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.

TLC2272A: Buffer follower not following

Part Number: TLC2272A
Other Parts Discussed in Thread: TLC2272,

I am investigating a production issue for a legacy design where a tlc2272acpwr opamp is configured in a buffer follower configuration with the output tied to - and the + input is driven by a another opamp at -2.0V in series with two 33 megohm resistors in series. The voltage at the + input is -0.29V. The voltage rails are at +5 and -5V. The output connects through 23.7k to an ADC input. On passing boards the ouput of the tlc2272 and the input of the ADC is -1.48V. The test limit requires ADC input is no lower than -1.65V. On failing boards the voltage is -1.66 to -1.69V. The connections between the components in the circuit have been verified and the components have been swapped but the failing boards continue to fail and no incorrect connections have been found. The circuit was simulated and voltage on the input matched the output over the range the circuit operates. When the input of the circuit goes high (3.6V) the output matches it and the simulation.

Are there conditions where the opamp won't behave like an opamp and a buffer follower output won't follow the input?  

  • Justin,

    First of all, your TLC2272A circuit works in the simulation as it should - in a buffer configuration output follows the input input until input common-mode voltage range limits it at +4V - see below.

    However, something does not make sense in your description:  how on a passing boards the output of the TLC2272 and the input of the ADC can be at  -1.48V while you drive the input of a buffer to -2V - they all should be at -2V as shown below.  

    Thus, I wonder what supplies do you use for your ADC?  Most ADC's do NOT work on +/-5V.

    Your decription of the failing board would make me believe that you have a damaged ESD protection diode between the Vinp and positive rail because -.29V at Vinp and Vin at -2V would require 51nA input current to flow thru 33M input resistor - see below. But since even a "passing" board does not make sense, I can't be sure what else is going on?  Please use the attached schematic instead of write up to show what is going on - a picture is worth thousand word.

    Justin circuit TLC2272A.TSC

  • Hi Justin,

    yes, please show us a schematic. Your description is really hard to read and, at the end, makes no sense...

    Kai
  • Here is the circuits connected to the buffer (J29). The RLF signal is -0.289V on a passing board and RLB is -1.485V. On a failing board RLF is -0.292V and RLB is -1.69V. RLF is driven by VLOV, which is -2.013V on passing and failing boards. RLF can be driven by analog switch J21 but under the test condition it isn't.

  • Hi Justin,

    I cannot see a schematic.

    Kai
  • Hi Justin,

    what is connected to X17?

    Kai
  • Hi Kai,

    The test calls X17 to be disconnected and other inputs on channels like the one partially displayed above X17 to be tied to ground.

    Justin
  • Hi Justin,

    the measurement values (-0.289V) make only sense, if it assumed that the voltmeter has an impedance of about 10M and forms a voltage divider in combination with R88, R263, R78 and R68. And there must flow a leakage current of more than (2.013V - 1.485V) / (66M + 21.5k + 21.5k) = 8nA. As the maximum input bias current of TLC2272A is well below 800p, this leakage current must have a different source. I guess, it has to do with the diode clamp circuitry. This makes sense, as there exists a current path of three pn-junctions in series to signal ground. It's very easy for a voltage of 2.013V to partially turn them on and cause a leakage current of more than 8nA.

    So, the issue does not seem to have anything to do with the TLC2272A itself, but with the diode clampling circuitry.

    Kai
  • Hi Kai,

    I had wondered if the clamp diodes starting to turn on or board contamination could account for small differences at the input.
    It looks like you're saying that the meter impedance is changing the voltage at the input (RLF) and the buffer follower is actually following but I can't see it because when I attach the meter because this changes the voltage.
    The production test is actually performed without a meter and the voltage is measured at the ADC further downstream (RLB connects through an RC lowpass filter).
    This points to the clamp circuit, the RF filter or contamination somewhere between VLOV and RLF.
    I can try having the parts swapped between good and bad boards and look into cleaning options.

    Justin
  • Justin

    We haven't heard back from you so we assume this resolved your issue. If not, just post another reply below.

    Thanks
    Dennis
  • Hi Justin,

    I would try to increase the clamp voltages +CLMP and -CLMP. This could be accomplished by using three diodes in series for V19 and V25 instead of only two. Or use the turn-on voltage of a LED, if you don't want to move too far away from 2 x 0.6V = 1.2V.

    Kai
  • Justin

    Did this resolve your issue? If not post another reply below.

    Thanks
    Dennis