• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 46 subscribers
  • Views 546 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

ADS7253: ADS7253 Signal oscillates during A/D conversion with Anti-aliasing filte

Eric QI
Intellectual 620 points
Part Number: ADS7253
Other Parts Discussed in Thread: OPA836, OPA2836, INA240

Hi,

Now i'm using ADS7253 for A/D conversion.

But the digital data after ADS7253 is distorted.I measured the analog signal of C17 by an oscilloscope.It's OK.

It is verified that the signal conversion fault is caused by the filte,especially anti-aliasing filte.

So my question is,

1、Please help to confirm if there is a problem with my operational amplifier part design（OPA836,shown below）.

2、Increasing resistance values (R30,R31) makes signal conversion more stable，how to choose this resistance value?

3、The rated frequency of my analog signal input is 20 KHz, please help solve this problem.

Thanks.

  • 0
    Intellectual 620 points

    To add, the following figure shows the digital signal read by the host computer.

    The second half of the sine wave is seriously disturbed, especially in the sine wave valley.Oddly enough, the signal in the part of the sine wave peak is very small.

  • 0 in reply to Eric QI
    TI__Guru 63995 points

    Hi Eric,

    A few comments.

    First, the procedure in the TI design your referencing chooses the proper R/C charge-bucket filter not as an anti-aliasing filter, but as part of the ADC drive circuit. Use the R52/C47 combination as your anti-aliasing filter, and use the R30/R31/C17 combination as the ADC charge-bucket drive filter.  The theory behind this is covered in depth in the TIPL videos below:
    https://training.ti.com/node/1139106?context=1139747-1140267-1128375-1139106

    The ringing is a classic op amp stability issue while driving the 22nF capacitive load.  The current 10Ohms of "isolation resistance" between the amplifier output and the capacitor is not sufficient to stabilize the circuit.  You could:

    A.) reduce the 22nF load to something closer to the 400pF recommended in the design guide, or
    B.) you'll have to increase the series resistance until the op amp is stable at the cost of distortion, or
    C.) you'll have to choose a different amplifier that can drive those types of capacitive loads directly.

  • 0 in reply to Collin Wells
    Intellectual 620 points

    Hi,colllin

    I'm sorry to reply after a long time.

    It is proved that the oscillation is caused by the large capacitive load driven by the operational amplifie of OPA2836.

    Cut-off frequency needs to be set at 100KHz. R52 and C57 seem to bring system instability because of INA240A1.

    I want to know how to improve this,whether or not the operational amplifier needs to be re-selected for it can drive

    large capacitive loads.   Or does the topology need to be redesigned?

    Thanks.

  • 0 in reply to Eric QI
    TI__Guru 63995 points

    Hi Eric,

    It has been a while, I hope you've been well.

    If the instability is starting at the output of the INA240 we'll need to fix that first or it will propagate through the following stages.

    Try decreasing the capacitor to 2.2nF and the resistor to 1k.  It will maintain the same cutoff frequency but will probably cure the issue in that stage.  You may need to slightly adjust the OPA836 output circuit values as well after stabilizing the INA240 circuit.  

  • 0 in reply to Collin Wells
    Intellectual 620 points

    Hi,collin

    I have a new problem.The actual output voltage range of the operational amplifier (OPA2836) is too narrow.

    The actual output high voltage of OPA2836 is about 4.2V.According to datasheet of OPA2836,output voltage

    range of operational amplifier is affected by saturation current.So how to limit the saturation current?

    Resistance of R30 and R31 should not be too large to prevent output oscillation.

  • 0 in reply to Eric QI
    TI__Guru 63995 points

    Hello,

    The situation you're experiencing is not an output limitation, but rather an input common-mode range limitation.  When you operate an amplifier as a unity-gain buffer (Vin- shorted to Vout), the input voltage applied to the IN+ pin is the common-mode voltage.  Take a look at the "Common Mode Input Range High" spec in the datasheet on page 12 for a +5V supply.  It's only 3.9V, so the fact you're getting 4.2V is a little above the expectation.  Either choose a rail-to-rail input op amp, switch to an inverting configuration, or put the amplifier in a gain of 2 and the common-mode voltage will remain in range for a 0-5V output voltage.

    Also, for the 2nd channel of the OPA836 connect VIN- to VOUT2 and then connect VIN2+ to GND to prevent unwanted operations from floating inputs.