Shop OBEX P1 Docs P2 Docs Learn Events
BS2 Accuracy — Parallax Forums

BS2 Accuracy

ArchiverArchiver Posts: 46,084
edited 2004-07-19 18:50 in General Discussion
If I build a tach circuit for a BS2 how is the best way to determine the
accuracy.

My idea was to use a standard industrial magnetic pickup, fed to a NPN
transistor, then to the stamp.
The stamp would time the pulses, then output to a 10 bit dac indicating RPM
in voltage of 0-5 vdc.

My flywheel would be rotating up to 1800 rpm, and I would count flywheel
teeth anywhere from 183 to 208 depending on the engine.

I figured since I'm using a 10 bit dac, I would make the design range good
for 2048 RPM so each step of the dac would be equiv. to 1 rpm.

For example, say I have an engine running @ 1500 rpm and it has 183 flywheel
teeth. = 274500 / 60 = 4575 hz
The Dac would output 3.66 volts or 1500 steps.

How is the best way to determine the accuracy of the measurement?



[noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]

Comments

  • ArchiverArchiver Posts: 46,084
    edited 2004-07-18 14:50
    Sorry, I meant the stamp would "Count" the pulses.....

    Original Message
    From: "Richard" <rwskinner@a...>
    To: <basicstamps@yahoogroups.com>
    Sent: Sunday, July 18, 2004 8:46 AM
    Subject: [noparse][[/noparse]basicstamps] BS2 Accuracy


    > If I build a tach circuit for a BS2 how is the best way to determine the
    > accuracy.
    >
    > My idea was to use a standard industrial magnetic pickup, fed to a NPN
    > transistor, then to the stamp.
    > The stamp would time the pulses, then output to a 10 bit dac indicating
    RPM
    > in voltage of 0-5 vdc.
    >
    > My flywheel would be rotating up to 1800 rpm, and I would count flywheel
    > teeth anywhere from 183 to 208 depending on the engine.
    >
    > I figured since I'm using a 10 bit dac, I would make the design range good
    > for 2048 RPM so each step of the dac would be equiv. to 1 rpm.
    >
    > For example, say I have an engine running @ 1500 rpm and it has 183
    flywheel
    > teeth. = 274500 / 60 = 4575 hz
    > The Dac would output 3.66 volts or 1500 steps.
    >
    > How is the best way to determine the accuracy of the measurement?
    >
    >
    >
    > [noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]
    >
    >
    >
    > 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.
    >
    > Yahoo! Groups Links
    >
    >
    >
    >
    >
    >
    > [noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]
    >
    >


    [noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]
  • ArchiverArchiver Posts: 46,084
    edited 2004-07-18 19:41
    --- In basicstamps@yahoogroups.com, "Richard" <rwskinner@a...> wrote:
    > [noparse][[/noparse]...] How is the best way to determine the accuracy of the
    > measurement?

    Generally, accuracy is determined by comparing a result against a
    another result from a device of known accuracy.

    When I built a Stamp-based controller for my diesel genset, I used a
    magnetic pickup mounted near the flywheel teeth as you describe. The
    output from the pickup was clipped and fed to a transistor to sharpen
    the waveform and then fed to the BS2. In my case, I displayed the
    RPM value on an LCD instead of generating a proportional analog
    voltage. I compared the displayed result to the value from a
    frequency counter.

    You can get more information on my project, including schematics and
    source code at:

    http://www.kinzers.com/don/GenSet
  • ArchiverArchiver Posts: 46,084
    edited 2004-07-19 04:30
    Richard,

    I recently made a tachometer for a friend's boat. It runs in the same sort
    of range that you are looking at. I used a magnet on the flywheel and a hall
    sensor with a built in Schmitt trigger.To verify the accuracy, I used an
    Optascope. It agrees perfectly with the Stamp.

    Jonathan

    www.madlabs.info


    Original Message
    From: "Richard" <rwskinner@a...>
    To: <basicstamps@yahoogroups.com>
    Sent: Sunday, July 18, 2004 6:46 AM
    Subject: [noparse][[/noparse]basicstamps] BS2 Accuracy


    > If I build a tach circuit for a BS2 how is the best way to determine the
    > accuracy.
    >
    > My idea was to use a standard industrial magnetic pickup, fed to a NPN
    > transistor, then to the stamp.
    > The stamp would time the pulses, then output to a 10 bit dac indicating
    RPM
    > in voltage of 0-5 vdc.
    >
    > My flywheel would be rotating up to 1800 rpm, and I would count flywheel
    > teeth anywhere from 183 to 208 depending on the engine.
    >
    > I figured since I'm using a 10 bit dac, I would make the design range good
    > for 2048 RPM so each step of the dac would be equiv. to 1 rpm.
    >
    > For example, say I have an engine running @ 1500 rpm and it has 183
    flywheel
    > teeth. = 274500 / 60 = 4575 hz
    > The Dac would output 3.66 volts or 1500 steps.
    >
    > How is the best way to determine the accuracy of the measurement?
    >
    >
    >
    > [noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]
    >
    >
    >
    > 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.
    >
    > Yahoo! Groups Links
    >
    >
    >
    >
    >
    >
  • ArchiverArchiver Posts: 46,084
    edited 2004-07-19 14:44
    The best way to determine the accuracy of your setup is to capare its
    output with a calibrated instrument.

    -- Jon Williams
    -- Applications Engineer, Parallax
    -- Dallas Office


    Original Message
    From: Richard [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=L0uPyszTqIwoTsqC6oW2tdmPpdjVyTyNUC5LpvcofKDWmCjGX38yahXemjFPBYNDbCZ7VNfyaQIyJ2GFvQl_]rwskinner@a...[/url
    Sent: Sunday, July 18, 2004 8:47 AM
    To: basicstamps@yahoogroups.com
    Subject: [noparse][[/noparse]basicstamps] BS2 Accuracy


    If I build a tach circuit for a BS2 how is the best way to determine the
    accuracy.

    My idea was to use a standard industrial magnetic pickup, fed to a NPN
    transistor, then to the stamp. The stamp would time the pulses, then
    output to a 10 bit dac indicating RPM in voltage of 0-5 vdc.

    My flywheel would be rotating up to 1800 rpm, and I would count
    flywheel teeth anywhere from 183 to 208 depending on the engine.

    I figured since I'm using a 10 bit dac, I would make the design range
    good for 2048 RPM so each step of the dac would be equiv. to 1 rpm.

    For example, say I have an engine running @ 1500 rpm and it has 183
    flywheel teeth. = 274500 / 60 = 4575 hz The Dac would output 3.66 volts
    or 1500 steps.

    How is the best way to determine the accuracy of the measurement?
  • ArchiverArchiver Posts: 46,084
    edited 2004-07-19 16:14
    You could do a worst case analysis.

    With the 10 bit DAC, that is 1024 steps, so with 2048rpm full scale,
    that is 2rpm per bit (not 1rpm!). Let's say you have a reading of
    1000. If the accuracy of the DAC is 1 bit, then your rpm could be
    anywhere from 998 to 1002. There may be other errors (linearity,
    offset, monotonicity) in the DAC that you need to consider. If you
    have an accurate voltmeter, you can use that to check the calibration
    of the DAC and to make adjustments in software if need be.

    The accuracy of the stamp resonator affects the accuracy of the COUNT
    command. These resonators are typically very good, but the
    specification is not as tight as a crystal. +/- 0.5 percent initial
    accuracy and +/- 0.3 percent change over a -20 to +80 temperature
    range. So just to take 0.5% of 1000 rpm, that is +/- 5 rpm. You can
    measure the accuracy of your ceramic resonator by feeding an accurate
    frequency source into a stamp pin, and COUNT for a long enough
    period. That way you can calibrate the COUNT command for that
    particular Stamp.

    Once you calibrate both the DAC and the COUNT command, those are the
    error sources in this system, assuming the magnetic pickup is working
    as it should.

    -- Tracy


    >If I build a tach circuit for a BS2 how is the best way to determine the
    >accuracy.
    >
    >My idea was to use a standard industrial magnetic pickup, fed to a NPN
    >transistor, then to the stamp.
    >The stamp would time the pulses, then output to a 10 bit dac indicating RPM
    >in voltage of 0-5 vdc.
    >
    >My flywheel would be rotating up to 1800 rpm, and I would count flywheel
    >teeth anywhere from 183 to 208 depending on the engine.
    >
    >I figured since I'm using a 10 bit dac, I would make the design range good
    >for 2048 RPM so each step of the dac would be equiv. to 1 rpm.
    >
    >For example, say I have an engine running @ 1500 rpm and it has 183 flywheel
    >teeth. = 274500 / 60 = 4575 hz
    >The Dac would output 3.66 volts or 1500 steps.
    >
    >How is the best way to determine the accuracy of the measurement?
  • ArchiverArchiver Posts: 46,084
    edited 2004-07-19 18:50
    You have answered most of it. Good catch on the number of steps. Half
    asleep I guess.

    Sorry, I asked my question incorrectly. I know how to compare readings
    between instruments, but I was curious to how accurate the stamp would be
    using the count command with my frequency range.

    The stamp would be pretty much dedicated to RPM and Voltage output. I'd
    like to get 2 samples per second and hold it within about 5 rpm accuracy .
    I could use offsets if needed since we "mainly" care about the 700-1500 RPM
    Range.




    Original Message
    From: "Tracy Allen" <tracy@e...>
    To: <basicstamps@yahoogroups.com>
    Sent: Monday, July 19, 2004 10:14 AM
    Subject: Re: [noparse][[/noparse]basicstamps] BS2 Accuracy


    > You could do a worst case analysis.
    >
    > With the 10 bit DAC, that is 1024 steps, so with 2048rpm full scale,
    > that is 2rpm per bit (not 1rpm!). Let's say you have a reading of
    > 1000. If the accuracy of the DAC is 1 bit, then your rpm could be
    > anywhere from 998 to 1002. There may be other errors (linearity,
    > offset, monotonicity) in the DAC that you need to consider. If you
    > have an accurate voltmeter, you can use that to check the calibration
    > of the DAC and to make adjustments in software if need be.
    >
    > The accuracy of the stamp resonator affects the accuracy of the COUNT
    > command. These resonators are typically very good, but the
    > specification is not as tight as a crystal. +/- 0.5 percent initial
    > accuracy and +/- 0.3 percent change over a -20 to +80 temperature
    > range. So just to take 0.5% of 1000 rpm, that is +/- 5 rpm. You can
    > measure the accuracy of your ceramic resonator by feeding an accurate
    > frequency source into a stamp pin, and COUNT for a long enough
    > period. That way you can calibrate the COUNT command for that
    > particular Stamp.
    >
    > Once you calibrate both the DAC and the COUNT command, those are the
    > error sources in this system, assuming the magnetic pickup is working
    > as it should.
    >
    > -- Tracy
    >
    >
    > >If I build a tach circuit for a BS2 how is the best way to determine the
    > >accuracy.
    > >
    > >My idea was to use a standard industrial magnetic pickup, fed to a NPN
    > >transistor, then to the stamp.
    > >The stamp would time the pulses, then output to a 10 bit dac indicating
    RPM
    > >in voltage of 0-5 vdc.
    > >
    > >My flywheel would be rotating up to 1800 rpm, and I would count flywheel
    > >teeth anywhere from 183 to 208 depending on the engine.
    > >
    > >I figured since I'm using a 10 bit dac, I would make the design range
    good
    > >for 2048 RPM so each step of the dac would be equiv. to 1 rpm.
    > >
    > >For example, say I have an engine running @ 1500 rpm and it has 183
    flywheel
    > >teeth. = 274500 / 60 = 4575 hz
    > >The Dac would output 3.66 volts or 1500 steps.
    > >
    > >How is the best way to determine the accuracy of the measurement?
    >
    >
    > 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.
    >
    > Yahoo! Groups Links
    >
    >
    >
    >
    >
    >
    > [noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]
    >
    >


    [noparse][[/noparse]Scanned by AwesomeNet Anti-Virus]
Sign In or Register to comment.