Skip to content

Reading spectra and waveforms

KPIs tell you that something is changing. Spectra and waveforms help you work out what. You view them on a measurement point in Condition.

Available on Treon Make (Treon Industrial Node X)

Spectra and waveforms are captured by the Treon Industrial Node X sensor used in the Treon Make solution. Treon Flow machines (Treon Industrial Node C) are KPI-only and have no spectra or waveforms to open — see Overview of solutions.

Waveform — the time domain

A waveform is the raw vibration signal over a short slice of time: amplitude on the vertical axis, time on the horizontal. It shows the signal as it actually happened — useful for seeing impacts, modulation, and the overall character of the vibration. Waveforms are captured per axis (X, Y, Z).

A raw acceleration waveform: about one second of the X-axis signal, captured at a G-range of 4 G

The same waveform shown full-width, with the X, Y and Z axis traces available from the legend below the chart

Spectrum — the frequency domain

A spectrum (FFT) breaks the same signal into the frequencies it contains: amplitude on the vertical axis, frequency on the horizontal. This is the core diagnostic tool, because different faults show up at characteristic frequencies tied to the machine's running speed and geometry:

The Spectrum (FFT) tab for a Treon Industrial Node X measurement point — amplitude against frequency for acceleration, velocity and magnetic field, with the Harmonics overlay on the velocity spectrum

You open it on the measurement point's Spectrum (FFT) tab, choosing the measurement by date and timestamp. On a healthy point the spectrum stays low and flat, as above; a developing fault raises peaks at the frequencies below.

  • Imbalance — a peak at 1× the running speed.
  • Misalignment — often 1× and 2× running speed, sometimes axial.
  • Looseness — multiple harmonics of running speed.
  • Bearing defects — peaks at the bearing's specific defect frequencies, usually high-frequency and often accompanied by rising kurtosis before they dominate the spectrum.
  • Gear problems — peaks at gear-mesh frequency and its sidebands.

Reading which frequencies grew, relative to the machine's running speed, points to the mechanism — which is why a spectrum turns "velocity RMS is rising" into "the coupling is misaligned."

Magnetic field — for electric motors

Treon Industrial Node X also captures a magnetic field signal. On electric motors this adds a diagnostic dimension that vibration alone can miss: electrical faults (broken rotor bars, winding or supply problems) leave a signature in the magnetic field. It is read the same way as a waveform — as a signal you compare against a healthy reference.

A magnetic-field waveform from a Treon Industrial Node X on an electric motor: the X axis, cycling between roughly 838 and 890 µT

The trace above is a healthy motor: a clean, regular cycle at a steady amplitude. Distortion of that shape, or sidebands around it in the corresponding spectrum, is what points at an electrical fault. Magnetic field is also what the platform derives RPM from, given a pole-pair count — so it is what lets spectrum peaks be read against the real running speed.

Using them together

Start from the KPI trend that flagged the machine, then open the spectrum to localise the fault frequency, and the waveform to confirm its character (steady vs impulsive). Comparison lets you line the same view up against a healthy period or a sister machine.