Blog
Estudio de caso: Prototipo Mecánico-Eléctrico de Monitor de Gas GLP
An IoT gas monitor is often seen as an electronics-led product. It usually involves sensors, a wireless module, a battery system, firmware, and data communication. From the outside, the enclosure may look like a secondary part of the design.
This project showed a different side of the problem.
When PS Electronics received the inquiry, we realized that the project was not only about building a circuit board or adding an enclosure after the electronics were defined. The device had to carry the cylinder’s weight, protect measurement accuracy, support a battery-powered wireless system, and remain realistic at prototype quantities.
This case study looks at how PS Electronics mapped out a product realization path from RFQ to buildable prototype, while keeping the next production stage in mind.
Antecedentes del proyecto
The client was an early-stage hardware team targeting household and small commercial LPG users in West Africa. They had a concept, a component direction, and a clear request covering electronics, firmware, structure, prototyping, and manufacturing support.
| Artículo | Descripción |
|---|---|
| Producto | Smart LPG cylinder gas-level monitor |
| First build | 3–10 prototypes |
| Next quote band | 100–1000 units |
| Core stack | Load cell + HX711 + ESP32/C3/S3 |
| Main risk | Load-bearing enclosure affecting measurement accuracy |
| Needed outcome | Prototype route now, production-ready path later |
What they did not yet have was a mechanical-electrical architecture that could turn the concept into a buildable device.
Core Challenge
In this product, the enclosure is part of the measuring system. If the sensor seat distorts, or if cable routing and housing geometry add side load, the reading can drift even when the PCB is working correctly.
A second constraint was the business stage. The first build was only 3–10 prototype units, with an explicit request for 100–1000 unit pricing next. This type of project can easily fall into an uncomfortable middle ground. The quantity is too small for a mold-first approach, but the requirement is too serious for a simple sample build. Many suppliers may see it as too early-stage to prioritize, but for the customer, this first build would decide whether the product could move forward.
The first task for PS Electronics was therefore not to quote the PCB, enclosure, and tooling as separate items. It was to define a practical product route.
The shortest way to describe the problem is this:
| Constraint | Engineering meaning |
|---|---|
| Load-cell measurement | No side load, bending, or torsion should enter the sensing path |
| Cylinder support | The enclosure had to carry real mechanical load, not only protect electronics |
| Battery-powered IoT | Battery volume, connector direction, antenna clearance, and cable routing became enclosure constraints |
| 3–10 prototypes | The first build had to validate architecture, not just appearance |
| 100–1000 units next | The design had to keep a production-ready path from the beginning |
How PS Electronics Structured the Project
PS Electronics treated the requirement as a system-integration problem led by mechanical definition. Before moving into fabrication, three things had to be locked first.
- Separate the Paths
The structure had to be divided into three paths: the load path, the sensing path, and the service path.
- The load path had to support the LPG cylinder.
- The sensing path had to protect the load cell from side load, bending, and torsion.
- The service path had to allow battery access, cable routing, assembly, and maintenance without disturbing measurement stability.
If these paths merged too early, the prototype could look complete but still produce unstable or misleading weight data.
- Use a Staged Prototype Route
At 3–10 units, the target was architecture validation, not production simulation.
CNC parts, printed fixtures, bent supports, or hybrid builds could be more useful than forcing a mold-first decision too early. The first build had to test the structure, load transfer, electronics envelope, and assembly logic before the design moved toward the next quantity level.
- Freeze the Electronics Envelope Early
Battery size, PCB position, connector direction, antenna keep-out, and sensor cable routing all affected the enclosure.
If these rules were not defined early, the housing could require late-stage redesign. For this reason, PS Electronics treated the electronics envelope and enclosure structure as one connected design problem from the beginning.
Engineering Logic Behind the Decisions
The key decisions were not arbitrary. They followed standard design rules already used in real load-cell, RF, and molded-part programs.
| Engineering basis | Practical impact here |
|---|---|
| Load path discipline | The enclosure had to keep side force, bending, and torsion away from the sensing path |
| Off-center loading | Real gas cylinders may not land perfectly centered every time |
| RF clearance | The antenna zone and enclosure geometry had to be considered together |
| Prototype-to-mold path | Wall thickness, ribs, draft, and radii had to leave room for later tooling and cost-down |
For this project, these rules all pointed to the same conclusion: the enclosure could not be treated as a final shell. It had to carry load, protect the sensor, support the electronics, and remain manufacturable.
Quality Control Had to Follow the Failure Path
For this type of product, quality control had to follow the failure path, not the org chart.
| Gate | What gets checked first | Why it matters |
|---|---|---|
| IQC | Sensor-mount dimensions, support parts, fasteners, sealing materials | Structural tolerance can become weighing error |
| Control de calidad en proceso | Load-path preservation, cable routing, enclosure fit, assembly order | Poor assembly can add mechanical bias into the reading |
| Cualquier cosa | Closure repeatability, post-assembly stability, RF behavior, loaded response | The assembled unit has to perform as a complete system |
Three additional checks were especially important:
- Incoming parts must preserve the sensor seat and support datum.
- Assembly cannot force cables, boards, or battery holders into stress-bearing positions.
- Final validation has to happen in the closed housing, under representative loading, not only on bench electronics.
What Changed for the Customer
The project moved from a broad hardware inquiry into a clearer product realization path.
| Before | After |
|---|---|
| Broad hardware inquiry | Clear prototype-to-production route |
| “Need design + prototype + quote” | Clear staged NPI logic |
| Risk of fragmented vendors | Integrated enclosure, electronics, and manufacturing coordination |
| Prototype-only thinking | Prototype route connected to 100–1000 unit production planning |
| Unclear first-build method | Practical options such as CNC, hybrid build, or mold-led planning could be evaluated based on project stage |
That shift matters. The customer no longer has to guess whether the first build should be CNC, hybrid, or mold-led. They no longer have to discover late that the housing is distorting the sensor path.
In this case, the founder stated the requirement directly: “Advise on manufacturing feasibility, provide a quote for producing 3–10 prototypes, and provide optional pricing for 100–1000 units.”
Consideraciones finales
For early-stage hardware products, the first prototype should not only answer “Can it be made?” It should also answer “Can this product move forward without being redesigned from the beginning?”
That was the key lesson in this project. For a weight-based IoT gas monitor, the enclosure was not just a housing, and the PCB was not the only engineering problem. The product had to be understood as a complete system from the start.
In other words, the starting point had to be designed with the endpoint in mind.
Preguntas frecuentes
Cuando el BGA principal, la memoria o la interfaz de alta densidad no se pueden enrutar limpiamente con orificios pasantes convencionales. Si el enrutamiento de escape comienza a forzar capas adicionales, un tamaño de placa más grande o una geometría de traza arriesgada, se debe revisar HDI desde el principio.
La prueba piloto confirmó si toda la cadena de fabricación podía soportar el diseño, no solo si se podía fabricar una muestra. Le dio al cliente datos reales de rendimiento y entrega antes de comprometerse con la producción mensual.
Andy es un profesional experimentado en la industria de PCBs con décadas de experiencia en fabricación, ensamblaje y soporte al cliente de PCBs. En PCBCool, lidera el equipo de marketing y ayuda a convertir la experiencia práctica de proyectos en contenido técnico útil para ingenieros, compradores y desarrolladores de productos.
