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.

High precision differential amplifier 50 kHz

Other Parts Discussed in Thread: LM833

Hi,

i need a differential amplifier with a very accurate gain of 2 at a frequency of 50 kHz (1.9999<gain<2). When assuming ideal resistors, my calculation leads to opamps with a GBP greater than 15 MHz. For a test i chose the LM833, even though its minimum GBP is too small. I have attached an Image of the acutal test board along with a simple schematic.

The circuit behaves in a way i did not expect. At 50 kHz the gain is too big. Its round about 2.04 instead of 2.

The resistors are high precision and a measurement with a preciscion impedance analyser results in the correct resistances.

At low frequencies (<10kHz) the gain is much closer to 2.

At frequencies above 50 kHz the gain rises even further and than falls at about 150 khz.

I would expect the frequency response to be much more flat and the amplifier to be much more accurate at 50 kHz.

Any help would be greatly appreciated!

Kind regards,

Markus Schyma

  • Additional infomation:

    Input Signal about 2 V peak, "pure" sine wave.

  • Markus,

    Gain bandwidth is typically 16MHz, At 50kHz op-amp gain will be about 320 [16MHz/50kHz] with a 90 degree phase shift. The loop gain is R3/R1+1 = 3. With just a [320/3] 107 'correction gain', a 1% to 2% error is not surprising.

  • Hi,

    thank you for the quick Response.

    I have calculated the closed loop gain according to the following equation:

    where:    α=2/3, β=1/3, AOL=90db=31622, f=50kHz and f0=200 Hz

    The closed Loop gain should therefore be at least 1.9992.

    I don't understand where the error is.

    Kind regards,

    Markus Schyma

  • Markus,

    I simulated using your numbers and the frequency response was indeed flat to 50kHz.
    Therefore I now suspect that the resistor divider is good, but the capacitance in the circuit is not frequency balanced. Even a few pf of mismatch could cause an error at 50kHz.

  • Hi,

    i had already feared it is a problem with parasitic capacitances, that's why i build it using a piece of fr4 with no ground plane paired with a relatively large setup.

    If it is so sensitive to parastics, how can i ensure the accuracy i need?

    May it be caused by parastics within the amplifier itself so that another one would do better?....but what to look for in the datasheet?

    Kind regards,

    Markus Schyma

  • Markus,

    The "Differential input capacitance" spec is typically 12pF. Op-amps with lower cap spec should be helpful. Lowering the resistor values will also help. 

    As another thought; is the output also pure sine wave? Amplifier non-linearity (distortion) could affect results.


  • Markus,

    First, make sure your measuring equipment has flat response.  Use a high-quality sine or square wave generator as a signal source.  Don't forget to calibrate your scope probe.

    Second, input a square wave into your circuit.  Does the output ring?  If so, the input capacitance of the op amp may be giving an underdamped response and frequency peaking.  You can try two things to fix this:  reduce the R values, or add a small C across the feedback resistor.  If the input C is the problem, you may see unacceptable unit-to-unit variations.  The best way to solve this is probably to reduce the R values.

    Another thing that can cause problems is poor power supply bypassing.