Noisy Instrumentation Amplifier INA129

Discussion in 'General Electronics Chat' started by Efry, Feb 22, 2009.

  1. Efry

    Thread Starter New Member

    Feb 22, 2009
    6
    0
    Has anyone had problems with these chips having an incredible amount of noise? For a circuit gain of 100 and a .089V DC signal generated by a 9V battery and a voltage divider, the output from the amp is really noisy. I measured it with my NI 6008 USB DAQ and I've attached a plot of about 1 secs of data I measured.

    Any help or input with these intrumentation Amplifiers would be appreciated.

    Thanks,

    Evan
     
  2. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Post the circuit to clear things up.
     
  3. Efry

    Thread Starter New Member

    Feb 22, 2009
    6
    0
    Mike,

    Thanks for your interest. The circuit I'm testing is attached in the jpg.
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Looks to me like you have your Vcc and Vee supplies reversed in polarity.
    That would not only make the output noisy, but smoky too. :eek:

    V2 has no ground reference; it's "floating".
     
  5. leftyretro

    Active Member

    Nov 25, 2008
    394
    2
    I though an instrumentation differential input op-amp didn't need to be referenced to the input sources common, that's one of it's biggest advantages, isn't it?

    Lefty
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    While the input voltage range exceeds the supply voltage range, it's still not infinite.

    Absolute max supply voltages: +18v/-18v
    Absolute max analog input voltages: +40v/-40v

    Our OP needs to include a path to ground in the input. Spice doesn't "like" floating nodes.
     
  7. Efry

    Thread Starter New Member

    Feb 22, 2009
    6
    0
    SGTWOOKIE,

    I think I wired the amp so that -12V goes to the top of the amp in the figure and +12v goes to the + input at the bottom of the block. V2 is a battery so when I try to tie it to ground in the TINA TI software it gives me an error message.
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    OK. I couldn't see the "+" on the bottom of the INA129 at the magnification of the schematic you posted.
    Try tying the - side of V2 to ground using a 1M resistor (1 Megohm). I simulated it using Tina-TI v7.

    When running the Transient Analysis, select "Zero initial values".
     
  9. Efry

    Thread Starter New Member

    Feb 22, 2009
    6
    0
    I tried adding the resistor and tying the v2 - pin to ground with it, but the resulting voltage is about half what it should be.

    Either way though, the computer analysis doesn't show a noisy signal when I do it without tying the v2 to ground, which makes me wonder if the TINA analysis is really accurate to my actual circuit. I am powering the amp with this power supply I have and the power is pretty clean, but do you think some caps on the vcc and vee pins would help reduce the noise?
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You should always use bypass capacitors to ground on Vcc/Vee pins.
    See the datasheet; in the Application Information section, it shows both pins bypassed to ground using 0.1uF non-polarized caps.
    Bypass caps won't make a difference in the simulation though, because the "wires" in the simulator are ideal conductors; 0 Ohms, no inductance or stray capacitance, and capable of carrying near-infinite current.

    Are you using metal film resistors? They are much less noisy than carbon or carbon film resistors.
     
  11. Efry

    Thread Starter New Member

    Feb 22, 2009
    6
    0
    I am using 1/8W metal resistors from NTE. I'll try the caps in the real circuit and see if that does anything. In the meantime I came across this article about Instrument Amplifiers with Indirect Current Feedback Architecture. Have you had any experience with these?


    http://www.en-genius.net/includes/files/acqt_052608.pdf
     
  12. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Let's go back to your circuit for a bit first.
    You say that you want a gain of 100. In that case, your R1 should be 499 Ohms instead of 461 Ohms. See the datasheet for the resistor selection table under "Application Information".

    Your resistive divider for the input creates an approximate 89.1mV differential across the instrumentation amplifier's input. (100 / (10k+100)) * 9v = 89.10891...mV

    In the simulation I put together, when I connected V3's negative terminal to ground via a 10M resistor, changed the 461 Ohm resistor to 499 Ohms and ran transient analysis, I obtained an output of approximately 8.91v with no visible noise. (Note that voltmeter V2 is inverted so that output polarity display is correct).
    [​IMG]

    The noise you're seeing on your ADC may be a grounding problem.
     
  13. Efry

    Thread Starter New Member

    Feb 22, 2009
    6
    0
    Thanks,

    I'll give that a try tonight when I get home. Yeah, the gain is more like 108, but I was just saying 100 for simplicity on that first post. I was using the 460 Ohm resistor because it was the closest I had at home.

    Evan
     
Loading...