Shop OBEX P1 Docs P2 Docs Learn Events
accelerometer issues and questions (Parallax/Memsic devices) — Parallax Forums

accelerometer issues and questions (Parallax/Memsic devices)

ArchiverArchiver Posts: 46,084
edited 2003-04-24 21:49 in General Discussion
I have been experimenting withe the Memsic 2125 accelerometer Demo kit from
Parallax and have noticed some quirks.
1.) The demo in the 2003 catalog uses a divisor of 16 supposedly to scale the
xmG and ymG values so that the ATN function will work. (it only works for
byte size variables ...not word size). But the example in the app note that
is shipped with the demo kit uses a divisor of 8. Why the difference? Which
is better or would it be better to develop a 16 bit ATN function? Would this
improve the accurracy or address the issue in part 2 below (certain angles
produce screwy results)?
2.) I am using the code for sensing 0-360 degrees of rotation, but the value
"blows up" in certain positions (especially around +/- 90 degress or so. Does
anyone know why this happens? Otherwise the ouput is fairly accurate and
repeatable.

Does anyone have experience with this type of device? Any help, suggestions,
or answers to my questions would be appreciated. Thanks!


[noparse][[/noparse]Non-text portions of this message have been removed]

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2003-04-23 23:41
    I think what you're describing as a "blowing up" of data in certain
    positions comes from the nature of the sensor. I'm not familiar with this
    particular sensor, but in general the piezoelectric material on each axis is
    most sensitive parallel to the horizon. Because of this, the resolution
    declines and the change per degree of tilt goes from a negligible amount to
    more than 10 times what it was per degree of tilt at +/- 1 g. You may want
    to check the data sheet and adjust the raw data output accordingly. I'm not
    quite sure about the ATN divisor you're talking about, but certainly 16 bit
    resolution would help..i've been using 12-16 bit ADC's

    >From: azeasi@a...
    >Reply-To: basicstamps@yahoogroups.com
    >To: basicstamps@yahoogroups.com
    >Subject: [noparse][[/noparse]basicstamps] Re: accelerometer issues and questions
    >(Parallax/Memsic devices)
    >Date: Tue, 22 Apr 2003 15:40:04 EDT
    >
    >I have been experimenting withe the Memsic 2125 accelerometer Demo kit from
    >Parallax and have noticed some quirks.
    >1.) The demo in the 2003 catalog uses a divisor of 16 supposedly to scale
    >the
    >xmG and ymG values so that the ATN function will work. (it only works for
    >byte size variables ...not word size). But the example in the app note that
    >is shipped with the demo kit uses a divisor of 8. Why the difference? Which
    >is better or would it be better to develop a 16 bit ATN function? Would
    >this
    >improve the accurracy or address the issue in part 2 below (certain angles
    >produce screwy results)?
    >2.) I am using the code for sensing 0-360 degrees of rotation, but the
    >value
    >"blows up" in certain positions (especially around +/- 90 degress or so.
    >Does
    >anyone know why this happens? Otherwise the ouput is fairly accurate and
    >repeatable.
    >
    >Does anyone have experience with this type of device? Any help,
    >suggestions,
    >or answers to my questions would be appreciated. Thanks!
    >
    >
    >[noparse][[/noparse]Non-text portions of this message have been removed]
    >
    >
    >To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    >from the same email address that you subscribed. Text in the Subject and
    >Body of the message will be ignored.
    >
    >
    >Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
    >
    >


    _________________________________________________________________
    STOP MORE SPAM with the new MSN 8 and get 2 months FREE*
    http://join.msn.com/?page=features/junkmail
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-24 06:40
    This may be saying the same thing a different way as Christian said it,
    but tilt measurements with accelerometers are based on the cosine of the
    angle of tilt with respect to the horizon. The cosine function shows
    the greatest change per degree near 0 degrees, and the least near 90
    degrees. This is true for all accelerometers, not just those based on
    piezoelectric material, i.e., it's the physics of gravitational
    components, not the sensor material.

    Dennis

    Original Message
    From: Christian Wentz [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=XmwMqO9bsFii3THmTbsFxq_NmMNNerI6CEfMypwhpw8rAB3hGwbvl7_sMvuVrBmPdNHDJPLxAN31qg]cwentz86@h...[/url
    Sent: Wednesday, April 23, 2003 3:41 PM
    To: basicstamps@yahoogroups.com
    Subject: Re: [noparse][[/noparse]basicstamps] Re: accelerometer issues and questions
    (Parallax/Memsic devices)


    I think what you're describing as a "blowing up" of data in certain
    positions comes from the nature of the sensor. I'm not familiar with
    this
    particular sensor, but in general the piezoelectric material on each
    axis is
    most sensitive parallel to the horizon. Because of this, the resolution
    declines and the change per degree of tilt goes from a negligible amount
    to
    more than 10 times what it was per degree of tilt at +/- 1 g. You may
    want
    to check the data sheet and adjust the raw data output accordingly. I'm
    not
    quite sure about the ATN divisor you're talking about, but certainly 16
    bit
    resolution would help..i've been using 12-16 bit ADC's

    >From: azeasi@a...
    >Reply-To: basicstamps@yahoogroups.com
    >To: basicstamps@yahoogroups.com
    >Subject: [noparse][[/noparse]basicstamps] Re: accelerometer issues and questions
    >(Parallax/Memsic devices)
    >Date: Tue, 22 Apr 2003 15:40:04 EDT
    >
    >I have been experimenting withe the Memsic 2125 accelerometer Demo kit
    >from Parallax and have noticed some quirks.
    >1.) The demo in the 2003 catalog uses a divisor of 16 supposedly to
    >scale
    >the
    >xmG and ymG values so that the ATN function will work. (it only works
    for
    >byte size variables ...not word size). But the example in the app note
    that
    >is shipped with the demo kit uses a divisor of 8. Why the difference?
    Which
    >is better or would it be better to develop a 16 bit ATN function? Would

    >this
    >improve the accurracy or address the issue in part 2 below (certain
    angles
    >produce screwy results)?
    >2.) I am using the code for sensing 0-360 degrees of rotation, but the
    >value
    >"blows up" in certain positions (especially around +/- 90 degress or
    so.
    >Does
    >anyone know why this happens? Otherwise the ouput is fairly accurate
    and
    >repeatable.
    >
    >Does anyone have experience with this type of device? Any help,
    >suggestions,
    >or answers to my questions would be appreciated. Thanks!
    >
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-24 21:49
    True, i suppose i didn't word that one the greatest-thanks for the
    clarification.
    Christian





    >From: "Dennis O'Leary" <doleary@e...>
    >Reply-To: basicstamps@yahoogroups.com
    >To: <basicstamps@yahoogroups.com>
    >Subject: RE: [noparse][[/noparse]basicstamps] Re: accelerometer issues and questions
    >(Parallax/Memsic devices)
    >Date: Wed, 23 Apr 2003 22:40:42 -0700
    >
    >This may be saying the same thing a different way as Christian said it,
    >but tilt measurements with accelerometers are based on the cosine of the
    >angle of tilt with respect to the horizon. The cosine function shows
    >the greatest change per degree near 0 degrees, and the least near 90
    >degrees. This is true for all accelerometers, not just those based on
    >piezoelectric material, i.e., it's the physics of gravitational
    >components, not the sensor material.
    >
    >Dennis
    >
    >
    Original Message
    >From: Christian Wentz [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=2boSTdzo49vccvCYL7r7pjxs-1jqAYdPYOfvrvWKo2kT__gOh4bZQVlxFhCloGwIA3CZVDUL_ZEEViCSzVHs]cwentz86@h...[/url
    >Sent: Wednesday, April 23, 2003 3:41 PM
    >To: basicstamps@yahoogroups.com
    >Subject: Re: [noparse][[/noparse]basicstamps] Re: accelerometer issues and questions
    >(Parallax/Memsic devices)
    >
    >
    >I think what you're describing as a "blowing up" of data in certain
    >positions comes from the nature of the sensor. I'm not familiar with
    >this
    >particular sensor, but in general the piezoelectric material on each
    >axis is
    >most sensitive parallel to the horizon. Because of this, the resolution
    >declines and the change per degree of tilt goes from a negligible amount
    >to
    >more than 10 times what it was per degree of tilt at +/- 1 g. You may
    >want
    >to check the data sheet and adjust the raw data output accordingly. I'm
    >not
    >quite sure about the ATN divisor you're talking about, but certainly 16
    >bit
    >resolution would help..i've been using 12-16 bit ADC's
    >
    > >From: azeasi@a...
    > >Reply-To: basicstamps@yahoogroups.com
    > >To: basicstamps@yahoogroups.com
    > >Subject: [noparse][[/noparse]basicstamps] Re: accelerometer issues and questions
    > >(Parallax/Memsic devices)
    > >Date: Tue, 22 Apr 2003 15:40:04 EDT
    > >
    > >I have been experimenting withe the Memsic 2125 accelerometer Demo kit
    > >from Parallax and have noticed some quirks.
    > >1.) The demo in the 2003 catalog uses a divisor of 16 supposedly to
    > >scale
    > >the
    > >xmG and ymG values so that the ATN function will work. (it only works
    >for
    > >byte size variables ...not word size). But the example in the app note
    >that
    > >is shipped with the demo kit uses a divisor of 8. Why the difference?
    >Which
    > >is better or would it be better to develop a 16 bit ATN function? Would
    >
    > >this
    > >improve the accurracy or address the issue in part 2 below (certain
    >angles
    > >produce screwy results)?
    > >2.) I am using the code for sensing 0-360 degrees of rotation, but the
    > >value
    > >"blows up" in certain positions (especially around +/- 90 degress or
    >so.
    > >Does
    > >anyone know why this happens? Otherwise the ouput is fairly accurate
    >and
    > >repeatable.
    > >
    > >Does anyone have experience with this type of device? Any help,
    > >suggestions,
    > >or answers to my questions would be appreciated. Thanks!
    > >
    >
    >
    >
    >To UNSUBSCRIBE, just send mail to:
    > basicstamps-unsubscribe@yahoogroups.com
    >from the same email address that you subscribed. Text in the Subject and
    >Body of the message will be ignored.
    >
    >
    >Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
    >
    >


    _________________________________________________________________
    Help STOP SPAM with the new MSN 8 and get 2 months FREE*
    http://join.msn.com/?page=features/junkmail
Sign In or Register to comment.