• State
  • Locked Locked
  • Replies 9 replies
  • Subscribers 47 subscribers
  • Views 5817 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

INA128 Sensitivity Issues

Jake Vonderharr
Prodigy 140 points
Other Parts Discussed in Thread: INA128

Hi,  I'm attempting to use an INA128 in a load cell instrumentation application.  I have been successful at setting a gain that will deliver a reasonable range of voltage response, but the problem I am having is that the sensitivity of the load cell on the low end doesn't seem to be where I would like it.  In example, adding one gram of weight to the load cell will provide ~3mV of change in voltage.  This is the best I have been able to achieve while adjusting gain values.  I need a larger value, as the micro-controller I'm using can only sense as low as 5mV increments.  The load cell I believe is balanced and to be hooked directly into the sense +/- inputs.  My input voltage is 5V, and reference is 2.5V.  Excitation voltage is 5V.

My ultimate goal is to amplify this load cell to allow a resolution of 0.5g reading and therefore need to successfully implement this amplifier or similar to allow this with the use of a microcontoller /ADC for further tasks.

Thanks for any insight provided!

Jake

  • 0
    TI__Mastermind 25805 points

    Hello Jake,

    Can you please post your schematic?  Also, can you please give us more information about the load cell (e.g. part number and/or data sheet)?

  • 0 in reply to Pete%20Semig
    Prodigy 140 points

    Pete,

    This is a prototype project.  The load cell is surplus from a scale, so I don't have any details.  It's made of two 3 lead strain gauges bridged into a 4 pin output of E-,S-,S+,E+.  the load cell was designed for a 2kg scale, and this is in the range we are interested in. The schematic I used is seen below.

  • 0 in reply to Jake Vonderharr
    Expert 1495 points

    Hello Jake,

    You first need to establish the load cell sensitivity. Typical scale load cells have a sensitivity of 2mV/V/ xx kg. For example, in case of a 2mV/V/10kg load cell, an excitation of 10V will produce a full scale output of 20mV.

    The load cell sensitivity can be measured experimentally using known excitation and known weights by a precision multimeter directly at its output. Once this is known, the amplifier gain can be chosen to suit the requirements of the analog to digital converter.

    You have been able to achieve 3mV/g output. I recon this includes the gain of the amplifier. I would suggest that you prepare a chart of the load cell output at various loads, without the amplifier connected. This will help you establish the input characteristics for your system. After ensuring that the load cell  output is linear throughout the range of interest, the gain of the amplifier may be adjusted to achieve the desired output.

    In case your load cell is damaged, the output may not follow a linear relation.

    Hope this helps..

    Jayant

  • 0 in reply to Jayant Deshpande
    Prodigy 140 points

    Hi Jayant,

    I followed your advice and experimented with some known weight values.  It appears that at an excitation of 3V, the cell shows a trend in change of 1mV/1kg of weight.  I would like to see a voltage of around 2V at that weight amplified, so would we be looking at a gain of 2000 then? This looks like ~25ohm for the INA128.  I had already set a gain around that point and didn't see the sensitivity I needed.  Would increasing the voltage to the amplifier increase my available voltage for finer increments?  I don't really want to do that...

    Thinking about voltage increments.  If I need 0.005mV of movement to sense for an increment of 1g (assuming a resolution of 1g), my ADC would likely need to be 12 bit to encompass more than 1024 grams for a 10 bit right?

  • 0 in reply to Jake Vonderharr
    Expert 1495 points

    Hi Jake,

    A gain of 2000 is too high. You will have all noise amplified 2000 times and at low inputs you will have difficulties separating the signal from the noise. I would recommend  higher value of excitation, with higher signal amplitude.

    More details about your application would help. Why do you need 2V at 1kg ? What is the significance of 5 microvolt for 1g increment ?

    Regards,

    Jayant

  • 0 in reply to Jayant Deshpande
    Prodigy 140 points

    Jayant,

    The Atmega328P IC can only sense at the minimum 5mV.  So if I were to put a 1g weight on the scale and it only returned a change in voltage of less than 5mV, the scale would still read zero.  I need to amplify it enough to sense that 1 gram, but also be able to read through a sweep of 2000g.

    Thanks,

    Jake

  • 0 in reply to Jake Vonderharr
    Expert 1495 points

    Hey Jake,

    It is not possible to make a 2000 count measurement using a 10 bit ADC. Seems you need to change your design approach.

    All the best,

    Jayant

  • 0 in reply to Jayant Deshpande
    Prodigy 140 points

    Jayant,

    So with the previous load cell outputs, is the INA128 the best choice?  Is there another option that may be better for this application?  I am still having trouble getting both the sensitivity and sweep dialed in with low noise affecting the amplified output.

    Thanks,

    Jake

  • 0 in reply to Jake Vonderharr
    Expert 1495 points

    Jake,

    The INA128 is fine.. you need a better ADC. Have you read http://www.ti.com/lit/wp/sbaa154/sbaa154.pdf ? It explains the signal conditioning requirements for load cells.

    Regards,

    Jayant