When an industrial OEM client approached us with a temperature transmitter design, they had already completed the schematic, selected the components, and built a handful of prototypes that worked beautifully in the lab. The design used an RTD sensor with a 24-bit delta-sigma ADC for high-precision 4-20mA loop-powered output. Everything checked out.
Then came the production order: 3,000 units, six weeks.
What happened next—and how we navigated the challenges together—offers a practical look at what it really takes to move a sensor-based industrial product from prototype to reliable volume production. This case study reflects our approach to temperature transmitter manufacturing and how we help clients bridge the gap between a working prototype and a manufacturable product.
The client’s engineering team had done their homework. The schematic followed a leading semiconductor manufacturer’s reference design for a 2-wire RTD transmitter. The ADC selection was appropriate. The BOM was complete. From a functional standpoint, the design was sound.
But when we received the Gerber files and started our pre-production review for temperature transmitter manufacturing, we spotted several issues that weren’t on their radar—not because the design was bad, but because the client’s team, like most hardware teams, didn’t have manufacturing experience.
Sensor Placement and Thermal Coupling
The RTD sensor was placed near the edge of the board, close to a voltage regulator. In the lab, this hadn’t caused noticeable issues—the prototypes were hand-assembled and tested under controlled conditions. But in production, where thousands of boards would go through reflow together, the thermal gradient between the sensor and the regulator would create inconsistent measurements across batches.
The sensor needed to be placed away from heat sources and on a dedicated thermal pad to minimize measurement offset. This wasn’t a design flaw—it was a design choice that worked for prototypes but wouldn’t scale reliably in temperature transmitter manufacturing.
Test Point Accessibility
The client had included test points on the board, but some critical signals—the ADC reference voltage and the sensor excitation path—were positioned in ways that would have made fixture design unnecessarily complex. We worked with them to relocate a few test points to the bottom side of the board, aligning them for a simpler, more reliable bed-of-nails fixture. A small adjustment that made the test process far more robust.
Component Spacing and SMT Compatibility
Some passive components were packed tightly around the ADC, leaving insufficient clearance for the pick-and-place nozzle. A human hand can place components in tight spots; a machine needs room to work. This would have caused placement errors and lower yield during SMT assembly.
We follow strict DFM guidelines to catch these issues early, and our SMT assembly process is designed to handle complex component placements with precision.
We didn’t reject the design and send the client back to the drawing board. That’s not how manufacturing partnerships work. Instead, we walked through each issue, explained why it mattered in production, and proposed practical solutions.
We recommended moving the RTD sensor to the opposite side of the board, away from the voltage regulator, and adding a thermal pad to improve heat sinking. The client’s team validated the change with a quick prototype spin, confirmed that accuracy remained within spec, and approved the update.
We worked with the client to identify the minimum set of signals that needed to be accessible for functional testing. We added test points for the ADC reference, the sensor excitation path, and the 4-20mA output loop. These were small pads placed on the bottom side of the board, positioned to align with a bed-of-nails test fixture.
We adjusted the placement of several passive components around the ADC to meet our SMT equipment’s clearance requirements. The changes were minor—0.2mm here, 0.3mm there—but they made the difference between a board that could be placed reliably and one that would generate placement errors.
For projects with tight timelines, our prototype to production support ensures these adjustments happen quickly without derailing the schedule. Our NPI and pilot production process is built to handle exactly this kind of iterative refinement.
With the design updated, we turned our attention to production testing.
The client had provided a detailed test protocol—step-by-step instructions for verifying the transmitter’s accuracy across its temperature range. But their protocol assumed a lab environment with a technician holding probes and logging results manually.
For temperature transmitter manufacturing, we needed something different: a fixture that could load the board, make contact with all test points simultaneously, run the test sequence automatically, and log pass/fail results in under 30 seconds per unit.
We designed and built a custom test fixture based on the client’s protocol. The fixture used spring-loaded pogo pins to contact the test points we’d added, connected to a multimeter and a current loop simulator. The operator placed the board in the fixture, closed the lid, and pressed a button. The fixture ran the test sequence, compared results against the client’s acceptance criteria, and displayed a pass/fail indicator.
The fixture also logged every test result with a unique serial number, giving the client full traceability for every unit shipped.
This is a core part of our box build assembly service, where we handle not just board assembly but full product integration and testing.
The client’s original calibration plan was straightforward but labor-intensive: each unit would be manually calibrated using a reference temperature source and a multimeter, with a technician adjusting the output trim until it fell within spec. At three minutes per unit, 3,000 units would have consumed 150 hours of technician time—two full weeks of dedicated labor, assuming no interruptions.
During our pre-production review, we flagged this to the client. We walked through the math: three minutes per unit, 3,000 units, plus setup and rework time. The risk wasn’t just the labor cost—it was the schedule risk. If any calibration step proved problematic, the entire delivery timeline would slip.
We proposed integrating the calibration process into the existing functional test cycle. Rather than a separate manual step, the calibration could run automatically as part of the 30-second test fixture sequence: the fixture would apply reference values, the board’s firmware would read the ADC output, calculate the calibration coefficients, and store them in EEPROM—all without operator intervention.
The client’s engineering team took ownership of the software side. They modified the board’s firmware to accept calibration commands over the test interface, added the calibration calculation routine, and implemented EEPROM storage for the coefficients.
On our side, we added a reference resistor bank to the test fixture to simulate RTD sensor values at known temperatures, and we integrated the calibration sequence into the existing test automation software. The fixture now applies the reference values, triggers the board’s calibration routine, and verifies that the stored coefficients produce the correct output—all within the same 30-second cycle.
Calibration time dropped from three minutes per unit to zero additional time—it was now embedded in the functional test. The client avoided two weeks of labor, and the automation eliminated the risk of human error in the calibration process.
For smaller production runs, we also offer low-volume PCB assembly with no minimum order quantity, making this kind of automated testing accessible even for pilot runs.
1. Prototypes lie.
A design that works perfectly on the bench often fails at volume—not because it’s wrong, but because production is a different environment. Testability, thermal management, and SMT compatibility matter just as much as circuit performance in temperature transmitter manufacturing.
2. Test fixtures are not optional.
If you can’t test a board quickly and reliably at volume, you don’t have a production-ready design. Plan for test access from the start, and involve your manufacturing partner early in the test fixture design process.
3. Calibration is a production problem, not just a design problem.
Manual calibration doesn’t scale. If your product requires calibration, design the production process around it—automate what you can, and simplify what you can’t.
4. The right manufacturing partner doesn’t just build boards; they help you build better products.
We didn’t just assemble this client’s boards. We helped them see their design through the lens of manufacturing—and together, we turned a good design into a great product.
Learn more about our PCB assembly services and turnkey PCB assembly for industrial sensor products.
Ready to move your temperature transmitter design into production? We offer free design-for-manufacturing reviews for new projects. Send us your files, and we’ll flag potential production risks before they become expensive problems.
Q: How do you ensure measurement accuracy in production?
A: We develop custom test jigs based on your test protocol, simulating real operating conditions with calibrated reference standards. Each board undergoes 100% functional test with serialized logging.
Q: Do you support mixed production of different board designs?
A: Yes. Our flexible SMT lines support high-mix, low-to-mid volume production with standardized changeover procedures.
