Access Official Technical PDF

Share Summary:LinkedInEmail

Embedded board test (EBT) can help engineering and operations teams address a persistent challenge: as components become more complex and PCB access becomes harder, conventional probing may not provide enough visibility for effective testing. Planning for testability early can preserve access to important signals, support clearer fault diagnosis and make board-level tests useful beyond manufacturing.

EBT combines methods such as Boundary Scan, processor-based test and chip-embedded instruments. These approaches can complement automated test equipment (ATE), rather than replace it. The right combination depends on the product, available test access and the coverage required.

Why embedded board test needs design-stage planning

Ball-grid array packages, fine-pitch connections, buried PCB traces and multilayer boards can limit physical access to circuit nodes. At the same time, high-speed interfaces and increasingly integrated devices make it harder to test every function with a probe or fixture alone. Design for testability (DfT) addresses these constraints by building test access and control into the product architecture.

IEEE 1149.1 Boundary Scan, commonly called JTAG, provides a way to access digital I/O through test logic inside compatible ICs. It can support connectivity checks for faults such as opens, shorts and stuck-at signals without requiring direct physical access to every node. The same test access port may also support in-system programming, built-in self-test access, debug and system-level testing.

Test planning should account for lifecycle needs, not only production inspection. A test developed from the schematic can support prototype verification, manufacturing diagnosis, programming, service and repair. Combining EBT with conventional ATE may be appropriate when one method alone does not provide the desired coverage or test performance.

Choose devices and prepare test resources early

Where device selection allows, prefer ICs with IEEE 1149.1 support and obtain accurate Boundary Scan Description Language (BSDL) files from manufacturers early in development. Check that each file is syntactically valid and review its compliance conditions and design warnings. For example, a BSDL file may specify logic levels required on control pins before Boundary Scan testing can operate as intended.

Also check whether devices provide built-in self-test (BIST). BIST can exercise internal circuitry, and some implementations can help test externally connected resources such as memory. Its results may be limited to pass-or-fail information, so it should be considered alongside methods that provide more detailed fault localization.

For programmable logic, review in-system configuration requirements as part of the test plan. IEEE 1532 provides a standard approach to programming compliant programmable logic devices, including the possibility of handling multiple devices from different vendors. Where board logic is still being designed, placing suitable logic in an FPGA may also provide test resources that discrete logic components do not offer.

Design a robust JTAG scan chain

A conventional scan chain connects each device’s TDO output to the next device’s TDI input. TCK and TMS are shared across the devices. Careful implementation matters: poor signal quality or unsuitable routing can disrupt TAP synchronization and produce intermittent scan errors that are difficult to diagnose.

Useful design practices include:

  • Standardize test access. Use a consistent connector and pinout across product variants where practical. Place the connector near a PCB edge and consider alternate access through an edge connector.
  • Plan control-pin states. Use pull resistors to establish functional-mode defaults for compliance, boot and test-control pins, while providing a way for the test system to set required test levels.
  • Protect signal integrity. Route TCK, TMS and reset signals with attention to ringing, overshoot, crosstalk and termination. Treat shared test-bus signals with the care required for high-speed nets.
  • Check drive and voltage compatibility. Longer chains and differing I/O voltage levels may require buffering, level shifting or separate chains. Confirm device requirements rather than assuming all parts can share one electrical interface.
  • Support diagnosis and variants. Test points on otherwise inaccessible TDO-to-TDI connections can help locate chain faults. Provide bypass paths where optional or unpopulated devices would otherwise interrupt the chain.

Devices that are noncompliant, use a different debug interface or have special operating requirements may be better placed in separate chains. A device that controls power to other scan-capable ICs also needs careful consideration: the power-controlling device may need to be configured before the other devices can be tested.

Make non-scan circuitry controllable

Boundary Scan coverage depends not only on the scan devices themselves, but also on the surrounding circuit. Make important reset, watchdog, power-control and clock signals accessible where possible. A watchdog that resets the board during a test, or a reset signal that prevents a device entering test mode, can appear as an unreliable scan chain.

For buses connected to non-Boundary Scan devices, provide a way to disable their outputs when needed. This helps avoid contention during connectivity tests. Keep logic clusters small and, where feasible, connect their inputs and outputs to Boundary Scan pins. Direct access to memory-device signals—including clock inputs for synchronous memory—can simplify test development and improve diagnostic detail.

Spare FPGA or processor pins may also provide useful access to enable, direction or other control signals. Even when this does not add substantial fault coverage, it may make failures easier to isolate.

Plan programming and extend coverage

In-system programming can be part of an EBT strategy, but access to programmable devices must be designed in. For Boundary Scan-based memory programming, scan-chain length and TCK frequency affect data throughput. The slowest device in a shared chain can constrain the usable clock rate, so a separate chain may be worth considering for a device with different programming requirements. Direct access to a Flash write-enable signal can also be considered when the design allows other drivers on that net to be disabled.

Boundary Scan is primarily a quasi-static method, so it does not address every dynamic or high-speed behavior. Coverage can be extended by combining it with processor-based test, embedded instruments, device-specific test modes such as NAND-tree tests, or suitable I/O loopbacks. ADCs and DACs may provide an indirect path for involving analogue circuitry, while optical indicators and switches can be included when they can be controlled and observed as part of the test setup.

Make testability a cross-functional decision

Effective embedded board test is a design decision shared by hardware engineers, test engineers, production, service and procurement. Early coordination helps teams agree on access points, scan-chain architecture, device requirements and lifecycle uses before layout decisions make changes costly.

The practical goal is not to maximize test features without regard to constraints. It is to preserve useful access, control and diagnostic options, then select a combination of test methods suited to the board. Thoughtful DfT can make prototype debugging, manufacturing checks, programming and field repair more manageable while helping teams target test effort where it provides the greatest value.

Access Official Technical PDF

About The Author