Load Cell Amplification Circuit

Discussion in 'The Projects Forum' started by fraser8, Sep 29, 2015.

  1. fraser8

    Thread Starter New Member

    May 19, 2015
    4
    0
    Hello,

    I'm building a handheld 5Kg weigh scale and have ordered a sample of sensors and am a little new to this. Attached is the datasheet.

    With ~5.09V to the sensor I've measured the following differential output. This does not appear as stated in the datasheet. The range should have been @0g a range of 0.29mV - 0.31mV and @5000g a range of 9.8065mV - 11.2535mV.

    Recorded Data:

    Sensor 1 @ 0g = 5.315460mV
    Sensor 1 @ 5000g = 17.226600mV
    Sensor 2@ 0g = 3.575580mV
    Sensor 2@ 5000g = 15.388600mV
    Sensor 3 @ 0g = 0.685243mV
    Sensor 3 @ 5000g = 12.290700mV
    Sensor 4 @ 0g = 6.298300mV
    Sensor 4 @ 5000g = 18.079000mV

    How can I compensate for this range, especially the offset? I have been using the INA126 and a gain of 277 would allow me to measure the 18mV Sensor 4 on a 10bit ADC at 5V. However, this limits my range to 1.75V-5V. I would obviously like to realize the full 0-5V of the 10bit ADC.

    Is there a potential design that allows me to build once, or does the board need to be customized for each sensor? I am also unsure how to trim the offset using the OPA237.

    Regards,
    Fraser
     
  2. joeyd999

    AAC Fanatic!

    Jun 6, 2011
    2,690
    2,755
    You should try to figure out why the load cell outputs are so far from the spec. This, IMHO, clearly indicates something wrong with either the cells or the way you are measuring them.

    With that said, I no longer use analog front ends for load cell processing. High resolution delta sigma converters have a wide enough dynamic range to allow all processing, including offset and scaling, in the digital realm.
     
  3. fraser8

    Thread Starter New Member

    May 19, 2015
    4
    0
    Thanks joeyd

    Do you have a suggestion for the converter? I happened to have the AD7780 evaluation board that I was using to get my readings. At a $17 price point it's a little much. What are your thoughts?
     
  4. ericgibbs

    AAC Fanatic!

    Jan 29, 2010
    2,503
    380
    hi,

    At 2mV/V for a 5V supply at 5kg, should give 10mV output with a 5kg loading. [ according to the datasheet]

    A Zero balance of 0.3mV +/-%FS is typical.

    How are you measuring the differential from the bridge.???

    The values you are posting are way off the spec.

    E
     
  5. joeyd999

    AAC Fanatic!

    Jun 6, 2011
    2,690
    2,755

    I've had great success with the TI ADS1242 (it includes a x128 digital PGA).

    Another that I've been looking at but haven't tried is the Nuvoton NAU7802. The price is excellent, but I am not sure I am comfortable with the specs yet.
     
  6. fraser8

    Thread Starter New Member

    May 19, 2015
    4
    0
    Thanks for the reply.

    I was using a Fluke to measure the mV and that was giving me one decimal of resolution. I have an AD7780 evaluation board that I was using to further measure the differential. This is over USB to my laptop.

    I have followed up with the supplier to get further support. But what you are suggesting is these load cells are out of spec. Am I correct in saying the Zero Balance should be no greater than 0.3mV?
     
  7. ericgibbs

    AAC Fanatic!

    Jan 29, 2010
    2,503
    380
    hi,
    The zero offset of 0.3mV plus the very small +/-0.1%FS is well exceeded from the test results you have posted.

    With the bridge connected to the INA126 and the gain set to 100, what Vout do you read when there is no load on the bridge.? [let me know]
    I assume that the INA is being dual powered.?
    Can you post a circuit showing the bridge and INA.?

    In normal circumstances the designer of bridge amp would provide a Zero pot to set the no load output to 0Vand also a Span pot for the scaling to FSD

    E
     
  8. fraser8

    Thread Starter New Member

    May 19, 2015
    4
    0
    I attached sensor 3 to the INA125. The readout on the Fluke is 2.923V.

    Doing the math will confirm this. The 100 ohm resistor will apply a gain of 805 based on the datasheet. The zero reading was 0.685243mV and so amplifying based on the previous reading will get me to 2.878V, which is close to what I'm seeing today.

    How do you suggest providing the zero? I am using single supply, as well as the single supply version of the INA125.
     
  9. ericgibbs

    AAC Fanatic!

    Jan 29, 2010
    2,503
    380
    hi,
    The single supply option should be connected as shown in Fig #4 of the datasheet, with Vref pin #5 at ~1,2V, see image

    I said to set the Gain to 100 , that would mean a 842R resistor.

    The above Zero reading of 0.685243mV * 805 [ which is the Gain value you have set] is ~ 0.5515V not 2.878V as you have posted for a no load on the bridge.????

    Configure the 1NA122 for 5V supply as shown, then set the gain for 100 times, using a 842R resistor for Rg.

    Re-measure and recheck your calculations.

    E
     
  10. ericgibbs

    AAC Fanatic!

    Jan 29, 2010
    2,503
    380
    hi,
    This is your bridge and INA126 in LTspice simualtion.
    The bridge test input signal is 0v thru 10mV for both plots.

    Note:
    Image Vref0.gif shows the effect of having pin#5 at 0V, ie: no Vref.
    So when operating with a single supply the lower level output voltage is non linear and close to zero volts.
    The gain is *500

    Image 2: Vref1v.gif is when Vref is at +1V, the output voltage swings from +1V thru +4V.
    Note: it will limit at +4V output due to the OPA design.
    The Gain is * ~270

    E
     
Loading...