• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 47 subscribers
  • Views 566 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.

TLV9062: How to minimize voltage spikes with TLV9062 in buffer mode?

Vincent Maurice
Prodigy 40 points
Part Number: TLV9062
Other Parts Discussed in Thread: OPA2374

Tool/software:

Hello,

I am currently working on a project where I am using the TLV9062 operational amplifier in a buffer configuration to read a voltage pulse with an MCU. The input pulse ranges from 0 to 2V, with a duration between 50 µs and 300 µs, at a frequency of 1 Hz.

To allow the MCU enough time to read the output voltage, I have added a 100 nF capacitive load at the output of the op-amp. However, I am experiencing significant voltage spikes at the output, particularly near 0V, while the spikes are less pronounced near 2V.

To improve stability, I added a resistor in the feedback loop, which slightly reduced the spikes, but the results are still not optimal. The spikes are too high near 0V and weaker near 2V.

When testing with the OPA2374, I noticed that the voltage spikes are much less pronounced. However, I understand that Texas Instruments recommends using the TLV9062 for new designs, and I would prefer to use it if possible.

Could you advise on the optimal circuit design to avoid these voltage spikes in my case? Is there a better configuration to stabilize the buffer (e.g., resistor values or additional components)? Alternatively, I am open to suggestions for other op-amp references that may be suitable for my application.

Thank you for your help!

  • 0
    TI__Guru**** 167546 points

    Hello Vincent, welcome to the TI e2e forums.

    You need a peak detector circuit to hold the peak input voltage for a time until MCU does a read. How does the MCU know when it is time to read? Is the 1Hz dependable enough when input pulses may be close to 0V which would look like a missing clock pulses.

  • 0 in reply to Ron Michallick
    Prodigy 40 points

    Hello Ron,

    Thank you for your feedback.

    To be more precise, I have a square wave signal with a pulse every second, lasting between 50 and 300 µs, with voltage levels ranging from 0.2V to 2V. The circuit stays at 0V the rest of the time.

    I chose to use an amplifier because the signal polarities can be reversed. Therefore, I included both a buffer and an inverter mode. For now, I am focusing on improving the reliability of the buffer mode.

    I periodically wake up the MCU to take measurements with the ADC. I continuously measure until the signal reaches zero, and as soon as a measurement exceeds a certain threshold, I capture the first value, since the signal rises very quickly. With this approach, I could even reduce the value of the charge capacitor at the output of the operational amplifier.

    However, when I simulate the Op-Amp circuit with a 50 µs pulse in LTspice, if I don't place a resistor in the feedback loop, for an input voltage of 1V, I get an output voltage of 1.6V. To fix this, I added a resistor, which solves the problem for a 1V input. But when the input voltage is higher, the output voltage becomes lower than the input voltage. I get the same result in physical tests.

  • 0 in reply to Vincent Maurice
    TI__Guru**** 167546 points

    Vincent,

    What I hear is that input is 0V most of he time. Once a second, the voltage changes to a voltage of [-2V to -0.2V] or [0.2V to +2V] and remains constant for at least 50us. I'm I understanding correctly? 

    For that I'd suggest an amplifier that converts [-2V to +2V] to [just under ADC range] and a window comparator [-0.1V to +0.1V] that generates an wake on interrupt signal. When window comp reports outside, do an ADC read. I assume you can wake and read in under 50us.

    What is the supply voltage, ADC range and ADC reference voltage?

  • 0 in reply to Ron Michallick
    Prodigy 40 points

    Yes, you understood correctly.
    The board's supply voltage is 3V, and the ADC is 12-bit (0 to 3V) with a reference voltage of 3V.
    For clarification, the pulse current comes from another external circuit.

  • +1 in reply to Vincent Maurice
    TI__Guru**** 167546 points

    Vincent,

    Lightly simulated it seems OK.

    ADC voltage with 0V input  = 1.496 (half of 3V) "b".

    Gain is -10k/14.7k "m"

    So ADC = VIN * m + b or ADC = VIN * -0.6803 + 1.497V

    When "Wake" goes low wake up and read ADC, then MCU can go back to sleep.

     

    Pulse scale and wakeup.TSC

  • 0 in reply to Ron Michallick
    Prodigy 40 points

    Thank you Ron for your kind reply.
    I will study and test your solution.
    Thank you for your help

  • 0 in reply to Ron Michallick
    Prodigy 40 points

    Hello Ron,

    After several tests with the simulator, the amplifier successfully tracks the input signal. However, I need a peak detection circuit to capture and hold the maximum voltage so the MCU can read it before it decays. The rise in input current is not constant, it can rise in two steps, and I want to avoid false negatives in case of intermediate current levels.

    How can I modify the schematic to achieve this?
    Thanks

  • 0 in reply to Vincent Maurice
    TI__Guru**** 167546 points

    Vincent,

    I put together two circuits to take a absolute value of a voltage then do a peak detection. The MCU should be able to reset the low leakage holding capacitor. Check the High-Z out leakage to gauge how much peak loss comes from the reset.

    ABS+PEAK+EXT discharge.TSC