Litmus has released Litmus Edge 4.4.0, a long-term support (LTS) version of its industrial edge data platform that adds asset hierarchy context to Digital Twins, a new Litmus Unify Connector and a single AI Suite page for managing models. The company says the release delivers data "contextualized" before it leaves the plant. Litmus Edge 4.4.0 is available for download from Litmus Central.
Digital Twins now carry each asset's hierarchy path and static metadata. That means the enterprise, site, area, line and machine a value belongs to travel with every message. Static attributes, dynamic attributes and flattened payloads each get their own topics, so a consumer can subscribe with wildcards to only what it needs. Litmus says analytics teams can compare a filler in one plant with a filler in another without a mapping spreadsheet.
For quality engineers, the Atlas Copco MicroTorque MTF 6000 support addresses a traceability gap. These low-torque controllers, used in electronics and medical device assembly, do not support MID 61, the Open Protocol message most drivers rely on for tightening results. In 4.4.0, the Open Protocol driver reads Last Tightening Results (MID 1201 and 1202) through dynamic PIDs that users can browse. It also adds Generic Subscribe (MID 8), connection result data (MID 2) and read-time polling (MID 80 and 81). Litmus says the result is a timestamped record of each tightening for traceability, audits and SPC.
The Litmus Unify Connector publishes data from Litmus Edge to Litmus Unify and builds topics from each asset's hierarchy path and metadata. It supports MQTT, MQTTS, WebSocket and secure WebSocket, and is activated with a Litmus Unify activation token. The existing Litmus Unify connector keeps running alongside it. Edge Cascading v2 links Litmus Edge instances so data can move from a plant-floor node to a node in the DMZ or at site level, and from there to enterprise and cloud systems.
On the AI side, one AI Suite page replaces the separate model page in Analytics. The section formerly named Connections is now called AI and holds connections to AI model providers, while a new TensorFlow section manages tensor models next to them.
The OPC UA Advanced Client driver can now call OPC UA methods, which are functions on the machine such as starting a job, resetting a counter or loading a recipe. Serial drivers also move from the Gen1.3 to the Gen2 driver architecture with full parity for register and data-type reads and writes. Both generations run side by side, so each connection can be migrated on its own schedule.
