ToolLife Edge
Change the tool before the part becomes scrap, from spindle current and acoustics alone.
- Category
- Tool condition monitoring
- Model
- Edge appliance + SaaS
- Core platform
- ALIA Signal Engine
- Status
- IN DEVELOPMENT
01 Problem
CNC shops scrap parts and crash spindles because tool wear is invisible until it is too late. Conservative shops change tools early and burn money on tooling, and aggressive shops change late and burn money on scrap.
The cost is measured in ruined parts, broken tools and machine hours lost to both.
02 Specification
This is the specification we are building toward. The status column shows what exists today.
- Real-time wear estimationIN PROGRESS
Tool wear is estimated in real time from spindle current and acoustic signatures alone.
IN PROGRESS - Non-invasive installPLANNED
Installation is a current clamp and a microphone, with no spindle encoder access and no machine modification.
PLANNED - Change-before-scrap recommendationsPLANNED
The system recommends a tool change in the window after useful life is spent and before the part is ruined.
PLANNED - Material and tool profile libraryPLANNED
Wear models are tuned per material and tool combination, and the library grows with use.
PLANNED - Breakage early warningPLANNED
Signatures that precede tool breakage raise an alert before the break, not after the crash.
PLANNED - Per-job tool-cost reportingPLANNED
Tooling cost is attributed per job, so quoting can reflect what a part actually consumes.
PLANNED - Machine-agnostic operationPLANNED
The system works across controls and machine brands because it never needs to talk to the control.
PLANNED - Fleet view across machinesPLANNED
Every monitored machine’s tool state is visible in one view across the shop.
PLANNED
Updated as capabilities ship. BUILT means working and validated; nothing gets that label early.
03 System
- 01Current and acoustic sensors
- 02Edge feature extraction
- 03Wear model
- 04Operator alerts
- 05Tool-cost ledger
04 Engineering
Signal core
ToolLife Edge is built on the ALIA Signal Engine. Spindle current and acoustic streams are normalized per machine, features robust to coolant noise and shop-floor acoustics are extracted at the edge, and wear and breakage models are maintained per material and tool combination in a growing profile library.
Deployment model
All inference runs on the edge appliance beside the machine. Alerts work without an internet connection, raw signal data stays in the shop, and the appliance never connects to the machine control, which is what makes the system machine-agnostic.
Integration surface
Installation is a clamp on the spindle feed and a microphone in the enclosure. Alerts reach operators at the machine, and tool-cost figures export into the job-costing spreadsheets shops already keep.
05 Why this is hard
The genuine problem is generalizing wear estimation across materials, tools and machines with limited labeled wear data. Ground truth requires measuring worn tools, which shops rarely record, and a model tuned to one machine can fail on the machine beside it. The engineering work is transfer across machine and material classes, plus a calibration path that builds each shop’s profile library from ordinary production instead of dedicated trials.
06 Design constraints
- Non-invasive install: a clamp and a microphone
- Never touches the machine control
- Alerts work without an internet connection
- Raw signals stay in the shop
- Works across controls and machine brands
07 Design partner program
Design partner program.
We are looking for a small number of design and pilot partners. A partner gets the problem solved for their operation first, direct influence on the specification above, and pilot pricing. In exchange we get reality: real constraints, real data, real operators. No external partner is engaged on this product yet.