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.

ADS131M08: maximum absolute current for internal reference

Part Number: ADS131M08

Dear TI,

I've read on this forum that the internal reference of the ADC is not meant to power external circuits without buffering. Could any damage occur if external circuitry started drawing up to 600 microamps from the REF pin, or is this safe and merely outside of the recommended operating conditions?

Kind regards.

  • Hello Damien,

    I just checked this on the bench and the device clearly can't start source much current at all (which is what I assume you mean by "external circuitry started drawing", as in the device is providing the current). I was able to cause the voltage on the reference to droop trying to source 10uA of current. Unless the specification is listed in the datasheet, you can be assured we didn't test, measure, or simulate this kind of specification. 

    To clarify a bit, we switch in physically different circuitry within the device when configuring the device to accept an external reference (i.e. when EXTREF_EN is flipped). So, the device can sink and source more current when the external reference is used but not when the device is configured for internal reference mode, because the current paths are different. This is why the situations from the related and this post are different.

    As a general rule, its almost never recommended to use a power rail produced by any IC unless it has a external current specification or the body of the datasheet says its okay. Sometimes, the body of a datasheet will say to not use a power rail externally and I can't find that text in our datasheet. I will be recommending we add this text in the next revision of the datasheet.

    Best,

    -Cole

  • Hello Cole,

    Thank you very much for checking on the bench, that level of support is very much appreciated! The 10uA figure does clarify how little current the internal reference can actually source.

    To rephrase the original question, I was looking at a potential corner case in our design where (a) the external reference has been physically disconnected, (b) EXTREF_EN is disabled (c) there is still a 2 kOhm load between REFIN and GND, which would normally be powered by the external reference.

    Conversely, another thing the datasheet could mention is whether any precaution (like power sequencing) is required to avoid applying an external reference voltage to the REFIN pin while the chip is still configured to use the internal reference. 

    Best,

    Damien

  • Hello Damien,

    Yes that makes sense, I'll push some of that feedback for future devices with external references.

    I did also pull 600uA for several seconds and no damage was sustained. The voltage returned to the nominal voltage when the load was removed which is what we expect to happen to the voltage on the rail while we're powering up. I can't claim that this was a test that can account for aging or drift, but it gives a lot of confidence because any Power On Reset (POR) timing (250us) should be significantly less energy applied to the pin.

    The general recommendation is to the establish the voltage on the REFIN pin sometime before t_POR. However, when the supplies are ramping, we just need to make sure that the absolute voltage on spec on the REFIN pin is followed (i.e. not greater than whatever the voltage of AVDD is + 0.3V, which would be changing while ramping). So there's really no issues of having a load on the pin when the the digital within the device isn't taking any action and requesting current from the internal reference. 

    If the external reference voltage can't be established between the t = t_90%supply to t = t_90%supply + t_POR, then I'd recommend taking the device into STANDBY as soon as you can while you wait for the reference voltage to be established.

    And finally, the delay between applying the voltage on the REFIN pin to flipping EXTREF_EN shouldn't produce any damage because the voltages should be very similar so energy doesn't really get applied to the pin. 

    Again, this is all stuff that should happen within a couple of milliseconds so I don't have an issues with damage as long as the absolute maximum ratings are followed. 

    Best,

    -Cole