This IoT device manufacturing project involved a small wearable device called TAG.
The customer called this product TAG.
It is a small wearable IoT device designed to be worn on the arm. The TAG uses UWB and LoRa technologies for indoor positioning and is intended for use in large warehouse environments, where it can help reduce the risk of collisions between workers and forklifts, particularly around blind corners.
The overall size of the product is approximately 10 × 6.5 × 1.3 cm.
The 13 mm thickness was an important product requirement for the customer.
The reason was mainly appearance. The customer wanted the TAG to look as thin as possible because a thinner product was considered more modern and more futuristic.
For the customer’s engineering team, however, the 13 mm target became a constraint that affected almost everything inside the enclosure.
The product needed to accommodate the PCBA, battery, E-ink display, buttons, connectors, DWM1001C UWB module, and LoRa module within a relatively small space.
As the design progressed, several details that looked reasonable when considered separately became more complicated when the actual parts were assembled together.
We saw several of these issues during sample assembly and manufacturing preparation.
The TAG uses an SMD tactile switch.
The product itself was required to meet IP65, including the button area. At the same time, the customer wanted the button to have a clear tactile feel and wanted to reduce the direct mechanical load applied to the SMD switch itself.
This was also an important concern internally because the customer had received complaints about button performance on earlier products. After long-term use, buttons could lose their original feel, become less responsive, or eventually fail.
The customer therefore had several requirements to deal with at the same time: waterproofing, tactile feedback, durability, cost, and limited internal space.
The customer did not want the SMD switch itself to be exposed directly through an opening in the housing.
Instead, the upper housing was designed with several U-shaped slots in the button area. A PVC membrane was then applied to the surface, with raised sections corresponding to the button positions.
This arrangement allowed the user to feel a defined button area while keeping the switch itself behind the housing and membrane rather than having the switch directly exposed to the user’s finger.
The first problem appeared when the housing mold was completed and samples were assembled.
Because of the U-shaped slots, the button sections were no longer as rigidly connected to the main body of the housing. These areas could move slightly up and down.
The clearance between the housing and the approximately 1 mm-high SMD tactile switch was already very limited because of the overall thickness target.
In some cases, the button area could therefore move downward slightly and touch the switch underneath. This could result in an unintended actuation.
The initial idea was to use a lower-profile SMD tactile switch.
We looked for alternatives in the range of approximately 0.6–0.8 mm in height.
The lower switches solved part of the clearance problem, but the test results created another issue.
With a thinner SMD tactile switch, the internal structure available for the spring mechanism is more limited. As the switch became shorter, the force feedback also became weaker.
This meant that reducing the switch height solved the mechanical clearance issue but made the button feel less satisfactory.
The final adjustment came from the mold supplier.
The plastic thickness of the upper housing was increased by approximately 0.2 mm to improve the rigidity of the button area and reduce the movement created by the U-shaped slots.
At the same time, a shallow area of approximately 0.1 mm was machined by CNC on the inside of the upper housing at the position corresponding to the SMD switch. This created additional clearance without requiring a lower switch.
The resulting structure allowed the original button height and tactile characteristics to remain while reducing the possibility of unintended actuation.
There was a manufacturing cost associated with this adjustment.
Because of the additional CNC operation, the housing cost increased from approximately USD 4.5 to USD 5.1 per set.
It was a relatively small mechanical change, but it became a direct manufacturing cost difference.
The E-ink display created another issue.
In the original mechanical arrangement, the PCBA and battery were positioned side by side rather than stacked vertically. The E-ink display was positioned above the battery.
The upper housing included a small step around the display opening. The display could sit neatly in this area, and the battery underneath provided additional support.
When the customer first checked the structure, the arrangement worked well.
There was one important difference between that early validation and the later production version, however.
The customer had originally tested the structure using a lithium battery pack purchased locally in their home country.
Later, after the project came to China, the customer asked us to help source several battery options. We evaluated approximately four to six battery samples before one version was selected for production.
The final battery met the requirements that had been defined for the project, but its thickness was not the same as the battery used during the customer’s original structural validation.
The final battery was thinner.
That difference was not particularly significant when looking at the battery by itself. It became significant when the complete product was assembled.
During assembly, we found that the E-ink display was no longer properly aligned with the opening in the upper housing.
The display could shift slightly because the thinner battery no longer provided the same support as the original battery used during the initial validation.
From the front, the display no longer sat evenly within the housing opening, with the alignment between the display and the surrounding bezel visibly off.
The first battery was tall enough to help support the display from underneath. After the thinner battery was selected, that support was reduced.
The original housing structure had therefore relied on a dimensional relationship that was no longer present in the final assembly.
The solution was not to redesign the entire housing.
Instead, the display fixing method was changed.
The PVC membrane supplier added adhesive backing at the display area so that the E-ink display could be fixed directly to the membrane and upper housing.
The assembly sequence was also changed.
Normally, the display would be connected to the PCBA first, and the complete PCBA with the display attached would then be installed into the housing.
For this product, we used a different sequence.
The display was first fixed separately to the upper housing with the PVC membrane. The display was then connected to the PCBA during the following assembly step.
For additional fixation, LOCTITE 431 was also used at the interface between the display and the housing.
A highly fluid instant adhesive was not suitable for this location because it could potentially flow into the E-ink display and damage it. A suitable low-flow adhesive was therefore used instead.
The final arrangement provided the required display fixation without changing the battery that had already been selected for production.
The screws used to secure the PCB to the lower housing created a different type of issue.
The customer originally selected pointed self-tapping screws.
Because the screws needed to form their own thread in the plastic housing, a conventional machine screw could not simply be substituted.
The difficulty was the thickness of the bottom housing.
Because the product was designed to remain thin, the bottom housing was also relatively thin.
During assembly, even the shortest suitable screw could place significant localized pressure on the plastic when tightened.
White marks appeared around some of the mounting points.
The screw itself was not necessarily incorrect. The problem was the combination of the available screw length, the thin housing, and the amount of tightening force applied during assembly.
There was no shorter version of the same type that could simply replace it.
The final solution was to increase the bottom housing thickness by approximately 0.2 mm.
We also changed the assembly control for the screws.
A torque-controlled screwdriver was introduced, and a specific tightening torque was added to the assembly SOP.
The purpose was not to tighten the screw as much as possible.
The torque needed to remain within a defined range: high enough to secure the PCB properly, but low enough to avoid excessive pressure from the screw on the thin plastic housing.
For many conventional PCB-to-plastic assemblies, this level of torque control would not normally be necessary.
In this case, however, the available space was limited and the housing was thin. The assembly process therefore needed to account for the mechanical effect of the screw as well as simply making sure that the PCB was securely fixed.
The connectors（JST ACHL connector) created another issue related to the limited internal space.
The customer wanted the connectors to be as flat and compact as possible.
This made sense from a mechanical layout perspective. Smaller connectors occupied less space and made it easier to keep the overall product within the 13 mm thickness target.
However, smaller connectors also meant smaller pins and smaller PCB pads.
That reduced the amount of PCB area available to mechanically retain the connector.
The issue became more obvious because the battery connector had another requirement.
After the battery’s male connector was inserted into the female connector on the PCB, the customer wanted additional adhesive fixation to improve the reliability of the connection.
The product then also went through ultrasonic welding during the later assembly process.
The small connector and small soldering area were therefore exposed to mechanical forces beyond normal insertion and removal.
During actual assembly, we found that the female connector could be pulled away from the PCB during the ultrasonic welding process.
The final adjustment was to increase the PCB pad area around the connector.
Red glue was also added to provide additional mechanical fixation.
This allowed the compact connector to remain in place while improving its ability to withstand the subsequent assembly process.
Again, the connector itself had not simply become a bad component.
The issue came from the relationship between connector size, PCB pad size, the mechanical structure, and the process that came later.
The original 13 mm requirement was a product-level target.
The customer wanted the TAG to be thin because of how the product should look and feel in the market.
By the time the product reached actual assembly, however, that single number had become many smaller requirements.
How much clearance was available around the SMD tactile switch?
How much movement could the U-shaped housing section tolerate?
How thick would the final battery actually be?
What would support the display from underneath?
How much space was available for the screw tip?
What tightening torque could the thin housing tolerate?
How small could the connector be while still leaving enough PCB pad area?
None of these questions existed entirely on their own.
The button supplier understood the switch.
The battery supplier understood the battery dimensions.
The tooling supplier understood the housing.
The connector supplier understood the connector.
The PCB manufacturer understood the PCB manufacturing requirements.
The customer’s engineering team had the broader product requirements.
The actual product was where all of these individual decisions eventually met.
That was where some of the issues became visible.
The battery itself was acceptable, but its different thickness affected the support relationship with the display.
The connector itself was acceptable, but the smaller format reduced the available PCB pad area and created another consideration during ultrasonic welding.
The self-tapping screw was also a normal component, but its interaction with a thin plastic housing meant that tightening torque needed to be controlled.
The button design addressed waterproofing and tactile requirements, but the flexibility introduced by the U-shaped slots created another interaction with the SMD switch underneath.
These were not necessarily cases where one person or one supplier simply made a mistake.
They were examples of different parts of the product interacting with one another.
Problems like these are not unusual in electronics manufacturing.
They can happen in almost any project, including projects that appear relatively straightforward. They are also not necessarily the result of a mistake by the engineer, the component supplier, the tooling supplier, or the EMS provider.
There are simply many details in a product, and not every interaction between components, materials, tooling, and assembly processes can be identified perfectly at the beginning.
This is one of the reasons why even relatively simple products still need prototypes, samples, and pilot production before volume manufacturing.
These stages provide opportunities to find issues that are difficult to see from drawings or individual component specifications.
For this TAG, several of the issues only became visible when the actual housing, battery, display, PCB, buttons, connectors, and screws were assembled together.
From the overall project perspective, the existence of these problems was not unusual.
What mattered more was when they were found, how many of them were found before volume production, and how quickly everyone involved could discuss them.
Fumax has been involved in a large number of NPI projects, and this is something we have seen many times.
For us, the role of manufacturing is not to assume that every problem can be prevented before the first physical sample exists.
It is to make sure that when something appears during prototyping, sampling, assembly, or pilot production, the customer knows about it in time.
We discuss what we see, explain the manufacturing impact as clearly as possible, and work with the customer and relevant suppliers to determine the next step.
The earlier a problem is found, the more options the project usually has to deal with it.
For an EMS partner, being transparent and timely about these issues is part of the job.
