Background
Thermastor has been designing and manufacturing HVAC and dehumidification equipment for over four decades. Their products handle demanding environments — crawl space moisture control, commercial building dehumidification, HVAC for structures requiring precise humidity management. The company sells into applications where equipment reliability directly affects building health and occupant comfort.
Thermastor understood the IoT monitoring value proposition for their customers. They built it into their product roadmap as a feature direction — connected monitoring capability that would allow end users to track equipment performance, detect degradation, and schedule maintenance proactively.
What created the initial Viaanix engagement was a more immediate problem: Thermastor’s own manufacturing floor.
Their production environment included stamping presses, coil forming equipment, welding stations, powder coat lines, and final assembly areas. Like most established manufacturers, the equipment had accumulated over decades — different vintages, different control systems, no unified monitoring layer.
Equipment failures were discovered when production stopped. Maintenance was calendar-based — service intervals set by equipment manuals rather than by actual condition. The maintenance team spent significant time on reactive repair rather than condition-based service.
The Challenge
Reactive Maintenance Across an Aging Equipment Mix
Thermastor’s production floor equipment ranged from 3-year-old CNC benders to 20-year-old stamping presses with relay-logic controllers that predate digital communication interfaces. Creating a monitoring layer across this mix required connecting both modern equipment with digital output and legacy equipment with no native connectivity.
The maintenance team’s recollection of recent equipment history was the primary diagnostic tool. When a press stopped, the lead maintenance technician’s first question was: “What was it doing before it stopped?” Without instrumented data, the answer was either the operator’s description or an educated guess.
No Leading Indicators for Equipment Health
Two specific failure modes had caused significant downtime in the previous 18 months:
Compressor failures in the test chamber HVAC units: Thermastor tests completed dehumidifier units in environmental test chambers before shipping. The chambers run HVAC equipment continuously. Two compressor failures in 14 months had resulted in chamber downtime totaling 6 days — plus the equipment damage and emergency repair costs.
Coil winding machine bearing failures: The automated coil winding machines had experienced 3 bearing failures over 18 months — each resulting in 4–8 hours of production downtime. The failures gave no obvious warning: the machine was running normally one shift, then stopped with a failed bearing the next.
Both failure types have detectable precursor signatures — elevated compressor current and temperature for the HVAC units, vibration signature changes for the bearings. Neither was being monitored.
The Solution
Phase 1: Test Chamber and Critical Equipment Monitoring
The first deployment phase focused on the equipment with the most costly recent failure history:
Test chamber HVAC compressors: Current transformers installed on each compressor motor feeder circuit, connected to VX-Olympus via industrial IoT gateways. Temperature sensors on compressor housings and discharge lines. VX-Olympus established a 30-day baseline for each compressor — average current draw at standard operating conditions.
Alert rules:
- Current exceeding baseline by 15% for 30+ minutes: warning alert to maintenance lead
- Current exceeding baseline by 25% for 15+ minutes: critical alert
- Compressor housing temperature above 180°F: immediate alert
Coil winding machines: Vibration sensors mounted on motor bearing housings and winding head housings. VX-Olympus Rule chains evaluated RMS vibration against a rolling 30-day baseline per machine.
Alert rules:
- Vibration exceeding baseline by 20% (sustained): watch alert — begin monitoring interval
- Vibration exceeding baseline by 35% (sustained 10+ minutes): warning alert — schedule inspection
- Vibration exceeding absolute threshold (equipment-specific per ISO 10816): critical alert — stop and inspect
Phase 2: Production Floor Equipment and Line Visibility
VX-Olympus connected to stamping presses and assembly stations through a combination of:
- Discrete I/O interfaces: Running/stopped/fault contacts from relay panels
- Current sensing: CT clamps on motor feeders for equipment without digital outputs
- OPC-UA: Newer equipment with Siemens PLCs exposed register data via OPC-UA
The floor manager dashboard showed all production zones with color-coded equipment status. Active faults surfaced immediately; equipment trending toward concern (elevated current or vibration) appeared in amber.
Maintenance work orders were generated from VX-Olympus alert events — either automatically for defined conditions or by maintenance staff converting an alert into a work order through the platform interface. Each work order carried the equipment ID, the alert that triggered it, and the relevant recent telemetry.
The Results
Test Chamber Compressor: One Failure Caught Before It Failed
Seven months after phase 1 deployment, a VX-Olympus current monitoring alert flagged one of the test chamber compressors drawing 22% above its baseline for 45 consecutive minutes during normal operating conditions. The vibration signature was also elevated but below the bearing failure threshold.
The maintenance team pulled the compressor for inspection. They found:
- A refrigerant charge below specification (causing the compressor to work harder)
- The beginning of internal valve wear (likely caused by the low charge condition)
The compressor was recharged and the valves were replaced during a planned weekend maintenance window. The repair cost: approximately $2,800.
The alternative, had the condition gone undetected: a complete compressor failure, emergency replacement, and test chamber downtime. The maintenance manager estimated the avoided cost at $12,000–$18,000.
Coil Winding Bearings: Pattern Identified
The vibration monitoring on coil winding machines identified a pattern the maintenance team had not recognized: bearing failures were preceded by a vibration elevation detectable 3–5 days before failure. The specific frequency signature of bearing outer race wear appeared consistently before each failure in the historical analysis of the pre-deployment period (performed retrospectively after understanding the sensor data patterns).
With continuous vibration monitoring, the maintenance team could schedule bearing replacement based on vibration signature evolution rather than calendar intervals — replacing bearings when their condition indicated they needed replacement, not when the calendar said to check them.
Impact: In the 12 months following deployment, no unplanned coil winding machine failures occurred. Three bearings were replaced proactively based on vibration pattern alerts — all during scheduled weekend maintenance windows. Production was not interrupted by any coil winder bearing failures.
Maintenance Labor Shift
Six months post-deployment, Thermastor’s maintenance manager reviewed the maintenance log and estimated:
- Reactive maintenance labor (responding to failures): dropped from ~65% to ~35% of maintenance hours
- Proactive maintenance (condition-based service, planned interventions): increased from ~35% to ~65% of maintenance hours
- Average maintenance call response time (from equipment stop to maintenance team on-site): unchanged (2–4 minutes for production stops) but the information available at the start of each call improved significantly — fault codes and recent telemetry in the VX-Olympus alert detail
An Unexpected Outcome: The Connected Product Reference
During the VX-Olympus deployment, Thermastor’s engineering team observed the monitoring layer they had built for their own operations. The question arose naturally: could a similar layer be built into their customer-facing connected product offering?
The answer was yes — and VX-Olympus’s white-label capability provided the architecture.
Thermastor began developing a connected monitoring portal for their dehumidifier product line: the same sensor types (temperature, current, vibration) installed in their manufacturing floor monitoring applied to field-deployed units, with VX-Olympus white-labeled as “Thermastor Connect.”
The customer-facing product — still in development at the time of this case study — would give Thermastor’s customers the same proactive maintenance capability that Thermastor had implemented for their own manufacturing operation.
Conclusion
Thermastor’s VX-Olympus deployment addressed a specific and costly manufacturing problem: equipment failures that could be detected in advance but weren’t, because the monitoring layer was missing.
The financial case was straightforward. The avoided test chamber failure in month 7 alone recovered the majority of the platform deployment cost. The coil winding bearing program eliminated the 3–4 annual bearing failure incidents that had been causing 4–8 hours of production downtime each.
The operational case was the more lasting value: a maintenance team that now has condition data behind every decision, and a production floor where equipment problems surface as amber indicators before they become red stops.
The product strategy value — using the deployment as a reference for Thermastor’s own connected product offering — was not anticipated when the project started. It emerged from the team’s direct experience with what continuous monitoring made possible.
Talk to our team about a condition monitoring deployment for your manufacturing environment.