Ask the experts

Understanding noise specifications


Meggitt datasheets for Endevco® ISOTRON® accelerometers consistently show noise floor or resolution units in terms of "equiv. g rms." Can you explain what that means?


"Equiv. g rms" simply means "equivalent g root mean squared." As an electronic instrument, an ISOTRON (IEPE-type) accelerometer outputs a voltage signal that is typically measured in millivolts (mV). If accelerometer sensitivity is known, for example, 100 mV/g, then we can interpret the mV signal in equivalent g units. Similarly, when specifying noise floor, rather than using units of mV, we use units of g.

So, why do datasheets express noise floor in terms of rms? It has to do with the random, probabilistic nature of electronic noise itself. (Please note that we are referring to the intrinsic electronic noise coming from the device itself, which can be specified, and not extrinsic electronic noise from outside sources, which cannot be specified.) By indicating physical quantities in rms units for random phenomenon like noise, users can utilize mathematical statistical tools for data collection. One direct result of this idea, for example, is the creation of a good rule of thumb for electronic measurements: given a voltage signal, do not try to resolve a peak value (i.e. mV pk) less than three times that of the rms noise floor. So, for example, if an ISOTRON® accelerometer has a specified noise floor (or resolution) of 0.02 equiv. g rms, it would be poor engineering practice to attempt to measure an amplitude of less than 0.06 g pk.

A full explanation of the concept of this idea and electronic noise issues are beyond the scope of this article. It is recommended that customers consult an electronics textbook on noise for a deeper understanding of these issues.