Fault Insertion Unit vs In-Circuit Fault Injection

Two complementary ways to inject faults in HIL testing
At the connector, with a fault insertion unit (FIU) — or inside the board, with in-circuit hardware fault injection. What each one reaches, and what evidence it gives for ISO 26262.

Connector-level fault insertion vs. in-circuit fault injection

Conventional HIL fault insertion units are switching matrices (relay or solid-state) placed between the test system and the DUT connector. They open lines or short them to ground, battery or another pin: an ohmic contact, on or off. They are designed for connector-level faults — the internal nodes of a multilayer PCB are out of their reach.

Novasom’s system is a true hardware signal injector. Through a dedicated bed-of-nails fixture it works directly on the board’s test points: it imposes voltage or current with 16-bit resolution on each node and reads the response on the same point. Hundreds of points in a few hundred milliseconds, every actuation and reading timestamped at 0.25 µs (4 MHz timebase).

Novasom Industries validation

Side by side

Conventional HIL fault insertionNovasom in-circuit fault injection
WhereDUT external connectorInternal test points, JTAG pads, internal buses
HowSwitching matrix: ohmic open / shortVoltage and current injection, 16-bit resolution
MeasurementSeparate instrumentStatic and dynamic readback on the same point
ScaleTens of channels per unit500+ internal points in the field
TimingSwitching time per channel5–6 ms per event, 200–300 ms per full sweep, 0.25 µs timestamp on every actuation and reading
Evidence for ASIL C/DSystem reaction to external faultsDiagnostic coverage and latent faults of internal safety mechanisms

When do you need in-circuit fault injection?

  • ASIL C/D programs where diagnostic coverage must be shown, not assumed
  • Safety mechanisms that can only be exercised from internal nodes
  • Multilayer boards where test pads are the only way in
  • FMEDA/FTA campaigns with thousands of scenarios to run deterministically

Both approaches work together: the FIU keeps covering the harness, the nail-bed fixture covers the board.

Frequently asked questions

Is a fault insertion unit enough for ISO 26262?
ISO 26262 does not prescribe one method. A FIU is the right tool for faults at the connector. Where a safety mechanism can only be exercised from an internal node, internal fault injection is the only way to produce experimental evidence.

Can we use both on the same HIL?
Yes. In-circuit fault injection is complementary to commercial HIL platforms and their FIUs, and can be controlled from the same test sequencer.

What is a bed-of-nails fixture for fault injection?
A fixture designed for one specific PCBA, with spring probes on its internal test pads. Ours is co-designed with your FMEDA, so every probe serves a fault scenario — reading and injecting on the same point.

Does hardware fault injection replace the FMEDA?
No. The analysis claims, the test proves: fault injection verifies the diagnostic coverage the FMEDA declares, on the real board.

See also: ADAS HIL testing · Fault Injection Systems & HIL

At your service 24/7

You ask, we will solve all your problems!

Live demo

See in-circuit fault injection running on our reference hardware, on site or remotely.

Technical session

Discuss your ASIL C/D program, your FMEDA and your test points with our engineers.

fault insertion unit