Signal

Who Signal is for

For electrical and mechanical engineers with a sensor capture and a suspicion. Bearing wear, harmonic distortion and mechanical resonance are frequently invisible in the time domain and unmistakable in the frequency domain.

What Signal does

Load a CSV capture and Signal gives you the spectrum: FFT and power spectral density, band-pass and band-stop filtering to isolate a region of interest, and a spectrogram when the frequency content changes over time. The workflow it is built for is the real one — look at the spectrum, notice something at a particular frequency, filter around it, and confirm. It pairs with Simulate, which generates faults whose ground truth you already know so you can validate the analysis before trusting it on real data.

• →FFT and power spectral density
• →Band-pass and band-stop filtering
• →Spectrogram view for time-varying content
• →Sensor CSV input

How Signal works

1. 01

   Load the capture

   Sensor data as CSV, from whatever acquisition setup you use.

2. 02

   Look at the spectrum

   FFT and power spectral density. This is where a fault signature usually announces itself.

3. 03

   Isolate the band

   Band-pass or band-stop to remove everything that is not the thing you are chasing.

4. 04

   Check the spectrogram

   When the frequency content changes over time, a static spectrum averages away the event you want.

What Signal does not do

Signal analyses captures you supply — there is no data acquisition, no instrument connection and no live streaming, so it works after the fact. Input is CSV, so proprietary acquisition formats need converting first. It shows you the spectrum and does not diagnose: identifying a peak as a bearing fault rather than a harmonic is engineering judgement, and the module deliberately does not pretend otherwise.

Common questions

Works well with