Hirose
Webinar Date: December 10, 2025
Key Takeaways
- Real-world signal and power routing challenges in humanoid robotics
- Connector strategies for motion, weight reduction, and durability under stress
- Approaches to high-speed data in vision, sensing, and AI processing
- How Hirose solutions enable miniaturization, modularity, and system integration
- Proven connector use-cases across humanoids, prosthetics, surgical systems, and assisted mobility
Frequently Asked Questions
As form factors keep shrinking, what typically becomes the first constraint for connectors—current, shielding, or mechanical retention, or any other?
- As products keep getting smaller and smaller, it involves less material which causes mechanical restraints in terms of retention between the socket and receptacle. If you are pulling on those wires, it’s a lot more difficult to keep those mated halves together. While the products continue to shrink, the distance between the pins are also reduced. Lower voltage and current ratings are also a factor. Innovating the proper locking mechanisms will account for the retention between the connectors. Driving the correct amount of voltage and current is something that is also being innovated.
What are some key features or characteristics for a connector to be used in humanoid robots?
- Combining multiple connectors into one solution reduces the size of the robot. This also minimizes the amount of wiring for multiple connections.
Can micro coax cables be terminated in the lab? Or do you need complex/expensive equipment to do so?
- There are hand tools to accommodate for the termination. There are also small prototype sample quantities that can be ordered for experimenting.
For humanoid head sensor clusters, when does micro-coax outperform flex-based differential pairs for MIPI or LVDS routing in terms of noise immunity and mechanical fatigue?
- Most engineers would likely pick micro-coax for both EMI noise and mechanical fatigue, though the fatigue may be not as significant advantage over flex, at least in the one-dimension flex cable can bend.
How do you see connector geometry evolving to handle increased noise coupling from multi-sensor fusion systems—especially when combining vision, IMU, and ToF sensors in tightly constrained humanoid heads?
- It’s early still, so far, we haven’t had to rethink connector fundamentals to meet our existing customer needs, if you have an application specific target connector spec., we’d love to consider it.
Do you offer native Altium footprints and 3D step files for the micro-coax, IX Industrial, FX30B, and hybrid power-signal connector families you covered today? And if so, where’s the preferred source—your website, SnapEDA, or an official library?
- We can support Altium or Allegro footprints, please request on our website or with your local Hirose Sales member.
You mentioned SFC up to 10 feet — how do you manage impedance and crosstalk over long flex harnesses routed through joints and torsion loops?
- For MIPI example or any 5-9 Gbps, with properly connectorized and shielded and GND line selection, we see both FPC and micro-coax satisfying several feet length runs, though 10 feet was mentioned as a “challenge” which it would be.
Is micro-coax still preferred over shielded FPC for internal AI/vision links inside moving robotic systems?
- I think most engineers prefer micro-coax due to its inherent 360 degree shielding per line. However, we see some customers choosing FPC cabling for costs, weight, and possibly other design criteria.
How does Hirose handle transitions from PCB microstrip to micro-coax in a way that preserves impedance and minimizes discontinuities?
- Check out our SMT board mount receptacles such as U.FL-R series, these provide a reliable transition from cable to microstrip.
What are the relative cost comparisons for cabling schemes?
- Ganged micro coax bundles more costly, multi-layer shielded FPC cables somewhat less costly, discrete wire and FFC flat cables least costly
What is the battery energy capacity requirement of the humanoid robots? At single charge, how many hours do we expect to operate? Assumption is just a normal and reasonable workload and operation. Hirose doesn’t influence battery selection or capacity requirements
- Hirose doesn’t influence battery selection or capacity requirements.
What AI processor can be used for humanoid robot?
- Hirose doesn’t directly influence processor choice.
What is the series of the hybrid bayonet push connect circular connectors?
- Our LF series is a bayonet lock with hybrid variations.
As a student, what are some skills/certifications we can get/work on to be competitive in this field of humanoid robots.
- Right now, majority of companies are exploring either the mechanical limitation of various hand grippers or the software that controls the thought process of real-time reactions. While we at Hirose do not offer software solutions, we offer solutions to make sure your signals can get from point A to point B even in environments with extreme limitations.
How do you manage thermal rise when running 3–10 A through such small SMT FPC connectors inside compact hand modules with limited airflow?
- In general, temperature rise is determined by the current run through the conductive contacts. While increasing contact area typically increases the current rating, we also need to make sure that the plastic can hold up to heat in its environment and that there is enough thermal dissipation through air exposure. We at Hirose are proud that we test thermal rise conditions by powering all pins until a 30-degree Celsius temperature rise is detected in an environment with still air.
How does the bayonet-lock variant of IX/HR30 handle torsional stress and shock loads inside humanoid wrists or elbows compared to friction-lock or latch-lock connectors?
- Latching locks are the most popular for internal connection points. Bayonet locking is less popular internally often due to an increase in price and the lesser need for IO connections, but can offer significant retention forces. Friction locks are not recommended for areas of expected shock because it can lead to partial or full disconnects.
3–5 A continuous on one power rail, with peaks up to 8–10 A at ~10–20% duty cycle?
- Assuming worst case scenario, 5A continuous + 10A @ 20% duty cycle would average out to about 7A. Temperature rise is not linear, so let’s make the assumption it causes a temperature rise equal to about 8A. Hirose’s current rating is based upon all pins being powered at the rated current, so finding a small SMT FPC connector rated to 10A or 15A would likely be a solution.
Are you exploring hybrid harnesses that combine micro-coax for vision/IMU with power conductors in the same physical cable, to reduce mechanical stiffness in humanoid torsos?
- Hirose is looking into a connector that can handle coax signals and power. While it may not except a wire with a single conductor that can do both, it can certainly accept a cable harness with coaxial wires and discrete power wires.
When combining SMP-style differential pairs with integrated power rails inside a single miniature connector, what design practices ensure that high-current rails (5–20A) don’t degrade signal integrity for multi-Gbps sensor lanes under vibration or bending?
- This is a very real and difficult problem in the interconnect industry. Added grounding pins or thermal isolation within the design of the connector can partially solve this issue. Ultimately, it comes down to the unique needs that the connector is designed for and what trade-offs are allowed in each design.
Do you recommend integrating sense lines or local regulation on each 1.5 A branch when using this kind of 1-to-2 adapter in humanoid robots, to compensate for voltage drop along long harness runs?
- Branch adapters, while advertised for an equal split, do not regulate voltage drops or current draws. The current is split based on the resistive load, so we highly encourage users to add their own regulatory systems.
When mixing DF-51 signal lines with DF-32 high-current rails inside the same humanoid joint or harness, what spacing or isolation practices do you recommend preventing EMI injection from 30–40 A loads into sensitive sensor lines?
- Isolating the power by securing it to a wall or other boundry can help, but this may also resitrct mobility. The signals could be further shielded using a larger cable insulator or by moving to a twisted differential pair. Finding the best solution is often solved on a case-by-case basis.
Referral Webinar Link
If you would like more information from this webinar, please review the details below.