AquaMeshAll case studies
Case study · MechTrack

Know which machine will fail before it does

MechTrack is a wireless sensor that listens to a machine three ways: vibration, ultrasound, and temperature. It mounts on the machines that keep a plant running and warns the maintenance team days ahead.

Illustrative scenario for a precision manufacturer: metal finishing, machining, coating, and heat treat. Modeled from industry benchmarks, not a verified customer result.

MechTrackSpindle bearing wearUltrasound and vibration rising at the same speed and load.
MechTrack sensor on every machine
Drag to rotate · scroll to zoom
The sensor

Three ways to hear a machine

MechTrack sensor
Vibration

Bearing wear, imbalance, misalignment, looseness.

Ultrasound

Early friction and poor lubrication, before vibration shows it.

Temperature

Overheating and cooling failures.

Mounts with a magnet or stud. Learns each machine's normal, then flags what changes.

Specs at a glance

Sensor and gateway

MechTrack sensor

MeasuresVibration, ultrasound, temperature, RPM, run time
VibrationUp to 25.6 kHz sampling, ±16 g or ±64 g
Ultrasound30 to 120 dBµV, sampled up to 82 kHz
Temperature±1 °C accuracy
RangeUp to 2 miles line of sight, about 1,000 ft indoors
Battery3 to 10 years
Built forIP67, magnet or stud mount, -40 °C to 70 °C

Site gateway

RoleReceives every sensor at a site; runs dashboards, alerts, and reports
CapacityOne gateway serves all sensors in range
Connectivity2 Ethernet ports, dual-band Wi-Fi, cellular option
IntegrationsMQTT, REST API, AWS, Azure, Node-RED
DataStays on your network

The full spec sheet is provided separately.

Engineering detail

What each signal tells you

FaultWhere it showsSignal
Unbalance1x running speed, radialVibration
Misalignment1x and 2x, often axial; heat at the couplingVibration, temperature
LoosenessHarmonics of running speedVibration
Bearing wearBearing defect frequencies, then a broadband riseUltrasound, then vibration, then temperature
Lubrication lossRising ultrasound amplitudeUltrasound
Gear wearGear mesh frequency with sidebandsVibration
Overload, cooling lossSustained temperature riseTemperature

Fault classes match the DOE FEMP O&M guide: vibration (Section 6.5.3), ultrasound (6.4), bearing temperature (6.2). Frequency locations follow standard vibration analysis practice.

From signal to alert

SampleVibration up to 25.6 kHz. Ultrasound up to 82 kHz. Temperature to ±1 °C. RPM and run time logged.
ExtractAcceleration, velocity, displacement, peak frequency. Raw time data kept for FFT.
LearnA baseline per machine and operating state. On-board machine learning flags abnormal vibration.
Alert900 MHz mesh to the gateway, 128-bit AES. Email alerts and reports. MQTT and REST out.

At 82 kHz sampling, ultrasound bandwidth reaches about 41 kHz, which covers the 28 to 32 kHz band DOE cites for early bearing warning.

Where it goes

If it spins or runs hot, it can go on it

One sensor for motors, spindles, pumps, fans, gearboxes, and compressors, across the whole plant.

Machining

  • CNC spindles
  • Grinders
  • Honing and polishing
  • Lathes and mills

Finishing and plating

  • Circulation and filtration pumps
  • Transfer hoists
  • Tumblers and vibratory finishers
  • Blast and shot-peen blowers
  • Rectifier cooling fans

Heat treat and forming

  • Vacuum pumps
  • Furnace blowers and quench pumps
  • Hydraulic presses and power units
  • Forge hydraulics and cooling fans

Plant utilities

  • Air compressors
  • Chillers and cooling towers
  • Dust collectors
  • Exhaust fans and scrubbers
  • Make-up air units

Wastewater

  • Process and transfer pumps
  • Mixers and agitators
  • Filter press drives
  • Blowers and aerators

Material handling

  • Conveyors and gearmotors
  • Overhead cranes and hoists
  • Robot and automation drives
  • Packaging and test equipment
Example

A spindle bearing, caught early

Signal

Ultrasound picks up added friction first. Days later, vibration at the bearing frequency starts to climb, at the same speed and load.

Finding

A developing bearing fault, well before any machine alarm.

Action

The maintenance lead approves a swap in the weekend changeover. No scrap, no emergency.

Anticipated results

What it is built to deliver

30 to 50%less unplanned downtime on monitored assets
10 to 25%lower maintenance cost on monitored assets

Targets are in line with the DOE benchmarks further down, and are confirmed against a customer baseline.

DOE benchmarks

What the published data says

25 to 30%lower maintenance costs
70 to 75%fewer breakdowns
35 to 45%less downtime

Industry average savings for predictive maintenance, as reported in the DOE FEMP O&M Best Practices Guide (Section 5.4). The guide attributes them to independent surveys and does not name the study.

Annual maintenance cost per horsepower

Reactive$18
Preventive$13
Predictive$9
Reliability centered$6

Comparison of four maintenance programs for pumps (Piotrowski 2001), as cited in the DOE FEMP guide, Sections 5.6 to 5.9.

Why rotating equipment

Machine-driven processes, including pumps, fans, compressed air, and materials handling, used 68% of U.S. manufacturing electricity in 2010 (DOE Advanced Manufacturing Office).

A temperature case from DOE

An infrared scan showed one pump's lower thrust bearing running hotter than the other. The pump was mounted vertically, which overloaded that bearing. Repeated failures cost $15,000 in repairs, before lost production (DOE FEMP guide, Section 6.2.6).

These are industry averages. We confirm them against your own baseline.

Sources: DOE FEMP O&M Best Practices Guide, Release 3.0, Chapter 5 · Chapter 6 · DOE Motor System Market Assessment

Want MechTrack on your machines?

Tell us your critical machines. We will show where MechTrack goes and what it should catch.