Jul11
What Is the Purpose of Testing Electronic Products? It’s Like Buying Insurance!

電子產品測試的目的所為何在?買保險也!

Workingbear has been in the electronics industry for several years and has noticed an interesting phenomenon. Almost every new manager will ask the same questions when they come on board: “Why do we need to perform ICT (In-Circuit Test) at the board level assembly?” and “Can we skip ICT tests?”

This is mainly because a “Keysight (Agilent) 3070” ICT test fixture can cost around US$15,000, equivalent to the price of a small car. And this cost is just for the fixture, excluding the test equipment. On the other hand, a simpler MDA (Manufacturing Defect Analyzer) fixture, such as the “TR-518,” only costs around US$1,700.

The following discussion focuses solely on board-level (PCBA) production. If we consider whole system production, which includes PCBAs, the considerations would be different, as whole system production also involves testing.

So, performing circuit tests after Printed Circuit Board Assembly (PCBA) is not only expensive but also time-consuming. From a lean manufacturing perspective, testing doesn’t seem to contribute directly to output. So why do companies still do it?

Before answering this question, it’s important to understand the PCB assembly test process. Typically, traditional circuit boards go through AOI (Auto Optical Inspection) after SMT assembly. AOI checks for defects like component misalignment, soldering shorts, tombstoning, and missing parts. After AOI, the boards undergo ICT or MDA circuit testing, followed by FVT (Functional Verification Tester) functional testing to ensure all the board’s functions are working correctly.

PCBA-process-flow

Among these, ICT fixtures are the most expensive, FVT takes the most time, and AOI isn’t used by every company and has limited detection rates. For those interested in further understanding these three types of PCB quality tests, you can refer to this article: “Discussion of Various Testing Methods After PCB Assembly: AOI/MDA/ICT/FVT/FCT.”

Here’s the problem: with so many testing methods available to ensure the quality of assembled circuit boards, can we skip the more expensive ICT/MDA tests? Isn’t AOI and FVT enough to guarantee the quality of the boards?

The issue is that both AOI and FVT have significant testing blind spots and cannot achieve 100% test coverage. That’s why multiple testing methods are necessary to achieve the highest possible test coverage.

But isn’t FVT a full functional test? How can it not achieve 100% test coverage?

This is because, during the circuit board design phase, there are bypass circuits and some components that FVT cannot test. These bypass components are not triggered during normal functional testing operations, such as overcurrent protection circuits, ESD protection circuits, or EMI shielding circuits.

Additionally, AOI and FVT cannot verify the resistance, capacitance, and inductance values of components on the circuit board. Only ICT/MDA can measure these values. In other words, ICT/MDA still has irreplaceable functions.

Now, let’s discuss an even hotter topic. Many managers view PCBA testing through the lens of production efficiency or Lean Manufacturing, believing that ICT or FVT testing has no productive value. Why not eliminate all PCBA testing steps and processes from the production line? Or just make the product perfect from the start, so no testing is needed, right?

This idea sounds great, and Workingbear would applaud it, but only in theory. If you’re in charge, unless you’re completely indifferent to quality issues after the product is shipped, removing testing is just inviting customer complaints, returns, and eventually, your own dismissal.

Workingbear can only say, “Testing” is like buying insurance for our products. Just as you drive carefully but can’t guarantee you won’t have an accident, you can’t ensure your products will never have defects. Sometimes accidents are unavoidable, whether caused by you or someone else. Can you guarantee that your products will always be flawless? Or that selling defective products won’t lead to complaints or returns? If not, then don’t even think about eliminating testing.

No matter what you do or how you streamline the process, you must ensure the quality and functionality of the product before delivering it to the customer.

Here’s what you can do: decide how much “insurance” to buy for your product, meaning how much testing is needed to ensure the quality of the product before it ships. So, how much testing is reasonable?

Workingbear believes this depends on the specific product. If it’s a product related to human life and safety (cars, airplanes, medical devices, and military products fall into this category), you should aim for 100% test coverage, no matter the cost, because you can’t afford to risk lives.

Next are products related to financial transactions. It’s recommended to also strive for 100% test coverage, as the potential financial losses for customers can be enormous, potentially sinking your entire company.

For general consumer products, there is some flexibility, and it depends on how much risk you’re willing to take. Some companies might indeed skip ICT/MDA testing.

For cheap toys, it might just come down to FVT testing. As long as it works, it’s fine. In some cases, companies might skip testing altogether and let customers do the testing. If it breaks, they just replace it—one for one. (That’s not what Workingbear suggests; it’s just something I overheard from someone else.)

Do you have your own thoughts on “testing”? Feel free to leave a comment and join the discussion.


Related Posts:

Leave a Reply

Your email address will not be published. Required fields are marked *