Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Action detect with accelerometer

Not open for further replies.


Junior Member level 2
Sep 18, 2013
Reaction score
Trophy points
Activity points
Hello guys,
I have a electronic and I want know when it's moved.
I did simple algorithm to eliminate natural error in accelerometer, but I´m problem with vibration.
If I put electronic in my hand, my natural vibration is read how action.
I only want detect when my electronic is displaced.
Anybody have idea for algorithm?

I use ADXL 345.

tnx very much

You need some filtering. Without more information, it's hard to offer more advice. For instance, what do you mean by your "natural vibration"? I, for one, do not vibrate. What type of motion do you want to detect? Amplitude? Frequency?

Hello Barry,

tnx very much for your help.
For instance, what do you mean by your "natural vibration
Are small movements. When you take a deep breath, maybe you move your body and accelerometer cath it.

I made a simple filter for eliminate natural variantion from accerometer, but when I knock on the table, my algorithm fail, because I want eliminate it.
My algorithm only calculates averages and calculates the variation of new measurements, excluding read errors.

It is used in a car repair shop. I only want have signal when worker move the electronics side to side of the room.

Tomorrow I will be implement a other filter to try to eliminate cases as vibrations on the table where this electronic positioned.
But I no have good idea for make it yet.

You need to determine what type of "signature" your signal of interest has and how it differs from signals you DON'T want to see. Does it have a particular duration? Amplitude? Frequency? Envelope? You may need more than just a filter. Maybe you need to do an FFT, or an integration or a double-integration.
Equipments that monitor body electric signals (e.g ECG) in general has band stop filters in the range of 35-40Hz, which is the typical frequency bandwidth of small muscular movements.
Not open for further replies.

Part and Inventory Search

Welcome to