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.

THS4521 into ADS1274 DC Level Shift

Other Parts Discussed in Thread: THS4521, ADS1274

We have a design with 3 THS4521 into an ADS1274, and we are seeing a digital shift in signal on about 3% of our production boards. We have not been able to narrow down the cause.  I've posted with this a schematic we've used for simulation for one channel, and raw samples plotted showing the issue.  The red channel is the one having the level shifts, and we have seen it have more than one level it shifts between.    It is not consistent which channel has the issue, as other boards have had the issue on the blue channel (the high frequency noise on the blue channel is from testing with open boards).  We've tried replacing the diff op amp multiple times, replaced capacitors, taken from one channel to another, wired over from power and ground of other channels, and we still cannot find out the cause of the discontinuities.  The shifts cause low frequency noise, and our application uses frequencies from 0.4 Hz to 2 kHz.  We have found that heat via a soldering iron on the pins of the part cause the issue to happen more consistently for a while, with it getting better over time, but still intermittently happening.

Any help would be appreciated. 

  • Hi Alex,

    I will need some time to simulate this circuit before I give you an answer. I should be able to look at it by the end of the week. Active filters can be complex.

    Can you try the circuit without C21 and C23. Are there really two different capacitors in these spots? They should be the same.

    Regards,
    Loren
  • Hi Alex,

    Can you confirm your schematic component values? The values of the resistors are not symmetric on the feedback path. When I simulate this circuit I get almost zero gain. What is the desired response for this circuit? Have you simulated it?

    Regards,
    Loren
  • Loren,

          The component values are what we are using, other than the 1G resistor we have left open.   I have simulated this circuit, we needed attenuation in the circuit, going from a single ended signal in to the differential to the ADC.    In order to meet input range requirements we had to change values from the initial design where we had matched feedback resistors, and used simulation to find a sweet spot that meets our input voltage range, 2-17 V in, but still gives us the low range linearity we need.  The circuit has worked pretty well, we just have the intermittent issue on a small number of boards.   The issue seems to be temperature related, and there may be a possibility that R156 or R18 may have been damaged.   The shift we see in the DC corresponds to only a 300 uV DC shift on the input, so if the resistance R18 or R156 changed by just a little due to a bad connection or damaged component, could that be causing the shifts we are seeing?

    Thanks,

    --Alex  

  • Hi Alex,

    I have been out of the office the last few days and I still have not had time to evaluate your circuit. 

    Some amplifier parameters like CMRR and second harmonic distortion are totally dependent on circuit symmetry.  Also, a floating input is an undefined operating state.  It is very hard to predict why a circuit like this works, let alone what could go wrong. 

    If you give me some more details of your application I may be able to propose a symmetrical circuit that has both inputs terminated.

    Regards,

    Loren

  • Loren, 

              Would there be any chance we could set up a phone call to discuss specifics of our application, there are quite a few variables involved that may be easier discussed via phone.

    Thanks,

    --Alex

  • Hi Alex,

    That may be possible. Are you working with a TI sales representative or FAE?

    Regards,
    Loren
  • Loren,
    We are not working with any right at the moment, I was looking for a contact to discuss the issue over the phone and got directed to just post my question to the forums instead.

    Thanks,
    --Alex
  • Hi Alex,

    I still don't have a good explanation for the DC shifts you are seeing. I have done some circuit simulations, the circuit looks like it could have problems with DC levels. Have you probed the amplifier pins to check their DC voltages compared to the datasheet ranges?

    E2E works well for many issues, but maybe not this one. I am working with some colleuges to arrange a call. Your applicaion is difficult because of the floating input and the assymetric feedback network. I am concerned that if we do fix or at least diagnose the observed issue the circuit will still be subject to issues in the future. A floatint input on the amplifier is an undefined state. It will naturally float to a value, but the mechanism by which it comes to equilibriam is tenuous at best. Things like input bias currents and external noise will have significatnly more impact on your circuit than a more traditional setup.

    The assymetric feedback network is less problematic, but still falls outside of what we would recommend. Common mode rejection and second order distortion performance will both be seriously compromised by this feedback network. A floating input will further complicate this design.

    Regards,
    Loren