Force transducer amplified with INA125

Discussion in 'The Projects Forum' started by silhas, Oct 15, 2013.

  1. silhas

    Thread Starter New Member

    Oct 15, 2013
    3
    0
    Hi there,

    I'm trying to learn a bit about sensors, and I'm currently stuck on the force transducer. I've got hold on a 2kN S9M (0,5 - 12 V Usupply) (http://www.hbm.com/fileadmin/mediapool/hbmdoc/technical/a3019.pdf)
    and I've tried two different approaches.

    1. With a INA 125 (http://www.ti.com/lit/ds/symlink/ina125.pdf):
    An instrumentation amplifier seems to be popular to boost the signal from a bridge circuit. A friend of mine composed this wiring diagram (attachement 1) but the result just seems to be random voltage output between +/- 10mV.

    2. Without an instrumentation amplifier:
    I then thought about just plugging it in and see the respons.
    -Black wire (excitation voltage (-)) to the - pool on the cell
    -Blue wire (excitation voltage (+)) to the + pool on the cell
    -White wire (Measurement signal (+)) positive output
    - Red wire (Measurement signal (-)) negative output output

    I've only got access to a 24V cell so I voltage divide it with two resistors (1K ohm each, might be way to low tho...).
    The result is a standstill on 1960mV, and the voltage over the force transducer is 4V.
    The force transducer should transmit 2 mV/V...

    I've attached both wiring diagrams. At the moment I don't really know what to do, any pointers to what might be wrong with the wiring diagrams would be a tremendious help.

    Regards
     
  2. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    According to the datasheet (INA125):

    Pin 2 (the sleep pin) should be connected to V+.
    You may need to buffer VrefOUT to drive the load cell (Fig. 4).
    Review Fig 5. and the description of single supply operation to make sure you are not busting common mode voltages etc.

    Why do you need to add a voltage divider? The top of the bridge is to be driven with VrefOUT, buffered if necessary.

    Have fun.
     
  3. silhas

    Thread Starter New Member

    Oct 15, 2013
    3
    0
    Thanks JohnInTX for the quick reply.

    Just to clear the misscommunication: The voltage divider is only for solution 2. I'll try to correct the spacing to avoid further missunderstandings.

    I'll attach PIN2 to my Usupply, that's a good start.

    If I understand this correctly a transistor will reduce the reference noice. Since I'm using Vref2.5 it shouldn't be to much of a problem ( "Reference noise is proportional to the reference voltage selected. With V REF = 2.5V, 0.1Hz to 10Hz peak-to-peak noise is approximately 9mVp-p." ). Also since I log all my data, souldn't I be able to filter it in a later stage?

    But in a learning aspect, how do I choose transistor? (I'm using Vref2.5)​

    Unfortunately I don't really understand this part, I've grounded PIN 3, 5 and 12. Aslong as I have a load between the Usupply and ground I should be safe?​
     
  4. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    An external transistor would be needed if the internal reference amplifier (pin 4 output) did not have enough current capacity to drive the bridge. Even if it does, the datasheet points out that external drive would reduce the temperature of the chip for less temperature effect on the internal reference. The TIP29 shown is a reasonable choice. Anything with the voltage and current capacity would do the trick. Characterized for low noise would be good as well. That said, its optional if the bridge voltage == Vref without it.

    As far as filtering - yup. Most load-cell stuff I've seen/designed incorporate some sort of RF filtering at the input to the IA. Some degree of low pass filtering is also incorporated into the firmware as well.


    You've grounded the correct pins but there are some other issues. (ref page 12 of the datasheet)

    First, the voltage on the +/- inputs must each be >1V above ground to get the amplifiers into their linear zone. If you are exciting the bridge with 2.5 volts, your null voltage on each leg will be 1.25V. This is OK but be sure that under full load, neither leg's input drops below 1V. Raise Vref if necessary.

    Second, and more problematic, is that the amplifier's output can only swing to within about 150mv of the V- rail. With dual supplies, this is not a problem but since V- is now ground, your minimum output voltage will be 150mv. Since you want a balanced bridge to read 0.000V out, this will be a problem. In effect, you chop off 150mv worth of measurement near 0. You might get away with it if the load cell (force transducer) has a built-in offset so that zero load is some voltage difference but how much usually falls into the category of manufacturing tolerance and, unless specified, should not be relied on. You can specify some offset for just these cases when designing the load-cell and there are IAs that are specified for single supply operation that will output a true 0V.

    Make sure your power and grounding are low impedance (including VrefCOM) and that the amp power is well decoupled.

    Good luck!

    EDIT: The Designer's Guide to Instrumentation Amplifiers from Analog Devices covers all these points and more in detail. Chapter 5 is pertinent. TI/Burr-Brown probably has lots of useful stuff as well; I'm more familiar with AD.
     
    Last edited: Oct 16, 2013
  5. silhas

    Thread Starter New Member

    Oct 15, 2013
    3
    0
    Thanks a bunch!

    Got some material to read up on, hopefully it'll work out :)
     
  6. JohnInTX

    Moderator

    Jun 26, 2012
    2,348
    1,029
    Of course it will work out. Go get 'em, tiger.
     
Loading...