> For the complete documentation index, see [llms.txt](https://help.tillit.cloud/tillit/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.tillit.cloud/tillit/knowledge-base/setting-up-tillit/causal-analysis/detailed-data-and-hardware-flow.md).

# Detailed Data & Hardware Flow

To understand how Causal Analysis works, it helps to trace the journey of an alarm signal from the physical machine to the Tillit user interface:

<figure><img src="https://352909507-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FrNT40XJFAFlAhUX8CcQS%2Fuploads%2FOdhIuFmmKiEzYDwO2AYE%2Fimage.png?alt=media&amp;token=ae85c92d-1f2e-41b3-b404-358146c3ad5e" alt=""><figcaption></figcaption></figure>

#### 1. Alarm Generation (PLC Level)

When a machine component experiences a fault (e.g., a motor overheats, a sensor is blocked, or a safety door opens), the PLC generates an internal fault bit or error code (e.g., ERR\_102).

#### 2. Tag Mapping & Capture (Ignition Level)

* In Ignition, engineers bind these PLC error codes to Ignition OPC tags.
* The Tillit Ignition SDK Module monitors these tags in real time.
* When the main bottleneck asset (e.g., the *Filler*) transitions to a Stopped state, the SDK automatically collects all active and cleared alarm signals triggered across related assets on the line within a predefined time window (e.g., the last 30 seconds).
* The SDK stores recent telemetry in a local circular buffer based on the configured Event Retention window (defaulting to 5,000 events) to prevent data loss during network latency.

#### 3. Payload Transmission (MQTT)

The collected context (containing timestamps, alarm codes, and machine states) is packaged into a structured payload and published to Tillit via MQTT.

#### 4. Logic Processing (Tillit Engine)

Upon receiving the payload, Tillit’s Causal Analysis engine evaluates the incoming data against the Active Rule Set.

The engine checks which rule conditions match the combination of active alarms, machine states, and timing.

#### 5. Automated Output (Tillit UI)

Tillit assigns the matching Event Reason and Machine Reason to the downtime record.

* **Event Reason** (*Why it happened*):
  * Root-cause tree hierarchy up to **5 levels deep**.
  * *Example:* Technical > Mechanical > Mechanical Malfunction > Mechanical Overload.
* **Machine Reason** (*Where it happened / Affected Component*):
  * Asset location hierarchy up to **3 levels deep**.
  * *Example:* De-palletizer > Hoist > Sweep Arm.

On the Downtime Tracking screen, operators see the suggested root cause marked with a lightbulb icon. Operators can accept suggestions individually or in bulk using Accept Causal Analysis Suggestion.
