1. This forum section is a read-only archive which contains old newsgroup posts. If you wish to post a query, please do so in one of our main forum sections (here). This way you will get a faster, better response from the members on Motherboard Point.

Tapping signal from digital scale's weight sensor

Discussion in 'Embedded' started by Cedi, May 24, 2006.

  1. Cedi

    Cedi Guest

    Hi Group,


    I intend to make a weight aqcuistion digital system. For the sensor I
    bought commercial digital scale that
    have range 0-5kg with 1gr resolution. (TANITA KD-403) After I opened
    the casing, I found that the load sensor
    and the processor connected by 4 wire: V+, V-, S+, and S- (this written
    on the pcb). I guessed that the sensor is strain
    gauge with bridge. So I did some research to tap signal from the
    sensor. I measure voltage (with ADC) between V+
    and V-, and the reading was toggling, between 0 and about 2.5V (with or
    without load on the gauge).
    I think this is intended for battery saving and for minimizing self
    heating effect on the sensor. And when I measured
    voltage between S+ and S-, the reading is always zero. what I don't
    understand is, the value is not changed as I change
    the load on the gauge. The theory I know says that (assume the sensor
    is really strain gauge) that the voltage across s+
    and s- would depend on the load given (because this is the one that
    indicates how strong force given to the sensor).
    Strange thing is the scale was still working perfectly (show a correct
    weight on its own lcd display) when I tap the signals.

    Hope someone has deeper knowledge on this. Is there possibilities that
    the sensor is not strain gauge?
    Or there's a tricky modulation using regular strain gauge so the
    voltage across s+ and s- doesn't need
    to depend on the load? Some pulse width modulation or delayed response?
    I'm really blank on this.

    Any help would be greatly appreciated


    thanks,
    Cedi
     
    Cedi, May 24, 2006
    #1
    1. Advertisements

  2. Cedi

    MK Guest

    A typical load cell will give an output of 2mV per volt of excitation at
    full scale. With 2.5V excitation you can expect about 1mV per kg of load
    from your scale - were you looking for such a small signal ?
    They may be pulsing the excitaion to save power or as part of an auto
    zeroing system.
    Try searching Linear Technology's web site for app notes about load cell
    signal conditioning and HBM's for specs on load cells.

    Michael Kellett

    www.mkesc.co.uk
     
    MK, May 24, 2006
    #2
    1. Advertisements

  3. This sounds like a chopper stabilised amplifier. Any DC amplifier will
    suffer from drift, but if you can use an AC connected amplifier, this
    should not be a problem.

    When driving the sensor bridge by a square wave, the retrieved sensor
    output voltage will switch between 0 V and some very low voltage (a
    few mV). This voltage swing is amplified by an AC amplifier and a
    synchronous detector is used to put the "0 V" setting as the reference
    setting and using the other value as the actual measurement value X.

    Any drift in the amplifier will affect both the "0 V" level as well as
    the "X" voltage level, but the difference X-0V is proportional to the
    value of interest.

    Paul
     
    Paul Keinanen, May 24, 2006
    #3
  4. Cedi

    David Tweed Guest

    Strain gauge signals are very low-level. Your voltmeter may not have
    the necessary sensitivity.

    This came up recently in a Circuit Cellar design contest, which was
    subsequently written up as an article:

    http://www.dtweed.com/circuitcellar/caj00191.htm#3288

    The author looked further downstream in the scale's own circuitry
    and found a PWM siganl on a marked test point whose duty cycle was
    linearly related to the weight on the scale. Perhaps your scale
    has something similar.

    -- Dave Tweed
     
    David Tweed, May 25, 2006
    #4
  5. Cedi

    David Tweed Guest

    Strain gauge signals are very low-level. Your voltmeter may not have
    the necessary sensitivity.

    This came up recently in a Circuit Cellar design contest, which was
    subsequently written up as an article:

    http://www.dtweed.com/circuitcellar/caj00191.htm#3288

    The author looked further downstream in the scale's own circuitry
    and found a PWM siganl on a marked test point whose duty cycle was
    linearly related to the weight on the scale. Perhaps your scale
    has something similar.

    -- Dave Tweed
     
    David Tweed, May 26, 2006
    #5
  6. Cedi

    Cedi Guest

    Thanks for the replies.

    I checked voltage across S+ and S- with Osciloscope and yes, I could
    see the signal. The signal was ranging between 0 to about 5mv (full
    scale). The signal was a bit noisy.

    I'm planning to amplify it to 0-5V to fit my ADC voltage range. And do
    the averaging to eliminate noise in processor. I guess I need a
    instrumentation amplifier to achieve this. Any suggestion what
    amplifiers that fit to this requirement?

    Could anybody direct me to web site links that useful for this
    applications?

    anyway come into my mind that maybe it's easier to tap the final result
    from digital scales's lcd, the lcd is 4 digits 7 segment. Physically
    connected to PCB processor by 16 conductors (elastomeric conductor). Is
    it feasible to do this?


    Thanks,
    Cedi
     
    Cedi, May 27, 2006
    #6
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.