Gen4 vibration sensor: longer raw captures, battery life and kurtosis/band RMS

Hi,

We’re evaluating the Gen4 vibration sensors for predictive maintenance on centrifugal pumps (~1,780 RPM, some on VFDs) in a below-grade plant room. Our data scientist has defined the data we need, and I’d appreciate guidance on a few points:

  1. Capture length: for bearing analysis we need about 1 s of contiguous raw data at 25.6 kS/s on one radial axis (~26,000 samples), and longer at low VFD speeds. The current maximum of 8,192 samples gives 0.32 s. Is there any way to capture longer, or to buffer back-to-back captures on the sensor with no gaps?

  2. Firmware calculations: would it be feasible to add kurtosis and band RMS (e.g. 1–6 kHz) to the processed data, through a firmware update or custom firmware?

  3. Battery life: what battery life would you expect with processed data every 15 minutes plus 8 raw captures a day (8,192 samples at 25.6 kS/s, single axis)? Alternatively, could you share the Gen4 current and duration values for processed and raw transmissions, so we can use your battery calculator?

Thanks,
Anthony

Hi Anthony,

  1. The max sample is capped at 8192 samples per axis. There is no way to increase it
  2. Yes, we are working on adding more parameters to processed data, and Kurtosis is one of them. We can add one as well if you can share a list of parameters you would like ot see
  3. Li-ion Primary battery Life Calculator - NCD.io
    Based on your data transmission frequency, the battery should last over a year. The sub-GHz radio is really good when it comes to long range and penetration, but it does consume a lot more power.

Great, thanks Bhaskar.

Kurtosis is one of the parameters we need. The other is band RMS with configurable band limits (e.g. 1–6 kHz) for bearing monitoring. If those two could be added, that would cover our processed data requirements. Do you have a rough timeframe for when these would be available in a firmware release? We’re planning a pilot and would like to know whether to expect them in time.

On battery life, understood. We’ll plan around annual changes and measure on evaluation units.

Could you also confirm whether we can trigger raw captures back-to-back in on-request mode, and roughly how long one 8,192-sample single-axis raw transmission takes?

Thanks again for the help.

I will check with the team and inform you of the live dates. When you request time-domain data, it will be sent for all three axes. You can only request it once per sample interval.