[Microbit-Python] Accelerometer orientation

[M.-A. Lemburg]

On 18.02.2016 19:56, M.-A. Lemburg wrote:
> On 18.02.2016 18:33, Radomir Dopieralski wrote:
>> On Thu, 18 Feb 2016 15:13:20 +0100
>> "M.-A. Lemburg" <mal at egenix.com> wrote:
>>
>>> On 18.02.2016 13:04, Loop Space wrote:  
>> [snip]
>>
>>
>>> The values of the meter go from -1024 to +1024.  
>>
>> Actually no, this is not true. The measurements are not normalized,
>> they just give you the acceleration measured in 1/1000th of a g. Thus,
>> with it lying flat and not moving you get 1g down.
>>
>> See https://en.wikipedia.org/wiki/Gravity_of_Earth
> 
> I'll have to recheck. When I did the experiments with the MB some
> weeks ago, those were the values I found (and it was not moving).

I ran another series of tests. +/-1024 is still the
figure I see when holding the MB vertically in x- and
y-direction (perhaps just a coincident).

-1024 is also the typical z-axis reading I get when holding
the MB horizontally with the LED panel up, 1024 when
turning the MB upside down (with the BBC logo up).

However, by shaking the MB it does produce much higher
values in both directions, corresponding to the g-force
measurements you mentioned.

>>> Now, let's assume you are holding the MB horizontally,
>>> with the LED panel up and the pins pointing to you.
>>>
>>> For the x-axis, it's +1024, if you hold the MB vertically with
>>> the B button pointing down and -1024 if you hold it in the
>>> other direction, with button A down. It's 0 if you hold the MB
>>> horizontally.  
>>
>> It's actually +1000 and -1000. Provided it doesn't otherwise move.
>>  
>>> For the y-axis, you get -1024 if you turn the MB up, so that
>>> the pins point to the sky, 0 when it's held horizontally and
>>> +1024, when turning the MB so that the pins point down.  
>>
>> Same here.
>>
>>> The "accelerate" part in accelerometer doesn't seem to
>>> have much meaning for x- and y-axis, i.e. they don't seem to
>>> measure acceleration, but instead orientation. Only the z-axis
>>> appears to actually work in terms of acceleration, since it
>>> changes its value depending on how you whether the MB up and
>>> and down.  
>>
>> Of course they measure acceleration. This is the whole point. The thing
>> is, gravity is impossible to tell from acceleration (Einstein even
>> claims it's pretty much the same thing), so you end up also measuring
>> gravity. Which can be used to tell the orientation while you are within
>> the gravity field of Earth (and not in a free fall).
> 
> Well, I know the term "acceleration" as referring to a change
> in velocity (first derivative of the latter). If a mass doesn't
> move, velocity is zero and thus acceleration is zero too.
> 
> But I see your point: the accelerometer is not measuring
> acceleration, it's measuring g-force and then the readings
> make sense again. My misunderstanding :-)
> 
>> All axes actually work in terms of acceleration -- you can see that by
>> shaking the microbit sideways. And you can see you can get much higher
>> values than 1024.
>>
>> Maybe we could even make a game that records the highest acceleration.
>> Then again, maybe not, because then kids would throw it against the
>> wall...
> 
> I'll try that ... uhm, better not :-)
> 
>>> I haven't found a good use for the z-axis yet.  
>>
>> All the axes are equally useful. You can use atan2 function to get
>> actual angle from them.
> 
> Right... giving me the orientation in space when the MB is
> not moving.
> 
> Anyways, it does work nicely for x- and y-axis:
> 
> https://www.youtube.com/watch?v=PwbfHcnkmNs
> 

-- 
Marc-Andre Lemburg
eGenix.com

Professional Python Services directly from the Experts (#1, Feb 19 2016)
>>> Python Projects, Coaching and Consulting ...  http://www.egenix.com/
>>> Python Database Interfaces ...           http://products.egenix.com/
>>> Plone/Zope Database Interfaces ...           http://zope.egenix.com/
________________________________________________________________________
2016-02-19: Released eGenix PyRun 2.1.2 ...       http://egenix.com/go88

::: We implement business ideas - efficiently in both time and costs :::

   eGenix.com Software, Skills and Services GmbH  Pastor-Loeh-Str.48
    D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg
           Registered at Amtsgericht Duesseldorf: HRB 46611
               http://www.egenix.com/company/contact/
                      http://www.malemburg.com/