The iron stays. The data layer modernizes.
Native FOCAS2 on Fanuc 16i, 18i, and 0i — not just the newest controllers. MTConnect, Brother HTTP, and Modbus TCP on whatever else you run. Modern operational visibility from the machines you already own, validated, and trust.
Live integrations: FOCAS2 (Fanuc 0i · 16i · 18i · 21i · 30i · 31i · 32i) · MTConnect · Brother HTTP · Modbus TCP — and OPC UA Client and Siemens S7 for the rest of the floor. FANUC MT-LINKi REST on the roadmap.
Real plants don't have one generation of controllers. They have five.
Real plants don't have one generation of controllers. They have five. A Fanuc 16i installed in 2009 next to a 32i from last year. A Brother S700Xd1 that's been running parts for eight years. A handful of older CNCs fronted by Modbus PLCs because the original controller's interface was never going to survive corporate IT review. Every one of those machines is making good parts.
Operators know them by feel. Tooling and fixtures are validated. The capital was depreciated years ago. And yet what corporate now wants — OEE numbers, alarm history, shift reports, audit trails — keeps getting pitched back as a hardware problem. The trap most plants fall into is accepting that modernization means replacement. It doesn't. What corporate wants is a data requirement, not a hardware requirement.
The iron can stay. The data layer is what needs to catch up. The plants that get there put one protocol-agnostic runtime in front of every controller — old and new — normalize every signal to one vocabulary at the edge, and let the existing systems keep doing their jobs. No machine comes offline. No operator changes a thing.
"The case for ripping them out exists only in vendor presentations."
How this differs from rip-and-replace modernization. The other way to get modern reporting off a brownfield floor is to forklift the controllers — replace the older CNCs with new iron that ships with a modern data interface. It works, eventually, on someone else's capital schedule: a multi-month rebuild project, re-validation of tooling and fixtures, operators relearning machines they knew by feel, and a controller-replacement bill that exists to solve a reporting gap. Elpis does the opposite. One protocol-agnostic runtime reads the controllers you already own — including the oldest Fanuc generations — over their native protocols, normalizes every signal to canonical vocabulary, and feeds the modern analytics layer corporate wants. The machine stays. The validation stays. The operator workflow stays. Only the data layer modernizes. A controller upgrade, if it ever comes, happens on your capital schedule — not as a precondition for visibility.
spindle_rpm in the pipeline. A cycle-complete signal off the 16i, the Brother, and the Mazak collapses the same way into a canonical cycle_time — the exact signal EREMOS V2 turns into OEE. Feed rate, parts count, tool number, axis positions, alarm codes — same names, same semantics, regardless of which controller generation produced the signal. The same dashboard works across every vintage on your floor.Your operators don't need to change a single behavior.
From EdgeConnect (edge runtime, Windows service today)
- FOCAS2 collector across every Fanuc generation — 0i, 16i, 18i, 21i, 30i, 31i, 32i. Axes, spindle, alarms, tool, production counters, programs. The protocol coverage does not quietly drop your oldest iron.
- MTConnect collector — for the newer multi-vendor CNCs already speaking the open standard.
- Brother HTTP collector — Brother S700Xd1 and similar models via the built-in web-monitoring interface.
- Modbus TCP collector — for older PLC-fronted CNCs, and the energy meters, drives, and instrumentation you may have already wired.
- Also today: OPC UA Client and Siemens S7 for the rest of the floor. On the roadmap: FANUC MT-LINKi REST. For the full protocol matrix with semantic modes → /edgeconnect.
- Canonical vocabulary across vendors and generations —
running,spindle_rpm,feed_rate,parts_count,cycle_time, axis positions (axes/x/absolute, etc.), tool number and offsets, alarm codes. Same names regardless of which controller generation produced them. - Read-only / observational by design — EdgeConnect polls the controller; it does not change control logic, operator HMIs, or alarm acknowledgment. The CNC runs independently of the platform's status.
- Per-route store-and-forward buffering — older plant networks aren't always reliable; signals queue at the source and replay in source order when connectivity returns.
- Three-way diagnostics — source / pipeline / sink. When something goes wrong on the floor or in IT, operators know where it broke.
- Connectivity Studio — web admin to add machines, configure tag maps, and run Test Connection probes before anything goes live. No command-line config required.
- Hash-chained configuration audit log — tamper-evident change history from day one, even where formal change-control was previously informal.
EdgeConnect ships as a Windows service today; brownfield plants typically run it on a small box that's already in the control cabinet, not on new hardware. A Linux runtime is near-term roadmap, arriving on the Edge Gateway appliance for plants that prefer a turnkey DIN-rail box. The appliance is an option, not a requirement.
From EREMOS V2 (intelligence layer, consuming the canonical stream)
- OEE Segments — RUNNING, PLANNED_STOP, UNPLANNED_STOP, IDLE, SETUP. Computed from edge-collected signals; auditable.
- Persistent alarm tracking with incident grouping — alarms from the oldest CNC on the floor become tracked records with open/close state, not just blinking lights on a machine's HMI.
- Tool-life ingestion — a dedicated path for tool-wear telemetry, so maintenance gets ahead of failures.
- Shift reports — PDF and Excel, built from edge-collected signals — the reports corporate has been asking for.
- Multi-tenant by design — one EREMOS V2 across multiple sites if you operate more than one plant; no data leakage.
- Dashboards split by device class — CNC, PLC, meter. Mixed-generation fleets render cleanly.
What every brownfield floor asks before scoping.
Will deploying this require us to take a machine offline?
No. EdgeConnect connects to the controller as a read-only client over its native protocol; polling happens alongside normal operation. No machine downtime is required for installation, and the platform doesn't touch control logic or operator workflow. If EdgeConnect is offline for maintenance, the CNC keeps producing parts — it operates entirely independently of the platform's status.
What about our oldest CNC — does it work with that?
If it's a Fanuc 0i / 16i / 18i / 21i / 30i / 31i / 32i, yes — FOCAS2 covers all those generations, not just the newest one or two. If it's older or proprietary and doesn't expose a native data interface, a Modbus TCP gateway in front of the controller bridges it to the platform. OPC UA Client and Siemens S7 cover the rest of the floor today. Bring the controller list to the scoping call and we'll confirm the collection path per machine before quoting — every brownfield floor has at least one weird controller that needs case-by-case evaluation.
Do operators need to retrain?
No. The platform reads from the controller; it doesn't change how the controller is operated. Operators continue using the same MPG, the same HMI, and the same parts programs they already know. Nothing about the modernization lands on the operator.
If the platform breaks, do our machines stop?
No. EdgeConnect is observational — the CNC operates independently of EdgeConnect's status. If the platform is offline for maintenance or the network drops, your floor keeps producing parts. On the network side, per-route store-and-forward queues every signal at the source with its quality code preserved and replays in source order when connectivity returns, and three-way diagnostics (source / pipeline / sink) surface immediately so the OT team sees exactly which leg was affected.
Can we start with one machine and decide whether to expand?
Yes — that's the recommended deployment pattern, and it's the point of an incremental, reversible rollout. Pick the controller you're most skeptical about, prove the platform against it on your real protocol, and expand only if it earns its place. No multi-year commitment, no forklift upgrade. The full rollout cadence is in "How brownfield plants typically roll this out" below.
How does this integrate with the SCADA, MES, or historian we already have?
Elpis sits beside them, not in place of them. EdgeConnect publishes the canonical stream to MQTT (or OPC UA Server) — standards your existing systems likely already consume — so your SCADA, MES, and historian keep their jobs and receive consistent canonical signals instead of vendor-specific ones. The modernization adds the cross-generation data layer; it doesn't replace the operational systems you've already invested in.
What changes when this lands.
- Modern visibility from machines you already own — OEE, alarms, and reports without replacing a single controller
- OEE reporting without capital expense on new iron — the platform earns its place in months, not over a multi-year capital cycle
- Audit trail from day one — hash-chained change history starts the moment you turn the platform on, without ripping out the existing configuration plane
- Mixed-generation fleets behave like one operational system — canonical vocabulary normalizes everything from a 2009 Fanuc to last year's Mazak
- Cut the multi-month rebuild project — incremental deployment, reversible at every step, no forklift upgrade
- Operator workflow unchanged — the machines operate the way operators already know
- The data layer outlasts the iron — when a controller does eventually change, on your capital schedule, the platform already handles whatever you put in
How brownfield plants typically roll this out.
Pick the controller you're most skeptical about. EdgeConnect installed on a small Windows box already in your control cabinet, polling that machine over FOCAS2 or Brother HTTP. Data flowing to a Mosquitto broker we set up alongside, or to your existing MQTT broker. EREMOS V2 displaying real cycle-time and alarm data, typically within the first few days. If it works for that machine, it'll work for the rest of your floor.
Add the rest of the machines on one line or cell. Tag-map authoring done together with your team — your operators know the names that matter. Shift-report templates configured against your actual shift schedule; OEE Segments aligned to how your plant already defines OEE.
Remaining CNCs onboarded — the newer ones, the Modbus-fronted older ones, anything PLC-fronted. Multi-site or multi-line aggregation in EREMOS V2 where applicable. Alerting routed to the channels your operations team already uses.
When you eventually do replace a controller — years from now, on your own capital schedule — the platform already handles whatever you put in. FOCAS2, MTConnect, Brother HTTP, Modbus TCP, OPC UA Client, and Siemens S7 all ship today; FANUC MT-LINKi REST is on the roadmap. Your modernization investment is in the data layer, not in the iron — so when the iron eventually changes, the data layer doesn't.
How it fits together on a brownfield floor.
Read-only on the floor; nothing depends on the cloud. EdgeConnect connects to every controller as a read-only client and never changes control logic — your machines run independently of the platform. It also runs offline by default: the license validates locally, with no phone-home. If your network or broker drops, per-route store-and-forward buffers locally and replays in source order on reconnect. Plants on an isolated network install and run the platform the same way as plants with internet; cloud connectivity is opt-in, not required.
Per-gateway identity + hash-chained configuration audit from day one. Each plant runs its own EdgeConnect runtime with a per-gateway UUID established at first start. Every change — a machine added, a tag-map edit, a threshold change — is captured with actor identity and timestamp in a tamper-evident, replay-ready audit chain, even on a floor where formal change-control was previously informal.
Read the full operational trust posture → /security
Looking for the same thing from another angle?
Bring us your oldest CNC.
Pick the controller you're most skeptical about. We'll scope a proof of value against that machine specifically — on its real protocol, against its real signals. Demos run on your real iron, not on a polished newer machine staged for the camera. No canned data, no slideware, no forklift upgrade.