Speed-aware baselines
Each node's flux sensor reads motor speed, and the gateway pulls line speed and order changes from the PLC, so spectra are compared at like speeds and an order change never looks like a fault.
Example engagement
Thirty-eight drives, one corrugator, fewer surprises
When one gearbox or blower fails, this representative box plant's whole corrugator stops. Wireless Pulsenode W nodes and analyst-verified alerts under Care Verified would turn those breakdowns into planned work.
Example engagement — a representative scenario built from typical deployment parameters. Figures are modeled, not reported results from a named client.
Corrugator monitoring at a glance
- Industry
- Packaging & paper
- Services
- Hardware
- 38 Pulsenode W wireless nodes (vibration, surface temperature, magnetic flux)
- 2 Gatepoint G2 gateways
- Read-only OPC UA link to the corrugator PLC for speed and order changes
- Work orders to the plant's CMMS via API
- Timeline
- 3-week Feasibility Sprint, 6-week Proof of Value on 25 assets, 6-week rollout to all 38
Modeled results
~35%
less unplanned downtime on the monitored assets (modeled)
38
critical assets watched with no new cabling
4 h
maximum time to analyst verification of an alert, business days
The situation
One machine, eight sections, no spare corrugator
Picture a box plant built around a single corrugator: two single facers, a glue machine, a double backer with steam-heated hot plates, then the slitter-scorer, cutoff knife and stacker. The printers, die cutters and folder-gluers downstream all wait on the board it makes. When the corrugator stops, the plant stops.
The maintenance team is skilled and small. Route-based vibration readings are scheduled monthly and happen when someone has time. A vacuum blower on a single facer loses a bearing on a Saturday night; a double-backer gearbox starts to whine and keeps running until it doesn't. Each failure costs a few hours of scramble, and a day or more when the part isn't on the shelf.
The machine is also hard to monitor. Its speed changes with nearly every order, steam and starch put heat and humidity everywhere, and the bearings that matter most sit behind guards, in places nobody wants to reach on a running machine.
Constraints
Monitoring a machine that never holds still
- Speed and orders
- Line speed changes with nearly every order, with frequent splices and order changes on every shift
- Heat and dust
- Steam-heated rolls and hot plates, starch paste, paper dust and high humidity at the wet end
- Access
- Many bearings sit behind guards; nothing may require stopping the machine or opening guards to take a reading
- Radio
- Steel frames, roll stock and racking scatter radio signals, so gateway positions come from a site radio survey
- IT/OT rules
- Gateways on the plant's OT network; corrugator PLC data read-only over OPC UA; one service account in the CMMS and no other new users
- Shifts
- Three shifts, five days, with Saturday overtime in busy months; one maintenance planner on days and an on-call technician at weekends
Approach
From two years of downtime logs to 38 watched assets
-
Feasibility Sprint
For machines, the Sprint starts with records rather than samples: two years of downtime logs and work orders, the critical-spares list and a walkdown with the maintenance lead. Ranking assets by failure cost and detectability lands on 38. The steam-heated corrugating-roll journals stay out of phase one; they run hotter than a battery-powered node should sit on.
-
Proof of Value
Pulsenode W nodes go on the 25 highest-ranked assets, with a loaner gateway. Success criteria, agreed in writing: baselines established for each speed band, every alert verified by an analyst, and every verified alert either confirmed at inspection or explained.
-
Rollout
The remaining 13 nodes, a second Gatepoint G2 where the radio survey called for one, CMMS integration, and training for the maintenance technicians and the planner.
-
Care Verified
A Perflens analyst verifies each alert within 4 hours on business days before it becomes a work order. Models are retrained quarterly as the machine and its baselines change. See Perflens Care.
What the system does
What 38 nodes and an analyst actually do
Named faults, not vibration levels
Imbalance at 1× running speed, misalignment at 2× with axial vibration, looseness harmonics, bearing defect frequencies in the enveloped spectrum and gear-mesh sidebands. Each alert names the likely fault.
Verified before it becomes work
Alert floods end monitoring programs; a technician who walks to three false alarms stops walking. Under Care Verified, an analyst reviews the spectrum and trend before anyone is sent to the machine.
Work orders where the team already works
Confirmed findings open a work order in the plant's CMMS with the asset, fault, severity and a recommended action window.
Battery life planned, not discovered
Hourly sampling gives battery life of up to 5 years. The eight most critical assets sample more often, so their battery swaps go into planned shutdowns.
No cable through a hot machine
IP68 wireless nodes mount on bonded pads in minutes, with no cable to pull through guards or over steam lines.
Coverage
The 38 monitored assets
| Asset group | Count | Faults watched for |
|---|---|---|
| Single-facer drives and gearboxes | 6 | Gear-mesh wear, misalignment, bearing defects |
| Single-facer vacuum blowers | 4 | Imbalance from dust build-up, looseness, bearing defects |
| Glue-machine and preheater drives | 4 | Bearing wear, misalignment |
| Double-backer drive train | 4 | Gearbox wear, belt and coupling problems, motor bearing defects |
| Slitter-scorer and cutoff-knife drives | 6 | Bearing defects, gear-mesh wear, impacting under load |
| Stacker and dry-end conveyor drives | 4 | Bearing defects, misalignment |
| Trim-system fans | 3 | Imbalance from fiber build-up, bearing defects |
| Steam, condensate and starch pumps | 7 | Cavitation, bearing defects, misalignment |
Modeled results
Modeled effect on unplanned downtime
~35%
less unplanned downtime on the 38 monitored assets
38
assets monitored without new cabling or a production stop
4 h
maximum time to analyst verification on business days, under Care Verified
Modeled on a downtime history typical of a single-corrugator box plant, where bearing, gearbox and blower failures make up most unplanned stops on these assets. Assumes most such failures give days to weeks of detectable warning and that the team acts within the recommended window. Failures with no vibration precursor, such as electrical trips and belt breaks from foreign objects, continue as before. Assets outside the 38 are not included.
Phase two
Phase two on the corrugator
- The corrugating-roll journals. They were excluded because of heat. A wired sensor arrangement designed for those temperatures is the first addition.
- Motor current. Pulsenode C in the motor control center on the largest drives adds current signature and harmonics to the vibration picture, separating electrical faults from mechanical ones.
- Steam traps. Pulsenode U on the traps serving the hot plates and preheaters finds traps failed open or shut, which is a steam-cost problem and a board-quality problem at once.
- Stops by cause. The PLC link already carries speed and order data; OEE tracking would turn it into downtime by cause and shift.
Related services
What sits behind this scenario
Predictive Maintenance & Condition Monitoring
Vibration, ultrasound, temperature and motor-current monitoring with fault diagnostics and human-verified alerts.
Read more
Perflens Care
The analyst desk, verification times and retraining schedule that decide whether alerts get trusted.
Read more
Packaging, Paper & Wood Products
Inspection and machine health for corrugators, converting lines and continuous paper processes.
Read moreFAQ
Questions we hear often
What happens to alerts at night and on weekends?
Care Verified analysts work business days. Severe alarms are not held for them; they go straight to the on-call technician by SMS, and the analyst reviews them the next business morning. Care Dedicated adds verification within 1 hour, 24/7.
What would 38 assets cost at published prices?
The asset monitoring subscription starts from $89 per asset per month (minimum 25 assets, 36-month term), including Pulsenodes, gateway, Console and Care Essentials. Care Verified adds $18 per asset per month, so 38 assets start from $4,066 per month. See pricing.
Start with the assets that stop the line
Bring two years of downtime records to a discovery call. We'll tell you which assets are worth monitoring, which aren't, and what verified alerts would cost.