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.

ADS1115: ADS1115 failing with no apparent misuse

Part Number: ADS1115

Hello,

I am using an ADS1115 A/D in a test application (details described below).  Twice now, this part has failed, but (so far) I have not found that I misusing the part in any way.  Unfortunately, I don't have exact information about how long it took the part to fail, but it's safe to assume it was running continuously for at least 40 hours.

Application conditions:

  • VDD = 5VDC
  • Connected to microcontroller via I2C bus.
  • Analog inputs are connected as follows:
    • AIN0:  Connected to 3.3VDC reference voltage
    • AIN1:  Connected to a 0.0V to 3.3V, ramp-shaped signal, with a very slow period (about 1 minute).
    • AIN2:  Connected to 2.5VDC reference voltage.
    • AIN3:  Connected to an analog signal, output from a device-under-test.  0V to 5V.
  • All measurements are single-ended, no differential measurements.
  • Measurements for signals connected to AIN0, AIN2, and AIN4 are single-shot measurements, PGA gain set to 2/3.
  • Measurement for signal connected to AIN1 is a continuous-mode measurement, PGA gain set to 1, sample rate set to 860 samples per second (maximum sample rate for this part).

Failure mode:

  • Device still communicates over I2C and reports A/D readings, but A/D readings are off by about 100-200 counts.  All four channels are reporting this same drift.
  • Device is much hotter than normal, as in, too hot to touch.  Spraying the IC with freeze spray causes the A/D readings to temporarily return to its former accuracy, but it quickly heats up and drifts again.

I'd like to know if you have any suggestions, or if you have seen anything like this before.  Would you expect any problems with continuously running the part at its maximum sample rate?  I didn't see anything in the ADS111x datasheet that said this would be a problem.  Any information or insight you can provide would be helpful.

Thanks very much!

Regards,

-Michael

  • Hi Michael,

    Welcome to the E2E forum!  The ADS1115 is a very popular device and is used in may different types of applications with no issues.  It would be helpful for me to see your schematic.  The ADS1115 is considered a low-power ADC, so if the device is getting hot, then I suspect that there is an Absolute Maximum Ratings violation.  The Absolute Maximum Ratings are given in the device datasheet in section 7.1.

    The most common issue I have seen with single-ended measurement are related to the ground connections of the various input voltages.  You should verify that the voltages at the input are correct by measuring each input voltage relative to the ADC ground.  The first thing I would check is the signal source providing the ramp.  This signal source output must stay within the range of the supply.  If the ramp goes below ground slightly or above AVDD the ESD diodes will conduct which violates the Absolute Maximum Ratings.  If this should happen, then the ADS1115 can be damaged if the input current through the input pin(s) exceed 10mA.  Your ADC inputs should have input filters that can be used to both limit input current and act as an antialiasing filter.

    At this point it is difficult to say if the ADS1115 is damaged.  If the ESD structure within the device is damaged, then most likely the device will get hot even if no input connections are made.  If the device doesn't get hot while running without the inputs connected, then I would try to connect one input source at a time (while running your same software routine) to see if there is one input source that seems to be affecting the device more than another.

    One other consideration is if the device does get too hot to touch and your spray it with freeze spray, it is possible to damage the device as the package could crack due to the thermal shock.

    Best regards,

    Bob B

  • I would absolutely second Bob's opinion.  I've seen plenty of ADC's damaged by ground offsets that weren't considered.  Based on the failure mode, there's a 99% chance that the input ESD protection diodes have been damaged by inputs out of the ADC VDD/GND Range.

  • Hi Bob,

    Thank you for your response.  I have attached a schematic showing the implementation of the ADS1115.

    For the ADS1115 that is running hot, it runs hot regardless of the analog input connections.  Even with the analog inputs disconnected, it starts to heat up as soon as 5VDC power is applied.

    Right now, I don't see any way that I'm violating the maximum ratings on the analog inputs.  The signals all measure as expected when I measure them using a digital multimeter.  The grounding scheme is very simple.  The device-under-test (DUT) is a passive device (potentiometer) and shares the same ground as the ADS1115.

  • Hi Michael,

    It sounds like the device getting hot was damaged.  If one of the ESD structures is damaged as a short, then current will flow from DVDD to GND through the short.  This can happen if one of the analog inputs is powered accidently before the 5V VDD supply is powered.

    From your drawing it is not clear which device package for the ADS1115 you are using.  Also it is not clear from the drawing if there are any resistance at the input to the ADC inputs.  A 1k Ohm resistor in series with the input will limit any possible input current to less than the absolute maximum (assuming the grounds are common and no input voltage exceeds the intended voltage).

    I would advise that the ADS1115 is always powered prior to applying any external input voltage.  I would also advise that if power is at the inputs you do not disconnect the 5V VDD power to the ADS1115 to reset the device while the external input is still powered.  Here you must also consider that any capacitance at the inputs will hold charge, so the inputs should be completely discharged before turning off the supply to VDD.

    As I stated in my earlier post, the ADS1115 has been used in a variety of different systems with no issues.

    Best regards,

    Bob B

  • Hi Bob,

    I agree that the failing ICs are most likely damaged, I'm trying to figure out what could have caused the damage.  If the most likely culprit is excess current through the analog inputs, then I can protect them with series resistors (as you suggested) and TVS diodes.  It's still mysterious as to how the maximum rating would have been exceeded, when no voltages above the maximum ratings are present.

    What about my initial question regarding the maximum sample rate?  We are running this device almost constantly at its maximum sample rate of 860 samples per second.  Would you expect that kind of stress to cause a failure?

    We are using the 10-pin DGS package, see the attached image from the datasheet.

  • Hi Michael,

    The ADS1115 is an oversampling ADC where the actual modulator frequency is the sampling rate (250kHz) regardless of the data output rate of 8sps or 860sps.  The ADS1115 can be run continuously without issue or overheating.

    As to the cause, it is difficult for me to troubleshoot a system that is not in front of me.  All I can say is what I have seen others do that can cause an issue.  I will review them again:

    • An overvoltage condition at the input occurred that caused EOS.
    • Grounds from all sources were not common to the ADS1115 ground.
    • Current not limited to the device using series input resistance.  An accidental overvoltage occurred that exceeds the absolute maximum input current of 10mA as a result.
    • Power applied to the inputs accidently while the ADS1115 was not powered.
    • An input signal source (such as a signal generator) that goes below ground or above VDD.

    Best regards,

    Bob B

  • Hi Bob,

    Okay, that's good to hear that I'm not abusing the ADS1115 just by keeping it running all the time at its maximum sample rate.  I figured that it was not an issue, but I wanted to make sure.

    We'll use your suggestions to make our design more robust, especially when it comes to protecting the analog inputs, and see what results we get.  Thanks very much for the help.

    -Michael