Shop OBEX P1 Docs P2 Docs Learn Events
Integrate me — Parallax Forums

Integrate me

ArchiverArchiver Posts: 46,084
edited 2003-05-01 22:59 in General Discussion
I have a rotational rate sensor which outputs a voltage proportional to
rotation speed (.11 mv per degrees / second). I want to integrate the
rate over time to give position (i.e. North, South East West and the
points in between). The BS2P will be performing other simultaneous
functions so how do I establish a known and steady sampling rate to
achieve an accurate integration? If possible I'd like to sample at about
10 Hz. Any ideas?

Thanks,
Chris

________________________________________________________________
The best thing to hit the internet in years - Juno SpeedBand!
Surf the web up to FIVE TIMES FASTER!
Only $14.95/ month - visit www.juno.com to sign up today!

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2003-04-28 22:58
    I don't know that I can help you, but that sounds like a neat sensor.
    Where did you get it? How much was it? Thanks!

    - Robert
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-29 06:28
    Accurate numerical integration of an angular velocity sensor to produce
    angular position is difficult at best, because any DC voltage offsets
    will also be integrated to result in positional errors. The longer the
    integration time, the more cumulative error. I attempt to do this
    computation, using Murata angular velocity sensors, for a project
    measuring human postural sway, but I limit the integration time to about
    20 seconds, and also detrend the positional data to remove the offset
    errors. This takes advantage of the "boundary conditions" of human
    postural sway, in order to limit the cumulative errors. So a steady
    smapling rate, for A/D conversion can be easily done, but it is not the
    main problem.

    An alternative approach is to use a digital compass, which has the
    advantage that it has an external positional reference (magnetic north
    pole), and interface that to a stamp. Other list members, and also Nuts
    & Volts, have experience with these.

    Dennis
    Original Message
    From: Christopher Dundorf [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=TZPk3BpzRpP9NIB5RCbMYPqH64lz5TtKJlWYQYMq7ypr_HI2r4rMeHWIfUB8421jxFZHQdhmHoci]cdundorf@j...[/url
    Sent: Monday, April 28, 2003 2:43 PM
    To: basicstamps@yahoogroups.com
    Subject: [noparse][[/noparse]basicstamps] Integrate me


    I have a rotational rate sensor which outputs a voltage proportional to
    rotation speed (.11 mv per degrees / second). I want to integrate the
    rate over time to give position (i.e. North, South East West and the
    points in between). The BS2P will be performing other simultaneous
    functions so how do I establish a known and steady sampling rate to
    achieve an accurate integration? If possible I'd like to sample at
    about 10 Hz. Any ideas?

    Thanks,
    Chris

    ________________________________________________________________
    The best thing to hit the internet in years - Juno SpeedBand! Surf the
    web up to FIVE TIMES FASTER! Only $14.95/ month - visit www.juno.com to
    sign up today!

    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/
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-29 19:00
    Chris,

    I'm using a through-hole version of the MuRata Gyrostar -- 03JA and
    03JB, but with the same sensor characteristics as the SMD versions in
    your web reference. I had to buy these in quantity (A & B versions)
    about 3 years ago when they were first marketed. Yes, I reset the
    integration. One method is based on using the MuRata in combination
    with Analog Devices ADXL202 linear sensors, using the latter to watch
    for a "null point" of human sway where the standing human as an
    "inverted pendulum model" passes through a vertical axis.

    I also have the ADXRS150, but haven't compared it with the MuRata yet.
    My guess is that it will have lower drift, because of the on-board
    temperature sensor, and the capability to then apply a temperature
    compensation table for drift in a Stamp or PIC. The trade-off is that
    you have to contend with a BGA (ball grid array) form factor.

    For accurate sampling rates using PICs, I use an external crystal
    oscillator, and divide down by 2s to reach a particular sampling rate.
    For example, using a PIC16C622A, I use an oscillator at 4.096 MHz, and
    divide down to a sampling rate of 1000 or 500 to trigger interrupts in
    the PIC. To get an accurate rate of 10 Hz, my approach would be to
    oversample at 500 Hz, apply some digital averaging (sliding boxcar
    filter) and then store every 50th point (decimation by 50) for an
    overall rate of 10 Hz. I don't know if this approach will work with a
    Basic Stamp, given its lack of a true interrupt capability, but others
    on this list will have some alternative approaches for the Stamp.

    Only reason I mentioned a digital compass, is that I've been impressed
    watching a robotic mower, which has a built-in compass, graze around my
    lawn.

    Dennis
    Original Message
    From: cdundorf@j... [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=4uYi7-JXmfy7fPizycV-rQMH2kNKcKE-BgmwcfTzDwv-zTLMQG5QtfQQvJ7Fd3zkoKPvpXDJXk0]cdundorf@j...[/url
    Sent: Tuesday, April 29, 2003 7:01 AM
    To: basicstamps@yahoogroups.com
    Cc: doleary@u...
    Subject: [noparse][[/noparse]basicstamps] RE: Integrate me



    Dennis,

    Unfortunately, a digital compass is not an option. It's sound like
    you've already spent sometime in this arena. I too am concerned about
    offset bias. Do you use the boundry conditions as "trip wire" to reset
    the intergration?

    You mentioned MuRata, are you using the MuRata GyroStar
    http://www.murata.com/catalog/s42e3.pdf? I'm running the Analog Devices
    ADXRS150. I am interested in your thoughts on how they compare, namely
    drift. I could find much else on the GyroStar specs.

    As for gathering the data integrate, do you have suggestions for how
    achieve a steady 10 Hz sampling rate using a Basic Stamp?

    Thanks,
    Chris

    <<<>>>


    Subject: RE: Integrate me

    Accurate numerical integration of an angular velocity sensor to produce
    angular position is difficult at best, because any DC voltage offsets
    will also be integrated to result in positional errors. The longer the
    integration time, the more cumulative error. I attempt to do this
    computation, using Murata angular velocity sensors, for a project
    measuring human postural sway, but I limit the integration time to about
    20 seconds, and also detrend the positional data to remove the offset
    errors. This takes advantage of the "boundary conditions" of human
    postural sway, in order to limit the cumulative errors. So a steady
    smapling rate, for A/D conversion can be easily done, but it is not the
    main problem.

    An alternative approach is to use a digital compass, which has the
    advantage that it has an external positional reference (magnetic north
    pole), and interface that to a stamp. Other list members, and also Nuts
    & Volts, have experience with these.

    Dennis
    Original Message
    From: Christopher Dundorf [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=4uYi7-JXmfy7fPizycV-rQMH2kNKcKE-BgmwcfTzDwv-zTLMQG5QtfQQvJ7Fd3zkoKPvpXDJXk0]cdundorf@j...[/url
    Sent: Monday, April 28, 2003 2:43 PM
    To: basicstamps@yahoogroups.com
    Subject: [noparse][[/noparse]basicstamps] Integrate me


    I have a rotational rate sensor which outputs a voltage proportional to
    rotation speed (.11 mv per degrees / second). I want to integrate the
    rate over time to give position (i.e. North, South East West and the
    points in between). The BS2P will be performing other simultaneous
    functions so how do I establish a known and steady sampling rate to
    achieve an accurate integration? If possible I'd like to sample at
    about 10 Hz. Any ideas?

    Thanks,
    Chris

    ________________________________________________________________
    The best thing to hit the internet in years - Juno SpeedBand! Surf the
    web up to FIVE TIMES FASTER! Only $14.95/ month - visit www.juno.com to
    sign up today!

    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/
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-29 22:00
    Dennis,

    Unfortunately, a digital compass is not an option. It's sound like you've
    already spent sometime in this arena. I too am concerned about offset bias. Do
    you use the boundry conditions as "trip wire" to reset the intergration?

    You mentioned MuRata, are you using the MuRata GyroStar
    http://www.murata.com/catalog/s42e3.pdf? I'm running the Analog Devices
    ADXRS150. I am interested in your thoughts on how they compare, namely drift.
    I could find much else on the GyroStar specs.

    As for gathering the data integrate, do you have suggestions for how achieve a
    steady 10 Hz sampling rate using a Basic Stamp?

    Thanks,
    Chris

    <<<>>>


    Subject: RE: Integrate me

    Accurate numerical integration of an angular velocity sensor to produce
    angular position is difficult at best, because any DC voltage offsets
    will also be integrated to result in positional errors. The longer the
    integration time, the more cumulative error. I attempt to do this
    computation, using Murata angular velocity sensors, for a project
    measuring human postural sway, but I limit the integration time to about
    20 seconds, and also detrend the positional data to remove the offset
    errors. This takes advantage of the "boundary conditions" of human
    postural sway, in order to limit the cumulative errors. So a steady
    smapling rate, for A/D conversion can be easily done, but it is not the
    main problem.

    An alternative approach is to use a digital compass, which has the
    advantage that it has an external positional reference (magnetic north
    pole), and interface that to a stamp. Other list members, and also Nuts
    & Volts, have experience with these.

    Dennis
    Original Message
    From: Christopher Dundorf [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=TYMU4re7rALyD6Su7duMEq1py_A8hYDCQCp_BsAbA1ucdjS44CWVPDwmY-Ws52mj7cO85FzkVN2u]cdundorf@j...[/url
    Sent: Monday, April 28, 2003 2:43 PM
    To: basicstamps@yahoogroups.com
    Subject: [noparse][[/noparse]basicstamps] Integrate me


    I have a rotational rate sensor which outputs a voltage proportional to
    rotation speed (.11 mv per degrees / second). I want to integrate the
    rate over time to give position (i.e. North, South East West and the
    points in between). The BS2P will be performing other simultaneous
    functions so how do I establish a known and steady sampling rate to
    achieve an accurate integration? If possible I'd like to sample at
    about 10 Hz. Any ideas?

    Thanks,
    Chris

    ________________________________________________________________
    The best thing to hit the internet in years - Juno SpeedBand!
    Surf the web up to FIVE TIMES FASTER!
    Only $14.95/ month - visit www.juno.com to sign up today!
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-29 22:06
    Hi Robert,

    I'm using the Analog Devices ADXRS150. The evaluation board is about $40 and
    easy to use. Check out
    http://www.analog.com/Analog_Root/productPage/productHome/0,2121,ADXRS150,00.htm\
    l

    Good Luck,
    Chris

    <<>>>

    I don't know that I can help you, but that sounds like a neat sensor.
    Where did you get it? How much was it? Thanks!

    - Robert



    ________________________________________________________________
    The best thing to hit the internet in years - Juno SpeedBand!
    Surf the web up to FIVE TIMES FASTER!
    Only $14.95/ month - visit www.juno.com to sign up today!
  • ArchiverArchiver Posts: 46,084
    edited 2003-04-30 11:35
    It really sounds like you just need a dedicated chip. I'd suggest
    using a dedicated PIC (or BS1 if you want to stick with P-BASIC) for
    the sampling and just have the BS2 talk to the PIC. - John

    I like that sensor idea, btw!

    --- In basicstamps@yahoogroups.com, Christopher Dundorf
    <cdundorf@j...> wrote:
    > I have a rotational rate sensor which outputs a voltage
    proportional to
    > rotation speed (.11 mv per degrees / second). I want to integrate
    the
    > rate over time to give position (i.e. North, South East West and
    the
    > points in between). The BS2P will be performing other simultaneous
    > functions so how do I establish a known and steady sampling rate to
    > achieve an accurate integration? If possible I'd like to sample
    at about
    > 10 Hz. Any ideas?
    >
    > Thanks,
    > Chris
    >
    > ________________________________________________________________
    > The best thing to hit the internet in years - Juno SpeedBand!
    > Surf the web up to FIVE TIMES FASTER!
    > Only $14.95/ month - visit www.juno.com to sign up today!
  • ArchiverArchiver Posts: 46,084
    edited 2003-05-01 17:52
    I have it also and I agree. However, for building into a project
    without using the evaluation board, the BGA structure of this chip
    presents problems, as compared with other available SMD or through hole
    sensors.
    Dennis
    Original Message
    From: cdundorf@j... [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=1jkVMpaT-AfPtZblGipPPnntZ40TS9Yr17v2IiaBVvDinNb6jyiZvuT6fkCFe2hCi5nCzXJx-_cRbIs]cdundorf@j...[/url
    Sent: Tuesday, April 29, 2003 7:07 AM
    To: uavscience@f...
    Cc: basicstamps@yahoogroups.com
    Subject: [noparse][[/noparse]basicstamps] Re: Integrate me



    Hi Robert,

    I'm using the Analog Devices ADXRS150. The evaluation board is about
    $40 and easy to use. Check out
    http://www.analog.com/Analog_Root/productPage/productHome/0,2121,ADXRS15
    0,00.html

    Good Luck,
    Chris

    <<>>>

    I don't know that I can help you, but that sounds like a neat sensor.
    Where did you get it? How much was it? Thanks!

    - Robert



    ________________________________________________________________
    The best thing to hit the internet in years - Juno SpeedBand! Surf the
    web up to FIVE TIMES FASTER! Only $14.95/ month - visit www.juno.com to
    sign up today!

    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/
  • ArchiverArchiver Posts: 46,084
    edited 2003-05-01 21:12
    Is there an adapter or other solution?
    > I have it also and I agree. However, for building into a project
    > without using the evaluation board, the BGA structure of this chip
    > presents problems, as compared with other available SMD or through
    > hole sensors. Dennis
    >
    AD's WEB site says $50, $100! from Newark.
    >
    > Hi Robert,
    >
    > I'm using the Analog Devices ADXRS150. The evaluation board is about
    > $40 and easy to use. Check out
    > http://www.analog.com/Analog_Root/productPage/productHome/0,2121,ADXRS
    > 15 0,00.html
    >
    > Good Luck,
    > Chris
  • ArchiverArchiver Posts: 46,084
    edited 2003-05-01 22:59
    Mike,

    Yes, there's always a solution. Last time I checked, a BGA machine cost
    about $35K. [noparse]:)[/noparse]

    Dennis
    Original Message
    From: Mike DeMetz [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=3ZnvGx67fD6TSGHweSXDkbanRs11tMxntGe4YN_JqCR6sWjtlg8PsJjjSPvJ9XhZNt49y04Ah2g]miked@e...[/url
    Sent: Thursday, May 01, 2003 1:13 PM
    To: basicstamps@yahoogroups.com
    Subject: [noparse][[/noparse]basicstamps] Re: Integrate me


    Is there an adapter or other solution?
    > I have it also and I agree. However, for building into a project
    > without using the evaluation board, the BGA structure of this chip
    > presents problems, as compared with other available SMD or through
    > hole sensors. Dennis
    >
    AD's WEB site says $50, $100! from Newark.
    >
    > Hi Robert,
    >
    > I'm using the Analog Devices ADXRS150. The evaluation board is about
    > $40 and easy to use. Check out
    > http://www.analog.com/Analog_Root/productPage/productHome/0,2121,ADXRS
    > 15 0,00.html
    >
    > Good Luck,
    > Chris


    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/
Sign In or Register to comment.