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.

LM2901: conflicting information about the maximum temperature allowed during operation of this device

Part Number: LM2901
Other Parts Discussed in Thread: LM339A, LM339, LM3302, , LM139-N, LM239-N, LM2903, LM2903-Q1

Hello,

It seems like there is some conflicting information about the maximum temperature allowed during operation of this device. The description in section 3 and the electrical characteristics in table 7.8 indicate a rating up to 125 C ambient temperature, while section 7.3 specifies recommended operation with a maximum junction temperature of 125 C. If both of these are true, then it would imply that no power dissipation is allowed at its maximum ambient operating temperature.

We have an condition where we believe that the device can get to approximately 134 C junction temperature (104 C case temperature). Does this actually violate the datasheet? If so, then can you provide any more information about operation in this mode between the recommended operating condition (125 C junction temperature) and the absolute maximum rating (150 C junction temperature)?

  • Hello David,

    For low-power devices like comparators and op-amps, which do not usually dissipate much heat, ambient temp and junction temp tend to be used interchangeably.

    The LM2901-N datasheet, which was the original National Semiconductor datasheet that everyone copied, has a specific note that the temperature must be de-rated to the junction temp:

    "For operating at high temperatures, the LM339/LM339A, LM2901, LM3302 must be derated based on a 125°C maximum junction
    temperature and a thermal resistance of 95°C/W which applies for the device soldered in a printed circuit board, operating in a still air
    ambient. The LM239-N and LM139-N must be derated based on a 150°C maximum junction temperature. The low bias dissipation and
    the “ON-OFF” characteristic of the outputs keeps the chip dissipation very small (PD≤100 mW), provided the output transistors are
    allowed to saturate."

    The National datasheet specifically states the recommended temperature operating range is junction temperature. So the actual ambient max temp would be a few degrees lower.

    The TI version of the LM2903 datasheet seems to imply Ambient Temperature range.

    So this seems to imply that there is allowance for self-heating "built-in" to the specs at an ambient temperature of 125C with no load (power supply spec). Worst case the die will be a few degrees higher than ambient due to quiescent current. I will check with the Quality folks to get their interpretation.

    In your case, I would operate as if it is a 125C maximum junction temperature to be safe, as detailed in the National datasheet. If so, then yes, you would be violating the datasheet specifications (Tj > 125), but not causing damage (<150C).

    We know this process and design is functional up to 150C, and have even released a Q1, grade-zero (150C) version of the dual (LM2903-Q1), so I am not worried about a failure at 135C.

    Is this a failure mode? Or normal operating conditions?

  • David,

    Do you have any more questions regarding this topic? If not, this thread will be closed shortly since it has been 7 days since the last response.

    Regards,
    Jonny
  • This information helps.

    The condition that we’re concerned about is a combination of a design flaw causing abnormal power dissipation (unintentionally using this device with +/- 15V supplies such that the output tries to drive to -15V in its low state, but a clamping diode on the output clamps to -1V and the comparator sinks larger than normal current), and also an environmental requirement to operate at hot ambient air such that local case temp is appx 104 C and calculated junction is 134 C. This environmental condition would be rare, though, and exceeding 125 C junction would be abnormal and is probably only a theoretical, not real, condition.
  • Hi David,

    Exceeding the 125C junction would mean that it no longer meets datasheet specs (may not meet ib, Vos, Iout, etc). But if you are in a fault condition, will that matter much since the output is externally "stuck" anyways? There should not be any 'damage' if the junction temperature is kept below the 150C max during a fault condition. Perhaps switch to the SOIC package to get more heat out if necessary?

    BTW: We know this family can function at 150C since we have released a grade-zero (150C) version of the dual (LM2903-Q1 "E")....If that makes you feel any better...