LM139: Increased bias current failure and recovery

Part Number: LM139

1) We have had recreated a fault where if the common voltage range is taken to outside the Vcc - 2 specification we will get a large input bias current, this is understood to be the failure mode and TI have conformed that.

However, the datasheet states that no damage will occur to the device. We think this is incorrect as we have found once the large bias current occurs there is no way recovering the device (bringing common mode into spec) from this large bias current. Can this be confirmed as error in the datasheet?


2) We are also seeing devices failing for unresponsive outputs. The device has 4 channels. If 2 of these channels were operated at Common mode greater than Vcc-2 would this have an effect to cause failure on channel 3 and 4, which are within the common voltage spec?


3) Power sequencing and input. Is there an increased chance of failure if the signal to the input terminals is present before the voltage rails have stabilised?

  • Hello James,

    TI offers many different versions of the LM139, so I would first like to confirm that we are talking about the same product.   My assumption is that you are talking about either the mil grade LM139 or space grade LM139 covered by the LM139AQML datasheet.   My answer will be based on that assumption, but please give us your exact part number to confirm the answer is for the correct product.

    The datasheet is correct in that taking the inputs negative is not destructive as long as the input currents do not exceed 50 mA as shown in the Abs Max table.  You mention large bias currents.  Exceding the 50 mA limit could damage the part.

    Power sequencing is not an issue as long as you stay within the Abs Max of the datasheet.  The voltages applied to the inputs and outputs are inpendent of the supply voltage.

    Again, these answers are for those parts covered by the LM139AQML datasheet.  If it is another version of the LM139, the answer is completey different.

  • In reply to Kirby Kruckmeyer96848:

    Thanks for responding Kirby

    We are currently using the LM139DG4.

    We received an answer from the forum previous which read:

    "Bringing the input above 3.5V pinches off the internal current source and current is redirected through the base (see the internal schematic in section 8.2) . The result is the input will tend to "float" to equilibrium around the internal bias point.
    The datasheet mentions that damage will not occur and that the output will be correct, but it does not mention that there will be a huge change in the bias current as you "drag" the input stage above the Vs-1.2V point.
    You need to attenuate the input slightly to avoid going above 3.5V."

    We are seeing the LM139 sinking 21uA (which is huge compared to the datasheet Input bias) which is causing the applied input voltage to drop from 5V to 3.7V

    However, we cannot recover the device by changing the Common Mode voltage to be Vcc-2, we are seeing the device has failed hard by continually sinking current in the uA where it should be nA.

    We are trying to understand why this should happen? Why does changing the Common mode within spec not recover the device?

  • In reply to James Allan:

    Hello James,

    It can get a bit confusing with the different versions of the LM139. Based on your part number, it looks like you are looking at a different LM139 than Kirby mentioned and it is covered by a completely different datasheet.

    Can you confirm if this datasheet is the one you are looking at? This will be the correct datasheet for your part.


    Bree Myers

  • In reply to James Allan:

    Hello James,
    I believe that you were looking at the wrong datasheet for the LM139DG4. The correct datasheet for this part is:
    For this part, the abs max negative limit is -0.3V and the inputs should not be taken below this level. The statement in the LM139AQML and LM139-N datasheets that no damage will occur from a negative pulse on the inputs is not applicable to the LM139DG4.