When a medical device moves from schematic to shippable product, the gap between a working prototype and a reliable, repeatable assembly is where most projects stall. We see it weekly: a design that performs flawlessly on the bench but fails in pilot production due to a component tolerance stack-up, a solder mask issue, or a test strategy that only catches 60% of latent defects. This article is our working checklist for de-risking medical device PCBA, written from the perspective of the engineers who review your files, challenge your BOM, and sit on the phone with you during the first article inspection.
We are not going to give you a generic list of "quality is important" bullet points. Instead, we will walk through the specific engineering decisions we make alongside our clients, from the first RFQ to the final functional test. Our goal is to give you a concrete framework to bring to your next design review, so you can ask the right questions before you commit to tooling or long-lead components.
Start with the RFQ: What We Need to See Beyond the Gerbers
The request for quote is not a formality. It is the first risk assessment. When we receive a medical device PCBA package, we are not only looking for the board outline and the netlist. We are looking for three specific documents that tell us how much engineering effort this project will require.
Technical reference: For related engineering context, see IPC standards and industry resources.
First, we need the fabrication drawing with explicit impedance requirements, surface finish specifications, and any blind or buried via stack-up details. Second, we need the assembly drawing with polarity markers, mechanical constraints, and a clear indication of which components are critical for placement accuracy. Third, and most often missing, we need the test strategy. Do you have a bed-of-nails fixture? Are you planning for flying probe? Do you have a firmware engineer available to write a boundary scan test? The answer to these questions changes how we plan the panelization and where we place fiducials.
Here is the operating question we ask every client at this stage: What is the failure mode you are most afraid of, and how will your test plan catch it? If the answer is "we haven't thought about it yet," we know we need to build in more design for test (DFT) review time. If the answer is a specific electrical parameter, like "we need to verify the output voltage ripple under load," then we can start planning the test hardware.
DFM: Where We Find the Hidden Costs
Design for manufacturability is not about making the board easier for us to build. It is about making the board more predictable to build. In medical electronics, predictability is the entire game. A 0.5% yield loss in a consumer gadget is annoying. A 0.5% yield loss in a glucose monitor is a recall.
Our DFM review focuses on three areas that cause the most rework in medical assemblies: thermal relief on ground planes, via-in-pad plugging, and component spacing for rework. Let us be specific. If you have a dense BGA on a board with thick copper planes, we will ask about the via-in-pad fill material. If you are using a no-clean flux that leaves a residue that interferes with an electrochemical sensor, we will recommend a different solder paste or a wash step. These are not "nice to have" suggestions. They are the difference between a stable process and a field failure.
We also look at component orientation. We have seen medical designs where a polarized capacitor is placed in a way that the silkscreen indicator is hidden under a mechanical bracket after assembly. That is a manual inspection error waiting to happen. We will flag it and ask you to rotate the part or move the bracket. It takes you ten minutes in your CAD tool. It saves us an hour of rework and you a potential field return.
Here is a practical table we use internally to triage DFM findings. It helps us communicate urgency without being alarmist.
| Severity | Definition | Example in Medical PCBA | Recommended Action |
|---|---|---|---|
| Critical | Will cause functional failure or safety hazard. | Missing solder mask between closely spaced high-voltage pins. | Stop production. Redesign or manual rework with engineering sign-off. |
| Major | Will likely cause yield loss or field reliability issue. | Unsupported through-hole connector on a board subject to vibration. | Add mounting holes or specify a different connector. |
| Minor | Will not fail but increases inspection time or cost. | Silkscreen text overlapping a test pad. | Adjust silkscreen on next rev. Acceptable for current build. |
We do not expect you to agree with every finding. We do expect you to respond to each one with a decision: accept, modify, or provide a rationale. That conversation is where the engineering happens.
BOM Engineering: The Long-Lead Item That Kills Schedules
The bill of materials is the most political document in electronics manufacturing. It is where procurement meets engineering, and where the "approved vendor list" often conflicts with the "available inventory list." For medical devices, we take a stricter approach than for commercial products. We want to know the end-of-life status for every active component, and we want to see an alternative source for anything that is single-sourced.
We are not asking you to dual-source everything. That is impractical, especially for specialized medical sensors or custom ASICs. But we are asking you to identify the single points of failure in your supply chain. For example, if you have a specific 10µF ceramic capacitor that is only made by one manufacturer in one package size, we will ask you to check the voltage derating. It sounds unrelated, but we have seen projects where a capacitor's DC bias characteristic caused a 30% capacitance drop at the operating voltage, leading to a power supply instability that only showed up in final test.
When you send us your BOM, please include the manufacturer part number, not only the distributor part number. We need to know the exact dielectric and tolerance. We also need to see your moisture sensitivity level (MSL) ratings for all ICs. If you have an MSL-5 part and the assembly house does not have the right baking schedule, you will get internal delamination that passes electrical test but fails in the field after thermal cycling.
We also recommend you review your BOM for passive component package sizes. We are seeing a trend toward 0201 and 01005 resistors in medical wearables to save space. That is fine for automated assembly, but it makes manual rework nearly impossible. If you have a circuit that might need adjustment during bring-up, consider using a larger package for the tuning resistor. It costs a fraction of a cent more and saves a significant amount of rework time.
For a deeper look at how to structure your board for reliability, we recommend reviewing the pcbManufacturing guidelines we have published separately. The key takeaway is that the PCB material and stack-up are part of the BOM, not only a mechanical afterthought.
Test Strategy: The Part Most Designers Skip
We have a saying in our shop: You do not have a product until you have a test that proves it works. For medical devices, this is not a slogan. It is a regulatory reality. But the test strategy is often the last thing designed, and it shows.
We see two common mistakes. The first is relying solely on a functional test that only checks the "happy path." The second is designing a test that is more complex than the product itself, which introduces its own failure modes. The middle ground is a layered test approach:
- In-circuit test (ICT) for passive component values and opens/shorts, if you have the volume to justify the fixture.
- Flying probe for low-volume or high-mix products, which gives us a good measure of solder joint integrity without a dedicated fixture.
- Functional test with a custom load and measurement script that validates the device's primary output signals.
- Boundary scan for digital boards with BGA parts, to verify interconnects without physical probing.
We also ask about burn-in. For some medical sensors, a short burn-in period at elevated temperature stabilizes the output and reduces drift. If your device has an analog front end, we will ask if you have characterized the drift over the first 24 hours of operation. If you have not, we will recommend adding a burn-in step to your production flow. It adds time to the manufacturing cycle, but it reduces the risk of a field failure that is much more expensive to manage.
Component Sourcing and Obsolescence: A Practical Approach
We work with our clients to build a risk register for their BOM. This is a simple spreadsheet that tracks each component's lifecycle status, lead time, and alternative sources. It is not a formal requirements management tool, but it serves the same purpose. We update it every time we place an order or receive a change notice from a manufacturer.
For example, if a major manufacturer announces an end-of-life for a popular microcontroller, we do not panic. We check the last-time-buy date, we look at the alternative pin-compatible parts, and we review the firmware impact. Sometimes the "drop-in" replacement is not actually drop-in because the ADC characteristics are slightly different. That is an engineering problem, not only a procurement problem. We flag it early so you can decide whether to accept the risk or redesign the board.
We also recommend using components that have a proven track record in medical applications. That does not mean you have to use the most expensive part. It means you should look at the part's thermal cycling data, its ESD rating, and its performance under humidity. The datasheet "operating temperature" is not the same as "reliability under stress."
For your next project, we suggest you ask your component supplier for the PCN history (process change notice) of any part you are considering for a high-reliability design. If the manufacturer has changed the die attach or the mold compound in the last two years, you want to know about it. It might be a benign change, or it might affect the part's long-term stability.
Panelization and Assembly: The Physical Realities
We have covered the design and the BOM. Now we get to the part where the rubber meets the road: the physical assembly. For medical devices, we prefer panelized boards with a solid rail rather than a mouse-bite breakaway. The solid rail gives us a stable surface for vacuum fixturing and reduces the risk of board flex during solder paste printing. Board flex is a silent killer of BGA solder joints. It can create micro-cracks that pass a quick electrical test but fail after thermal cycling.
We also pay close attention to the solder paste stencil design. For a medical board with a mix of fine-pitch QFPs and large thermal pads, we will often use a step stencil to control the solder volume. Too much paste on the thermal pad can cause the component to float, leading to poor coplanarity. Too little paste on the QFP leads can cause opens. This is a process engineering task, not only a "send it to the stencil house" task. We review the stencil apertures against the component datasheet recommendations and adjust for the specific board finish.
Finally, we talk about cleanliness. Medical devices often have stricter ionic contamination limits than commercial electronics. If your assembly uses a no-clean flux, we will verify that the residue is within your specified limits. If you require a wash process, we will check that the wash chemistry is compatible with your components and that we have the ability to dry the boards completely. Trapped moisture under a BGA is a reliability hazard.
How to Use This Checklist in Your Next Project
We have given you a lot of detail. Here is how to turn it into action. Before you send out your next RFQ for a medical device PCBA, do a self-review against this list:
- Do you have a test plan that catches the specific failure modes you are worried about?
- Have you reviewed your BOM for single-sourced items and long-lead components?
- Have you checked your DFM for via-in-pad, thermal relief, and rework access?
- Have you specified the cleanliness requirements for your assembly?
- Have you considered burn-in or stabilization for your analog circuitry?
If you can answer "yes" to all five, you are in good shape. If you are unsure about any of them, that is exactly the conversation you should have with your manufacturing partner before you commit to a production schedule. We are happy to review your package and give you our honest assessment. You can reach us through our quote page, and we will respond with specific questions, not generic pleasantries.
We also offer pcba services that go beyond simple assembly. Our engineering team can work with you on DFT, test fixture design, and reliability analysis. We have seen too many good medical devices fail because of a small oversight in the manufacturing process. Our job is to help you find those oversights before they become field failures.
The path to a lower-risk medical device is not about finding a magical contract manufacturer. It is about the engineering dialogue between your team and ours. Bring us your concerns, your test data, and your worst-case scenarios. We will bring our experience with what goes wrong on the line and how to prevent it. That is how we engineer a lower-risk path together.
FAQ
What is the most common DFM issue you find in medical device PCBA designs?
The most common issue we see is inadequate thermal relief on ground and power planes connected to through-hole components. This makes soldering difficult and can lead to cold solder joints. The second most common issue is placing test points too close to tall components, making it impossible to probe them in a fixture. We always recommend designing the test points and the mechanical layout together.
How do you handle component obsolescence for long-life medical devices?
We build a risk register for the BOM at the start of the project. We track the lifecycle status of every active component and identify single-sourced parts. When a manufacturer issues an end-of-life notice, we work with the client to evaluate the last-time-buy options, find an alternative source, or redesign the affected circuit. The key is to have the conversation before the part goes obsolete, not after.
What is the difference between a functional test and an in-circuit test for medical PCBA?
An in-circuit test (ICT) uses a bed-of-nails fixture to check the electrical properties of individual components, like resistance, capacitance, and diode drops. It verifies that the components are placed correctly and soldered properly. A functional test powers up the board and verifies that it performs its intended function, like measuring a signal or controlling an output. We recommend using both for medical devices: ICT to catch assembly defects and functional test to catch design or firmware issues.
