Machine-side witness
Runs at the machine edge where timing, scan windows, signal transitions, and process context can be preserved close to the source.
A.I.R.O.N. flagship evidence system
Ghost Busting™
A live digital twin comparator for the problems that disappear before maintenance can prove them. A.I.R.O.N. mirrors expected behavior against actual machine behavior every scan, locks the surrounding context, and traps the ghost before it vanishes.
1983 Ghost Hunting origin • modern A.I.R.O.N. / V.A.U.L.T. comparator • humans remain authoritative
Digital Twin ComparatorTruth Window Armed
Select ScopeOperator, technician, or engineer selects the suspect signal, step, station, circuit, or process window.
Mirror Expected StateA.I.R.O.N. maps the expected behavior from the digital twin, logic path, sequence step, or process model.
Compare Every ScanExpected behavior is compared against actual live machine behavior scan by scan.
Context LockWhen the live machine misses a beat, the surrounding truth window is preserved.
Ghost TrappedEvidence is presented so humans can verify, repair, improve, and preserve the lesson.
Core definition
Expected versus actual — every scan.
Ghost Busting™ is not a human interview function. It is a live digital twin comparator. The expected machine, circuit, station, or process behavior is mirrored at the edge and compared against actual live behavior. When the two disagree, A.I.R.O.N. captures the mismatch before downstream symptoms bury the initiating cause.
Expected digital twin state = Actual live machine state
If equal: continue watching
If not equal: capture the ghost
Position statement: Ghost Busting™ is preserved truth, bounded interpretation, and repeatable visibility. It captures proof; humans decide what to do with it.
Controller/HMI mirror
Can host or support local comparison logic where the controller, HMI, and I/O truth already live together.
Observe before altering
Ghost Busting™ begins in read-only or shadow mode unless a controlled, engineered, validated action path is explicitly approved.
Origin and evolution
Ghost Hunting became Ghost Busting™.
The original 1983 field discipline was called Ghost Hunting. It began as practical shop-floor evidence preservation at Iroquois Tool Systems under Walter W. Dingfelder, using a mobile, ad hoc diagnostic platform built from commercially available instrumentation to catch transient machine behavior before it disappeared.
The original field discipline
Timers, counters, comparators, indicators, test points, and practical witness logic were used to prove whether hidden events actually happened.
Meaning: we are looking for the ghost.
The modern captured-proof discipline
A.I.R.O.N. and V.A.U.L.T. make the witness native: context lock, digital twin comparison, evidence bundles, and institutional learning.
Meaning: we nailed it. It is in the trap.
The tools changed. The doctrine did not. Capture what others miss, preserve the moment, strengthen the team, and keep Continuous Improvement alive under real conditions.
AI-assisted implementation
A.I.R.O.N. can turn logic into trap-ready sequence steps.
With AI, A.I.R.O.N. can rapidly create a digital twin of nearly any logic platform, filter the logic into processes, and break those processes down into individual sequence steps for Ghost Busting™ monitoring. That allows the operator to select the cycle stage where the ghost is suspected and arm a trap during operation.
1
Import / read logic truthPLC, HMI, I/O, drives, historian, gateway, or supported platform data becomes the source for expected behavior.2
Filter into processesA.I.R.O.N. groups logic into stations, cycles, machine states, handoffs, permissives, and process windows.3
Break into stepsEach process becomes individual sequence steps with expected commands, confirmations, timers, counters, and outcomes.4
Operator selects stageThe operator chooses the suspect cycle stage and arms the Ghost Busting™ trap without waiting for maintenance to witness the fault.5
Compare every scanThe comparator watches expected versus actual behavior inside the selected step, circuit, station, or process window.6
Trap the mismatchLate, early, missing, false, weak, bounced, or broken events are captured with their surrounding context.7
Present evidenceA.I.R.O.N. shows where intent stopped becoming outcome and what discipline should verify next.8
Humans decidePeople retain safety authority, repair responsibility, CI judgment, and final operational decisions.Context lock
It does not stop the machine. It stops the loss of information.
Context lock preserves the bounded truth window around the event: the machine state, command, confirmation, timing, recipe, operator mode, inputs, outputs, timers, counters, permissives, analog values, previous event, and next expected event.
Capture the first mismatch
The initiating ghost is often hidden upstream of the visible fault. Context lock preserves the earliest useful break in continuity.
Preserve the surrounding truth
Signals, sequence position, timing, operating mode, process context, and related dependencies are held together instead of scattered.
Make the handoff evidentiary
The operator no longer has to reproduce the problem from memory. Maintenance receives a preserved event record.
Original ghost traps
Counters, timers, and comparators were the bridge.
The 1983 method used deterministic field logic before modern AI language existed. Counters proved what happened. Timers proved when it happened. Comparators proved whether behavior matched the expected pattern. Modern Ghost Busting™ keeps the same discipline and gives it edge memory, broader communication, time alignment, and permanent institutional learning.
What happened?
Did the input, one-shot, transition, count, or companion event actually occur?
When did it happen?
Was the response late, early, drifting, settling, sticking, bouncing, or waiting too long?
Did reality match expectation?
Did the live circuit, state, station, or process match the expected pattern?
Evidence workflow
Observe → Detect → Context Lock → Correlate → Rank → Present Evidence → Verify Fix.
1
ObserveRead existing machine truth from PLC tags, sensors, drives, I/O, gateways, historians, HMI data, and process context.2
DetectFind dropouts, chatter, jitter, delay drift, response lag, sequence violations, and legal-but-wrong behavior.3
Context LockPreserve the machine step, recipe, mode, command, confirmation, time window, and related conditions.4
CorrelateConnect related signals, timers, counters, permissives, mechanisms, product conditions, and environmental factors.5
RankA.I.R.O.N. ranks likely contributors and shows which discipline should verify first.6
Present EvidenceThe team sees what executed, what failed to carry forward, and where continuity broke.7
Verify FixAfter repair or adjustment, the system watches whether the anomaly signature disappears.8
Preserve LearningEvidence bundles and CI records become institutional memory instead of fading tribal knowledge.Human authority
A.I.R.O.N. captures evidence. People decide.
Ghost Busting™ does not replace operators, technicians, engineers, supervisors, safety authority, or CI leadership. It makes humans faster to the truth and less likely to miss the ghost.
- observes machine truth
- compares expected vs. actual behavior
- locks context
- preserves evidence
- ranks likely contributors
- supports verification
- safety authority
- lockout/tagout responsibility
- repair and adjustment judgment
- process-change approval
- CI interpretation
- final operational decisions
Read-only first. Ghost Busting™ observes, compares, records, and alerts. It should not silently take over control, force outputs, bypass safety, or modify machine behavior unless explicitly engineered, validated, and authorized.
Universal industrial intelligence
Not a foundry trick. A discipline for any repeatable process.
Food / beverage / packaging
Micro-stops, photoeye chatter, capper delays, valve response drift, encoder slip, and recipe-transition trouble.
Utilities / air / cooling
Pump cavitation onset, fan imbalance, VFD nuisance trips, flow switch bounce, sensor drift, and load/weather correlation.
Foundry / high-energy
Intermittent interlocks, contactor chatter, heat soak behavior, cycle-start deviations, melt steps, and safety sequencing.
Discrete manufacturing
Clamp/confirm timing, pneumatic seal behavior, robot handshake delays, sensor bounce, and exact sequence-window proof.
Oil / chemical / batch
Valve stiction, flow instability, instrument dropout, subtle lag during phase transitions, viscosity shifts, and environmental effects.
Any recurring failure pattern
Any machine, process, cell, subsystem, utility, or repeat problem can become a Ghost Busting™ target when there is enough signal truth to preserve.
Prediction loop
Ghost Busting™ catches the whisper.
Ghost Busting™
Captures the deviation, proves continuity, locks context, preserves evidence, and ranks likely causes.
Predictive Breakpoint™
Projects when a repeating deviation is becoming important enough to investigate, notify, or act.
Fortune Teller™
Turns patterns, baselines, and history into forward-looking guidance for operations and maintenance.
Institutional memory: People may move on. Learning does not. The tools changed. The doctrine did not.
From hidden problem to proven path
Your machines ALWAYS whisper. With A.I.R.O.N.™ you’ll NEVER hear them scream.
Start with the recurring issue that wastes the most time, causes the most doubt, or never happens while the right person is watching.