Accelerometer filtering suggestions

Discussion in 'General Electronics Chat' started by ProudlyCanuck, Nov 4, 2013.

  1. ProudlyCanuck

    Thread Starter New Member

    Nov 4, 2013
    Hi there,

    Engineering student here, sorry if the question seems obvious.

    I am trying to get the best accuracy I can out of an ADXL377 accelerometer. The data sheet suggests passive filtering, but would it be possible to get better accuracy if i implemented a few Sallen-Key low pass filters as well? I have listed some information about the ADXL377 below.

    Thanks in advance!

    I have attached the datasheet, but is a summary of the noise characteristics so you don't have to go digging through it:

    White Gaussian Noise, contributes equally at all frequencies
    Noise Density: 2.7 mg/sqrt(hz)

    They recommend the following caps to limit bandwidth at the pins:

    Cap(uF) Bandwidth (Hz)
    0.1 50
    0.05 100
    0.025 200
    0.01 500
    0.005 1000

    with a minimum cap of 1000 pF.
  2. MikeML

    AAC Fanatic!

    Oct 2, 2009
    Do you have any idea what the spectrum from the accelerometer looks like?
  3. crutschow


    Mar 14, 2008
    As MikeML noted, the frequency of the accelerometer signal is the main factor in determining the amount and type of filtering you need. What is the maximum frequency of interest?
  4. GopherT

    AAC Fanatic!

    Nov 23, 2012
    Filtering will also slow the response time (time constant) to stabilize a new reading at a new position.