Don’t Ship Bad Units: How Factories Test Electronics
If you’re developing an electronic product, you’ve probably tested your prototype by hand checking voltages, probing signals, or just seeing if it turns on.
That part’s easy.
But what about when you’re building a hundred or ten thousand units?
How do you reliably test every board without wasting hours, missing hidden faults, or slowing down production?
It’s one of the most critical and most overlooked steps.
This article walks through how real-world testing is done at every stage, from early prototypes to high-volume production.
Prototyping Stage
In the early stage, you’re doing everything manually. Multimeter, oscilloscope, maybe a logic analyzer.
You just need to validate the core functionality.
And for small volumes, that’s fine.
But even at this point, it helps to think ahead especially when it comes to test points.
Just like you need access to program the board, you need access to measure and test it.
That means exposing critical signals, power rails, and communication lines with test pads or headers.
And don’t forget about mechanical testing too. Does the enclosure fit? Do the connectors align? Catching those issues now will save you big headaches later.
Small Batch Production
Once you’re building dozens or hundreds of units maybe for beta testers, crowdfunding, or a pilot run manual testing just doesn’t scale.
This is where test jigs come in, with lots of pogo pins. It’s called bed-of-nails testing.
These are custom fixtures that press against your board’s test pads using spring-loaded pogo pins.
Once the board’s in place, the jig powers it up and runs automated tests: checking voltages, I2C communication, sensor outputs, LEDs, buttons, and more.
You can even combine this with in-circuit programming flashing firmware and testing functionality all in one shot.
The key here is automation. Instead of probing things manually, a microcontroller or PC interface runs a test script, logs the results, and tells the operator whether to accept or reject the board.
To make this work well, you’ve got to design your board with testability in mind:
- Add test pads for all key signals
- Label them clearly in your design files
- Space them consistently for alignment with pogo pins
- Avoid placing tall components near test points
High-Volume Production
When you’re building thousands or even millions of units, testing becomes part of the production line itself.
The factory will have automated test stations at different stages:
ICT (In-Circuit Test) checks for shorts, opens, and component values before the board is fully assembled.
FCT (Functional Test) powers up the board and simulates real-world use to confirm that everything works as expected.
Each one has pros and cons.
ICT is fast and reliable, but it needs a custom bed-of-nails fixture, which can be expensive and slow to develop, especially for complex or low-volume boards.
FCT is more flexible and lets you test the actual firmware and peripherals. But it might miss subtle hardware issues, like a missing pull-up resistor or the wrong cap value.
Most of the time, factories use both: ICT to catch build problems early, and FCT to make sure everything actually works.
But depending on your budget and volume, you might choose just one.
You’ll also see Automated Optical Inspection (AOI) at this stage.
AOI uses cameras and image recognition to check solder joints, part orientation, polarity, and component placement.
It’s fast, contactless, and great for catching visual defects before boards move on to more expensive test steps.
Another option is flying probe testing.
Instead of using a fixture, flying probe testers use a small number of moving probes usually 2 to 4 that “fly” around the board to contact test points and vias one at a time.
They measure things like continuity, resistance, voltage, and diode drops.
They’re super flexible, no fixture needed but slower than ICT, especially for big boards or large volumes.
So they’re typically used for:
- Low to mid-volume runs
- High-mix environments where lots of different boards are produced
- Design validation and early testing before building a full test fixture
And sometimes, test fixtures include clever sensors too.
For example, a simple photosensor placed near an LED can confirm whether it lights up when it’s supposed to.
That gives you an automated pass/fail check without needing a human to watch the board.
At this scale, your design choices really matter:
- Add a test mode to your firmware that runs self-checks or exposes internal data
- Use serial output or LED blink codes to show test results without needing a display
- Add test hooks to trigger hardware behavior
- And always make your test points accessible and consistent
Factories often combine all of these techniques depending on cost, throughput, and how complex your product is.
Post-Production Testing and Reliability
Even after production, testing doesn’t stop.
You might need to do:
- Burn-in testing to catch early-life failures
- Environmental testing like thermal cycling, vibration, or humidity
- EMI pre-compliance testing to catch noise issues before certification
You can also use sample-based testing maybe testing 1 out of every 10 or 100 units to keep an eye on long-term quality.
And don’t forget return analysis. If a product fails in the field, have a plan to inspect it, log the failure, and figure out what went wrong. That’s how you improve over time.
Final Thoughts
If you’re serious about bringing a product to market, testing isn’t optional.
It’s how you catch the issues prototypes miss, avoid costly returns, and protect your reputation.
Start thinking about testability as early as your first schematic and you’ll save yourself thousands later.
