Industrial electronics manufacturing can involve several stages moving at the same time when a product is still under development.
While the main controller boards are still being debugged, the supply chain may already be preparing production materials. While the customer is validating the first samples, another part of the product may be going through a separate development process. And while the engineering team is still working on the product, the manufacturing side may already need to think about testing, cost and production volume.
This was the situation when a well-known Middle Eastern industrial cooling equipment company approached Fumax during the development of a new floor-standing cooling appliance with Wi-Fi and 4G connectivity.
The product would use a touch-control interface on the front panel and would eventually be assembled as a complete finished appliance.
Fumax initially joined the project for the manufacturing of three main electronic boards:
The timing was important.
The product was still being developed when Fumax became involved. There was no finished production package waiting to be handed over.
The first objective was therefore to help move the three boards through the prototype stage while keeping the future production requirements visible.
At the same time, another development track would soon begin for the touch-control interface.
The project eventually took roughly 10 months from the early prototype stage to mass production, with the different workstreams moving toward the same final target: enough electronics for 5,000 finished cooling products.
The first manufacturing work focused on the three controller and communication boards.
The customer and Fumax were working across a 6.5-hour time difference.
The customer also did not have a dedicated PCBA hardware laboratory for this type of prototype work, and obtaining small electronic components locally was not especially convenient.
This mattered during debugging.
A prototype engineer may need to try a different resistor value, capacitor, diode or IC and continue testing immediately.
But the customer’s internal purchasing process was not necessarily designed around that kind of small, repeated engineering request.
The customer later told us that even a simple request to purchase a few resistors could take four to five days.
For a product still in the debugging stage, that waiting time could easily become part of the engineering schedule.
Fumax therefore became involved in both sides of the prototype activity.
The customer could use online video to show the actual hardware behavior and test conditions. Fumax engineers could review the situation remotely while the required components were sourced through the local Chinese electronics supply chain.
Instead of treating component sourcing as a separate step after debugging, the two activities could move together.
A component could be identified during a debugging session, sourced locally, and then used for the next round of testing without waiting for a normal overseas purchasing cycle.
The customer later estimated that this approach shortened the overall debugging process by approximately three weeks.
The important part was not that Fumax was replacing the customer’s engineering team.
The customer remained responsible for the product design and development.
Fumax was providing a manufacturing-side environment where prototype hardware, component sourcing and physical verification could move in parallel.
There was another issue that appeared naturally during prototype development.
Debugging a new electronic product often involves small purchases.
One test may require a different resistor.
Another may require a replacement IC.
A third may require several capacitors or connectors.
Individually, these items are inexpensive.
But if every small debugging purchase has to be handled as a separate commercial transaction, the process can become unnecessarily slow.
A quotation needs to be prepared.
Someone needs to approve it.
A purchase request may need to be created.
Then the material is purchased.
During the prototype stage, neither side wanted to create that level of administration for every small experiment.
So Fumax discussed the issue directly with the customer’s department manager through a video meeting.
The two sides agreed on a US$500 Debug Credit.
The arrangement was simple.
When cumulative debugging-related expenses remained within $500, Fumax would record the purchases and settle them together later, with the corresponding details.
If the total went above $500, the additional expense would be discussed case by case.
This arrangement allowed the prototype work to continue without turning each resistor, capacitor or IC into a separate commercial discussion.
As the prototype work progressed, the customer ordered the first samples of the three main boards.
The initial sample quantity was:
10 pieces per board.
At the same time, Fumax had already prepared quotations for different production volumes:
10 units / 500 units / 5,000 units
That meant the project was being looked at from two directions at once.
The immediate question was whether the prototype and samples worked correctly.
But there was already a visible production target behind them.
The customer could see the cost structure of the future production quantities, including the MCU and other major components.
That became useful later when the project moved into the trial-order stage and the customer raised a cost-reduction requirement.
The three boards were therefore not being treated as isolated prototype builds.
They were the first part of a manufacturing program whose eventual production volume was already visible.
After the initial sample stage, the project moved into trial production.
The customer told Fumax that they were facing cost-down pressure as the project moved from prototype development toward electronics manufacturing.
Instead of treating the existing BOM cost as fixed, both sides went through the component list together.
Because the quotation had already been prepared at several production levels, the customer had visibility into the pricing of the main components, including the MCU.
That made the discussion more concrete.
The question became:
Which parts could realistically be changed, and which parts needed to stay?
Some of the opportunities came from the customer’s own sourcing channels.
The customer selected domestic alternatives for several categories, including:
The customer also decided to purchase the MCU directly and supply it to Fumax.
The final result was an approximately 30% reduction in BOM cost.
It was not a case of Fumax simply finding a cheaper part for every line.
Some alternatives came from the customer.
Some components remained as specified.
And some material became customer-supplied.
The important point was that the BOM could be reviewed as a shared manufacturing cost rather than treated as a fixed list that neither side could question.
The trial order itself was not the only thing being discussed.
From the prototype and sample work, the team had enough information to see that the basic product design was already relatively mature.
The trial order still had to validate the selected alternatives and the current configuration, but it was becoming possible to plan beyond it.
That was particularly relevant because several components in the product had longer or more complicated supply requirements.
The product used a Quectel 4G module, a special Panasonic capacitor, industrial-grade power ICs and plug-in adjustable resistors.
Some parts also had to meet requirements related to the product’s IP65 application.
For certain original components, lead time could reach approximately eight weeks.
At the same time, the project was being developed during the COVID-era supply chain disruptions, when component availability and lead times were much less predictable than under normal conditions.
So the production planning could not simply start after the trial order was finished.
Fumax discussed the possibility of looking at the annual production requirement earlier.
The idea was not to commit the customer to a production quantity before the product was ready.
It was to distinguish between different types of material planning.
PCB material and approved alternatives could be purchased according to the immediate trial-order requirement.
Long-lead-time items could be identified earlier.
And where the design was already considered mature, larger purchasing quantities could be evaluated for the annual production plan.
The customer agreed to move forward with that approach.
While the three main boards were moving through this process, another workstream was beginning on the same product.
The front-panel touch interface was about to become its own development track.
The customer knew what they wanted the finished product to look and feel like.
The front panel needed to provide a modern touch interface, but the customer’s team was less familiar with the underlying implementation of this type of appliance-style touch control.
There was also a practical concern.
The customer was used to European development methods and was concerned that applying the same approach to this relatively simple interface could result in an unnecessarily expensive solution.
Instead of starting with a detailed technical specification, Fumax asked the customer to show what they actually wanted.
The customer searched on Amazon for products with similar interfaces and sent the relevant product links and videos to the engineering team.
That gave Fumax something concrete to analyze.
After reviewing the interaction shown in the reference products, the engineering team determined that the required interface was actually much simpler than a full touchscreen architecture.
The display could use a 7-segment display, while the user input could be implemented with a capacitive touch-key structure.
The touch sensing layer could be built with a flexible circuit using printed conductive silver electrodes.
This was particularly suitable for the product’s front panel.
The decorative panel itself did not have physical buttons. The button positions were represented by graphics on the panel, while the silver-ink FPC was placed behind the corresponding areas.
When the user touched the marked area, the sensing electrode detected the change in capacitance and the touch-control circuit could identify the corresponding input.
The approach gave the customer the type of interface they wanted without requiring a much more complicated architecture.
The next step was not to leave the customer with a collection of separate components.
The display and touch circuitry were integrated into a single module.
From the customer’s side, this simplified the integration work because the physical touch layer did not have to be developed as a separate system alongside the display.
By integrating the display and touch circuitry into one module, the customer could interface with it as a single functional unit instead of developing the touch layer as a separate subsystem.
The customer provided a physical reference sample.
Rather than waiting to create every piece of the touch structure from the beginning, the customer asked Fumax to use reverse engineering to obtain an initial set of source files and accelerate the starting point.
The customer was also careful about the final product design.
The source data created from the reference sample was not intended to become the customer’s final layout.
The customer provided the pin definitions in advance and retained control over the final PCB layout and firmware.
Fumax could therefore work on the touch module in parallel while the customer continued its own layout and firmware development.
That separation was useful for the schedule.
Neither side had to finish the entire product before the other could start.
The touch-control development moved quickly because several activities were handled in parallel.
The reverse-engineered starting data was prepared.
The touch structure was developed.
The display supplier was coordinated.
And the tooling requirements were discussed early.
The screen tooling was confirmed before the rest of the project had completely finished.
After approximately four weeks, the first screen/touch-module samples were ready and sent to the customer.
The customer tested the interface and confirmed that the expected touch interaction worked.
The work then moved to the complete touch-control board.
Approximately two weeks later, the first complete touch-control board samples were ready for the customer.
At this point, the touch-control development had moved from:
reference product → initial source data → touch module → screen sample → complete control-board sample
while the main controller boards were still progressing through their own trial-order and production-preparation activities.
At the same time, the three main PCBA programs were continuing forward.
The component alternatives had to be validated.
The cost-reduction decisions had to be incorporated.
The long-lead-time components had to be planned.
And another question was becoming more important:
How would the functional test work once the production quantity reached 5,000 units?
The customer’s original test method was useful for engineering verification.
The available firmware was primarily intended for the finished product.
To check different operating conditions, the operator had to create different shorting combinations on the plug-in terminals.
Those combinations caused the MCU to execute different actions and operate different relay combinations.
For a small number of engineering samples, this was workable.
The same method became much less attractive when multiplied by thousands of boards.
The complete test cycle was taking approximately 3–4 minutes per board.
The issue was not that the test method was technically wrong.
It could verify the required functions.
The question was what happened when the same procedure had to be repeated at production scale.
With a planned quantity of 5,000 units, the test time accumulated quickly.
Fumax therefore reported the actual test duration to the customer and calculated the approximate production impact.
The customer quickly recognized the significance.
A procedure that was acceptable for prototype and engineering validation needed to be adapted before mass production.
The next step was therefore discussed before the 5K production run.
The customer would prepare a production-oriented firmware and test script.
Fumax would develop the physical test fixture needed to execute that process efficiently on the production line.
The customer worked on the production firmware and script.
Fumax worked on the customized fixture.
The objective was to remove the repeated manual shorting steps from the operator’s process and make the test sequence repeatable.
The fixture was designed so that the single-pole and double-pole controller boards could be loaded together.
The operator could place the two boards into the fixture and start the test from the computer.
The script then controlled the test sequence automatically.
The final setup could test:
two boards in approximately two minutes.
That was a very different production process from manually testing one board at a time through multiple terminal combinations.
More importantly, the test architecture was now aligned with the expected production volume rather than simply extending the original development procedure into mass production.
Fumax then prepared a pre-production sample using the locked production configuration.
The customer reviewed the finished function through video before the full production run.
At the same time, the customer was moving forward with UL certification.
Fumax went through the required factory audit and successfully passed the inspection.
The touch-control project had started during the trial-order stage of the single-pole and double-pole controller boards.
After the screen and touch-module samples were approved, the complete touch-control board also had to enter the same production schedule.
The customer therefore needed the electronics for the finished product, not just the controller boards.
The final production requirement included:
The communication electronics were also included in the production program according to the final configuration.
The three electronics assemblies therefore had to be ready for the customer’s finished-product assembly at roughly the same stage.
The final production test system was also in place.
The customer had provided the production-oriented firmware and test script.
Fumax had completed the customized fixture.
Instead of manually creating different shorting combinations at the terminals, the operator could place the single-pole and double-pole boards into the fixture together and start the test from the computer.
The production script then ran the required sequence automatically.
The completed cycle took approximately:
Two minutes for two boards.
The test process that had originally been used for engineering verification had now become a repeatable production station for the mass-production stage.
With the design data locked, production materials prepared, the touch-control board ready, the production test system in place and the UL audit completed, the project moved into mass production.
The customer received the required electronics for approximately 5,000 complete cooling products.
The project had taken roughly 10 months from the early prototype stage to mass production.
During that period, the project also crossed several Chinese holiday periods, including the National Day holiday, the May holiday and Chinese New Year.
The project was also carried out during the COVID-era supply-chain environment, when component availability and lead times were less predictable than normal.
Despite the different schedules involved, the separate workstreams had reached the same production target.
The customer received:
5,000 single-pole controller boards
5,000 double-pole controller boards
5,000 touch-control boards
The electronics required for the complete cooling product were therefore in place for the customer’s final assembly.
The project started with three controller and communication boards still in development.
Approximately ten months later, the corresponding electronics production program was ready for 5,000 finished products.
