Instrument Amplifier - Temperature Drift

Discussion in 'General Electronics Chat' started by Digital_Mike, Feb 5, 2010.

  1. Digital_Mike

    Thread Starter New Member

    Mar 6, 2008
    I am using an instrument amplifier Identical to the one shown in this link:

    All of the resistors are 10K with the exception of the gain resistor which is 150ohm. If these resistors hold a tight tolerance that should provide me with a gain of ~134.33333. Great! that being said.. my input signal is in the range of 0 to 24mV, and i'm expecting an output from 0.7 to 4.096. The input is coming from a Pressure sensor that is excited with ~12V, and the mV output is biased around 6.000 - 6.020. Every thing seems to work fine, however, when i change the temperature of the device the output starts to change. I'm not sure what is the main cause for this drift. I know the resistors have a Temperature Coefficient of Resistance (TCR) of 50ppm/C and are meant for high temperature applications up to 200C. and I am using a High temperature Version of the OP284 by Analog Devices.

    I am getting the output i am expecting, at ambient temperature. it just varies when temperature changes. Does any body know Which components would be causing this to happen?

    I know the regulator which provides Excitation Drifts from 12.080V to 12.025V over the range of 0-60C. This regulator is also driving the Opamps.

    I also know the resistors do not have a high TCR at only 50ppm/degC.

    I'm not sure about the input impedance and output impedance of the OpAmps.

    Has anybody tried this before and experienced the same problem?

    When I use an Stable external Voltage source to excite my sensors, then ramp the temperature of the board, i get the same result. (Result being a shift in output voltage).

    I'd be open to suggestions on what may help make this circuit less suseptable to variations with temperature.

  2. Blackbull

    Well-Known Member

    Jul 26, 2008
    I have had problems with thermocouples caused by joints (wires and plugs) leading to the input pins, joints have to be kept to a minimum, but then perhaps this is not your problem.
  3. t_n_k

    AAC Fanatic!

    Mar 6, 2009
    Have you quantified the drift? You've said it changes, but it's not clear to what extent.
  4. someonesdad

    Senior Member

    Jul 7, 2009
    Blackbull's suggestion is a good one that might not be obvious. One technique is to use a source of heating or cooling on individual components to see if you can spot which component in the circuit exhibits the most sensitivity to temperature. This can be done with a can of freeze spray or a soldering iron. If there are multiple op amps in a single package, however, you may need to use separate ICs to be able to know which op amp it is (assuming it's one of them).

    Once the thermal culprit is found, one technique to fix it is to offset it with a component that responds in the opposite way to temperature. Another technique is to build the whole thing in a thermally insulated enclosure and keep it at an elevated temperature (this was done e.g. in older HP differential voltmeters).
  5. Duane P Wetick

    AAC Fanatic!

    Apr 23, 2009
    What is the tolerance of the 10K resistors (R)? I have seen drfit problems in differential amps when these values are not closely matched (.01%).
    Try selecting values within this tolerance if you cannot get .01% units.

    Cheers, DPW [ Everything has limitations...and I hate limitations.]
  6. ifixit

    AAC Fanatic!

    Nov 20, 2008
    What opamp are you using?

    The input offset of the opamps can change with temp. Look in the spec to see if it would be significant over the 60C temp range.
  7. creakndale

    Active Member

    Mar 13, 2009
    It is virtually impossible to build a discrete instrumentation amplifier that is immune to temperature drifts. Google:
    "The Instrumentation Amplifier Handbook"
    by Neil P. Albaugh.

    See Page 24 of the PDF document: Internal vs. External Gain Setting Resistors

  8. russ_hensel

    AAC Fanatic!

    Jan 11, 2009
    A bit off the point, but if you excite with ac then you can filter out the drift in offset. Of course at some point you will have to rectify the signal. And some sensors do not excite with AC in a nice way.