From first article to final report without the inspection bottleneck
Willowark builds measurement, monitoring, and quality software for machine shops and metal producers: automated capture from CNC controls and gauges, in-process measurement, tool condition monitoring, and inspection reporting that assembles itself instead of being typed twice. The bottleneck we remove is almost never the spindle. It is the inspection bench and the paperwork stack behind it.
The approach is to measure where the part is, not where the CMM is. Touch probes on the machine, post-process gauging at the cell, and in-line optical measurement all feed one quality database, so control charts populate automatically and drift raises an alarm before a feature walks out of tolerance. Spindle load and current signatures give an early read on tool wear and breakage. Everything is tied to job, machine, tool, operator, and material lot, which is what makes the resulting reports defensible.
Most shops we work with are high-mix and low-volume, which changes what automation is worth. A fixture or a program that pays off on a million parts does not pay off on fifty, so the first project is usually the thing that repeats across every job: getting utilization and cycle data off the controls, or getting measurements into a database without a person typing them. Defense and aerospace shops add export-control constraints on where data can live, so systems are usually built to run on servers inside the building, on a network your IT team controls.
Reviewed
Sound familiar?
If you've said any of these, we should talk.
“The CMM is our bottleneck. Parts sit in a rack waiting to be measured.”
We move routine dimensional checks off the CMM using on-machine probing and automated post-process gauges, reserving the CMM for features that genuinely need it. Results flow into SPC automatically, so releasing a lot no longer depends on one machine and one programmer being free.
“We only find the drift after the whole lot is cut.”
Trending probe results, gauge readings, and tool offset changes turns a slow walk toward the tolerance limit into a visible slope. The alarm fires on the trend rather than the first bad part, which is the difference between adjusting an offset and scrapping a lot.
“The first-article paperwork takes longer than running the job.”
We build FAIR software that carries the ballooned drawing and pulls measured values directly from the gauges and CMM output that produced them, generating AS9102 forms with certifications attached. Transcription errors disappear because the number is never retyped.
“Nobody knows how much life a tool has left before we change it.”
Cutting time and cycle counts per tool come straight off the control, so changes happen on accumulated use rather than a shift-end guess. Spindle load signatures add a second signal that catches a chipped or breaking tool between scheduled changes.
“Half our spindles are idle and we cannot say which ones or why.”
We pull in-cycle, idle, and alarm state from each control and show utilization by machine, shift, and job on one screen. The first weeks of data usually show that idle time is dominated by a few causes — waiting on inspection, waiting on material, setup — which turns a vague capacity problem into a scheduling one.
How this industry actually runs
The operation as we usually find it.
In most shops the constraint is measurement, not cutting. Parts come off a VMC or mill-turn and queue for a CMM that runs one piece at a time while an operator waits on a green light before releasing the lot. Meanwhile the process moves: a worn insert pushes a bore diameter until it leaves a half-thousandth band, and nobody sees it until the next scheduled check. Aerospace and defense work adds AS9100 requirements — ballooned drawings, AS9102 first-article reports, material certifications, and traceability from heat lot to shipped part. Mills and metal producers face the same problem continuously rather than per part, holding gauge, width, and centerline on hot material where contact measurement is not an option.
Machine signals to the people who decide
Components:
- PLCs & sensors (counts, states, current)
- Legacy machine (dry contact / clamp)
- Edge gateway (normalize, buffer)
- Production dashboard (downtime, OEE)
- Alerts & reports (who acts, when)
Connections:
- PLCs & sensors to Edge gateway (EtherNet/IP, Modbus)
- Legacy machine to Edge gateway
- Edge gateway to Production dashboard (MQTT)
- Edge gateway to Alerts & reports
What we build
Starting projects that fit Metals & Machining.
- CNC data collection from Fanuc FOCAS, Siemens, Haas, Okuma, and MTConnect-capable controls
- On-machine probing and post-process gauge integration feeding SPC without manual entry
- CMM output parsing into a quality database with Cpk trending by feature and machine
- Tool wear and breakage detection using spindle load and current signature analysis
- First-article inspection software with ballooned drawings, AS9102 forms, and certification attachments
- Heat-lot and serial traceability from raw material receipt through final shipment
- Vision and non-contact measurement for part presence, orientation, burr detection, and in-line dimensions
- Machine utilization dashboards showing in-cycle, idle, and alarm state by machine, shift, and job
Capabilities we bring
Related work
Where this is running today.
Radar centering system for steel mills
A steel-mill systems provider · Steel manufacturing
Radar-based centering system for steel mills
End-to-end engineering of a radar sensing system that measures and centers material on steel mill lines — from equipment assessment through hardware selection, electrical engineering, software, installation, and commissioning.
Read the case study →X-ray thickness gauge
An industrial measurement company · Industrial measurement
X-ray thickness gauge
A complete X-ray based thickness gauge: equipment assessment, hardware selection, electrical and hardware engineering, API integration, all software, and commissioning — delivered as a working instrument.
Read the case study →Working in Metals & Machining?
Tell us the line.
What runs by hand, what is not connected, what you are trying to build. An engineer replies within one business day with whether and how we would approach it.
Common questions
What Metals & Machining teams ask first.
Can you get useful data off older CNC controls?
Usually yes. Many legacy Fanuc and Siemens controls expose enough over Ethernet or serial for cycle, alarm, program, and offset data, and where they do not we add sensing for in-cycle and idle state. Less detailed than MTConnect, but enough for downtime, utilization, and cycle analysis.
Will in-process gauging let us retire the CMM?
No, and be skeptical of anyone who says otherwise. On-machine probing shares the machine's own geometric errors, so it is excellent for control and adjustment but not a substitute for independent verification. The gain is that the CMM stops doing routine work it was never the right tool for.
How do you prove an automated measurement is trustworthy?
The same way you prove any gauge: with a study. We run repeatability and reproducibility trials against your reference method on real parts across the tolerance range, document the results, and set acceptance criteria before handover. If the system eats too much of the tolerance, we change the approach.
We do ITAR and CMMC work. Can any of this be cloud-based?
It does not have to be. Everything we build here can run on servers inside your facility on a network segment your IT team controls, with no outbound connection required for the system to function. Where a shop chooses a cloud environment, it is typically one already scoped into its compliance program, and we design to fit that boundary rather than introduce a new one. Data residency and access control are settled in the requirements, not afterward.
Can vision measure machined parts accurately enough to replace a gauge?
For some features, yes; for others, no. Non-contact optical measurement handles presence, orientation, hole patterns, edge profiles, and burrs well, and can hold useful accuracy on larger dimensions with telecentric optics and controlled lighting. It typically struggles with tight bores, deep features, and highly reflective surfaces, where a probe or a dedicated gauge is still the right tool. We answer feature by feature against your drawing rather than in general.
Strategy. Software. Systems.
Engineering for Metals & Machining.
Describe the problem in your own words. An engineer reads it — not a sales script — and tells you plainly what it would take.