Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation.

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation.

1,379 Views
mattconnolly
Contributor I

The MMA7361 appears to act as a inclinometer while at a standstill.  My application will have unpredictable angles of tilt applied to the sensor.  This then affects the analog output because the MMA7361 incorporates both the acceleration value and the "tilt" value in it's final output.  Anyone have a suggestion that might kick-start a solution?

Labels (1)
0 Kudos
5 Replies

932 Views
Joshevelle
Senior Contributor I

Hello Matt,

The MMA7361 is a three axis accelerometer (X, Y and Z). So you're basically measuring three acceleration vectors. In order for you to measure the total acceleration, you would need to calculate the resultant of adding those three vectors.

if you _only_ need the magnitude of the total acceleration vector and not the direction of it, then it would be as easy as:

A= sqrt(x^2 + y^2 + z^2);

if you need the direction of that vector, then is a whole different story. Let me know if you need that.

Hope it helps.

Josh

0 Kudos

932 Views
mattconnolly
Contributor I

Josh,   First, thank you for your time.  My goal is to measure the acceleration of the y axis.  However, the sensor will be subject to random angles along the y axis possibly as great as say +-15 deg.  My thoughts are that the z axis can be used to correct for the angle error.  The z axis should deliver it's greatest voltage at the point where the y axis is "level" - in this case I am assuming the sensor is at standstill.  But when the y axis begins to tilt doesn't the z value decrease by the same amount regardless if the angle of y goes positive or negative?  Hope you can follow my rambling on that one.

If the above is true, then the actual state of the sensor is unknown (I think).  Is it really in motion and experiencing acceleration or is it just tilted or possibly some combination of the two.  I believe this can be overcome by intentionally inclining the sensor along the y axis to some known fixed angle from the start, say 20 deg.  This then would allow the sensor to tilt it's +-15 deg. but the z axis value would have a definite meaning as to positive or negative tilt.  Then some math applied to the y axis and z axis values should deliver the true acceleration of the sensor irrespective of any random angle.  I have not yet even considered how to solve that. 

Now that you have come this far - I would appreciate any type of feedback you may offer.  Even if it says that I am completely missing the mark.  The sensitive nature of this project requires that I am somewhat vague - would be much easier if you knew the very simple goal at hand. 

Finally,,,, my ultimate goal would be to do all of this in the analog world, but may have to resort to a microcontroller.  

Thanks again - and be well.Matt Connolly

  From: Joshevelle <admin@community.freescale.com>

To: Matt Connolly <matt_cnnlly@yahoo.com>

Sent: Wednesday, January 14, 2015 1:38 PM

Subject: Re: - Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation.

|

|

Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation.

reply from Joshevelle in Sensors - View the full discussionHello Matt, The MMA7361 is a three axis accelerometer (X, Y and Z). So you're basically measuring three acceleration vectors. In order for you to measure the total acceleration, you would need to calculate the resultant of adding those three vectors.  if you only need the magnitude of the total acceleration vector and not the direction of it, then it would be as easy as: A= sqrt(x2 + y2 + z^2); if you need the direction of that vector, then is a whole different story. Let me know if you need that.  Hope it helps.Josh

Did your question get answered? If so, say thanks by clicking Correct Answer in the community thread!

Reply to this message by replying to this email, or go to the message on Freescale Community

Start a new discussion in Sensors by email or at Freescale Community

Following Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation. in these streams: Inbox

0 Kudos

932 Views
Joshevelle
Senior Contributor I

Hello Matt,

You're very welcome.

Calculate the acceleration requires basic understanding of vectors. You're right about Z axis decreasing while the sensor is rotating and, as you noticed, that is because of the tilt angle but is not linear and it won't decrease the same amount of the X or Z axis, it would totally depend of the magnitude and direction of the total acceleration vector.

I just did a quick search for a tutorial and I would highly recommend you to read the following Instructable written by Gadget Gangster:

http://www.instructables.com/id/Accelerometer-Gyro-Tutorial/step1/The-Accelerometer/

I would recommend you to use at least a small microcontroller rather than just analog devices. A single uC can save you time (money), PCB space (money), smaller BOM (money) since it can make the whole maths involved in this application, otherwise you would require a bunch of operational amplifiers and comparators, that you later need to adjust with trimpots (expensive trimpots).

-Josh

0 Kudos

932 Views
mattconnolly
Contributor I

Thank you Josh.  Most kind of you to take time and give energy to my little problem.  Maybe you will see the commercial product very soon - it should be BIG. 

Best regards and be well.Matt

  From: Joshevelle <admin@community.freescale.com>

To: Matt Connolly <matt_cnnlly@yahoo.com>

Sent: Monday, January 19, 2015 2:34 PM

Subject: Re: - Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation.

|

|

Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation.

reply from Joshevelle in Sensors - View the full discussionHello Matt, You're very welcome.  Calculate the acceleration requires basic understanding of vectors. You're right about Z axis decreasing while the sensor is rotating and, as you noticed, that is because of the tilt angle but is not linear and it won't decrease the same amount of the X or Z axis, it would totally depend of the magnitude and direction of the total acceleration vector.  I just did a quick search for a tutorial and I would highly recommend you to read the following Instructable written by Gadget Gangster: http://www.instructables.com/id/Accelerometer-Gyro-Tutorial/step1/The-Accelerometer/ I would recommend you to use at least a small microcontroller rather than just analog devices. A single uC can save you time (money), PCB space (money), smaller BOM (money) since it can make the whole maths involved in this application, otherwise you would require a bunch of operational amplifiers and comparators, that you later need to adjust with trimpots (expensive trimpots).

 -Josh

Did your question get answered? If so, say thanks by clicking Correct Answer in the community thread!

Reply to this message by replying to this email, or go to the message on Freescale Community

Start a new discussion in Sensors by email or at Freescale Community

Following Is it possible to eliminate the effect of tilt angle while using the MMA7361? I am seeking an analog acceleration value that remains predictable under variable sensor orientation. in these streams: Inbox

0 Kudos

932 Views
Joshevelle
Senior Contributor I

You're very welcome.

Once you have the finished product, post it here so we can know what is this about.

-Josh

0 Kudos