DexterLab

🚨 New downloadable modules coming soon📘 Electrical Signaling & PHY Interfaces — new overview📘 Electrical I/O Standards — new overview📘 Integration between Theory and Design Library in progress

Test Points and Observability

Introduction

Test points improve controllability and observability in logic regions that ATPG cannot reach efficiently.
They reduce pattern count, increase coverage, and enable high‑quality testing in both deterministic ATPG and LBIST environments.

Why Test Points Matter

• Reduce pattern count

Hard‑to‑control nodes force ATPG to generate long, inefficient pattern sequences.
A control point breaks the bottleneck and simplifies activation.

• Improve coverage

Unobservable nodes prevent fault effects from propagating to scan outputs.
An observe point exposes the internal value to the scan chain.

• Enable testing of hard‑to‑reach logic

Deep combinational cones, gated clocks, and reconvergent fan‑out regions often require structural help.

• Support LBIST quality

LBIST relies on pseudo‑random patterns.
Test points increase the probability that random patterns activate and observe faults.

Types of Test Points

Control Points

Force a node to a known value during test mode.
Used when controllability is low.

  • AND‑type control point → forces logic 0
  • OR‑type control point → forces logic 1
  • Inserted on selected internal nets
  • Enabled only in test mode

Observe Points

Expose an internal node to the scan chain.
Used when observability is low.

  • Multiplex the internal signal into a scan FF
  • No impact on functional timing
  • Active only in test mode

Hybrid Points

Combine control and observe behavior.
Useful in extremely hard‑to‑test regions or for LBIST optimization.

Automatic Insertion Strategies

Modern DFT tools identify:

  • low‑controllability nodes
  • low‑observability nodes
  • random‑pattern‑resistant faults
  • reconvergent fan‑out regions
  • safety‑critical logic requiring guaranteed coverage

The tool then inserts the minimal number of test points to meet coverage or safety targets.

Safety Connection

Test points are essential in safety‑critical designs (ISO 26262, IEC 61508):

  • Increase diagnostic coverage
  • Reduce undetected dangerous faults
  • Improve latent fault detection
  • Support safety case justification
  • Enable ASIL‑D LBIST quality

They provide structural guarantees that ATPG or LBIST alone cannot achieve.

Diagrams

Figure 1: Control Point (AND‑type / OR-type)

A control point improves controllability by overriding a hard‑to‑control internal node during test mode.
The AND‑type control point forces the node to 0, while the OR‑type forces it to 1.
In functional mode the control point is transparent and does not affect downstream logic.


Figure 2: Observe Point

An observe point improves observability by routing an internal node to a scan flip‑flop during test mode.
In the MUX‑based implementation, a normally observable path is reused to observe a hard‑to‑observe node, selected by TEST_EN.
In the dedicated‑FF implementation, a separate scan flip‑flop is added to observe the node without affecting any functional path.


Figure 3: Hybrid Point

An hybrid point combines a control point and an observe point on the same internal node.
The control point (AND or OR type) improves controllability by forcing the node to 0 or 1 in test mode, while the dedicated observe flip‑flop taps the node upstream of the control logic to preserve its true value for observation.
This structure enhances both controllability and observability without altering the functional path.


Related Pages