouts into three-dimensional interconnection systems, and the design question becomes how to route signals through a structure that moves rather than across a surface that does not. That shift requires the flex materials, bend geometry, and coverlay placement to be treated as engineering decisions, as discussed in a capability review, rather than as mechanical packaging details.

The Hardest Part Is No Longer Making It Smaller

High-density design is well understood in consumer electronics. The difficulty in a robot terminal is that miniaturisation must coexist with system-level reliability across several domains.

Mechanical durability. Flexible circuits carrying signals to and from a moving gimbal accumulate bend cycles for the life of the device. Conductor fatigue, adhesive degradation, and coverlay stiffness all determine whether the connection survives. Copper foil selection matters here: rolled and annealed foil resists fatigue far better than electrodeposited foil, and trace orientation relative to the bend axis changes the strain a conductor experiences.

Thermal load. A high-performance processor, a charging system, a display, and now four motors concentrate heat in a small volume. Copper weight, thermal via arrays, and component placement determine whether heat spreads safely or creates local hot spots that shift the behaviour of nearby analogue sensing circuits.

Electrical coexistence. Motor switching noise and microvolt-level sensor signals occupy the same enclosure. Ground partitioning, loop area control, and physical separation between power and measurement circuits decide whether the control loop behaves predictably. Without deliberate design, the noise appears as a positioning error rather than as an obvious electrical fault.

Reliability under repeated use. A mechanically active phone is used more intensively than a static one. Connectors at the gimbal interface, solder joints under the main processor, and flex circuits crossing the moving boundary all accumulate wear. Failure modes that would take years to appear in a conventional handset can surface within a product cycle.

Those four requirements converge on the same manufacturing disciplines: process control on lamination, registration, drilling, and plating; controlled assembly covering paste volume, placement accuracy, and reflow; and inspection that reaches joints no camera can see. Paste inspection, optical inspection, X-ray, and electrical test each catch failure modes the others miss, and traceability connects the results to individual units so that any anomaly can be bounded. Bringing those stages together with the board fabrication process, rather than splitting them across suppliers, matters more than usual when a design combines rigid boards, flex, and rigid-flex in one product and iterates quickly. A structured test flow is what makes the difference between a prototype that demonstrates the concept and a product that can be shipped in volume.

What This Means for Product Teams

The robot terminal category points to several practical changes in how hardware should be planned.

Treat interconnection as an architectural decision. In a device with moving parts, deciding what runs on a rigid board, what runs on flex, and where rigid-flex transitions occur determines assembly complexity, weight, and reliability. Making that decision late, after the mechanical design is fixed, forces compromises that appear as yield loss rather than as design defects.

Confirm material and process capability before layout. A stackup that assumes a microvia size or line width the factory cannot hold in volume will pass first article and fail in production. Verifying achievable microvia dimensions, minimum line width, aspect ratio, and impedance tolerance converts assumptions into design inputs.

Model the flex cycle life explicitly. Bend radius, copper thickness, foil type, and trace orientation should be selected against an expected cycle count rather than inherited from a previous product. Where a design requires a tight radius, the mitigation choices have electrical consequences that must be resolved during layout.

Plan for noise before it appears. Adding actuation to a handset introduces switching noise into a system that also contains sensitive radio and sensor circuits. Grounding strategy, partitioning, and component placement should be designed deliberately, because correcting them after the mechanical design is frozen is expensive.

Keep assembly and test inside the same quality system. A device that combines dense rigid boards, flexible interconnects, and miniature components has failure modes concentrated at the interfaces between processes. Coordinating them under one supplier, under one quality system, removes the boundary where responsibility for a defect becomes ambiguous.

The broader significance of a phone with a gimbal and sixty-eight internal cables is that it marks a change in what a consumer device is. The innovation frontier is no longer purely computational; it is the fusion of computation, sensing, and actuation inside a single enclosure. That fusion does not reduce the importance of the printed circuit board. It makes the board the platform on which the fusion either works reliably or does not, and the engineering effort required to get it right now sits as much in interconnect, materials, and manufacturing consistency as in the silicon itself.

Frequently Asked Questions

Why does adding a gimbal to a phone complicate the PCB? It introduces actuation, which requires closed-loop control with low-noise sensing, adds moving connections that must survive repeated flexing, and consumes internal volume that would otherwise be available for the board.

How is space recovered in such a design? Through higher-order HDI and any-layer structures with laser microvias, fine-line traces in the densest regions, and flexible or rigid-flex interconnects that replace part of the internal cable bundle.

Why does flex copper foil choice matter? Rolled and annealed copper foil has a grain structure that resists fatigue cracking far better than electrodeposited foil, which matters for any connection that flexes repeatedly during the product’s life.

What electrical risks arise from combining motors and sensitive circuits? Switching noise from motor drive can couple into radio frequency and sensor paths. Ground partitioning, loop area control and physical separation between power and measurement circuits are the primary countermeasures.

What should be verified before production? That the required microvia and line width are routine for the chosen process, that flex cycle life has been modelled for the actual bend geometry, and that inspection and test cover hidden joints on the assembled device.