# How to derive an average output from a triaxial accelerometer

1. Feb 8, 2012

### TS@QS

Hi
I am measuring the acceleration of an oscillating body with a triaxial accelerometer. I want to derive some kind of vector average of the 3 channel output. Someone suggested using
x= √(a squared + b squared + c squared)
... where x is the vector average and a, b and c are the accelerometry readings for each channel. Someone else told me that this isn't ideal but I don't really understand why not. Please help - is it better to just take the mean of all 3... or even just to choose the largest amplitude channel and use that?
Thanks

2. Feb 8, 2012

### MRFMengineer

I think your method is just fine, kind of depends on what your needs are. You lose the "sign" of the acceleration with that method, just results in a magnitude. You could also look at the min and max values, or min and max amplitudes on each channel.

3. Feb 8, 2012

### TS@QS

thanks - well , i'm using the acceleration as a surrogate for amplitude of displacement. I want to compare the oscillatory movement before and after applying a damping system to decide if there is an 'average' damping effect overall.

4. Feb 9, 2012

### MRFMengineer

Since you are just looking for a relative change, looking at the highest amplitude channel probably works fine. Acquire all the data you can, then mine it for whatever looks most useful.

5. Feb 9, 2012

### TS@QS

thanks, I'll try that.
I was just wondering why i was told that there are better more sophisticated methods and what these could feasibly entail?

6. Feb 9, 2012

### MRFMengineer

You could certainly get into looking at impulse and frequency response for different damping. Vibrational response is a whole field in and of its own.