• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 1839 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LM397: Low level output

Pol REYMOND
Intellectual 1730 points
Part Number: LM397
Other Parts Discussed in Thread: TL331, LM339

Hello team,

Customer had several issues on LM397 device (DC1649) during In Situ Test. Indeed, they noticed a shift on the Output_Low which exceeds the 400mV limit (with a pull up 5V, 500Ohms). The transistor saturates. With 4.8V, the issue does not occur.

It is interesting to notice that, in the past, the average output voltage was about 300mV and has recently increased to 480mV.

So, do you think that something during manufacturing step could justify this shift? (die/fab changing for example).

Thank you for your help.

Best regards

  • 0
    TI__Guru 53846 points

    Hello Pol,

    The 400mV table limit is with a 4mA sinking current.

    5V across 500 ohms is 10mA. At about 10mA, you are on the knee of the short circuit current limit (~13mA). This can be seen in Figures 3 and 7.

    I will check to see if there have been any changes or issues to the device.

    In the mean time they can try the TL331 - the second source of the LM397. It was designed by a different company (at the time) on a different process - so it is independent of the LM397.

  • 0 in reply to Paul Grohe
    TI__Guru 53846 points
    OOPS...Typo... TLV331 should be TL331
  • 0
    TI__Guru 53846 points
    Hello Pol,

    There have not been any fab or test changes to the LM397.

    Can you get the top and bottom markings off the package? Data codes are on the bottom of small packages. Also take note of any dashes or lines on the edges of the package as these are also codes. Get photos if possible.

    It is also helpful if you get a photo of any packaging labels - from the reel and shipping box.
  • 0 in reply to Paul Grohe
    Prodigy 30 points

    Hello Paul ,

    I'm the Component engineer in charge with this issue (Customer). To answer your question, there is no marking on parts except "C397". However, you will find in attached, the reel label picture. There are several informations about the parts.

  • 0 in reply to LAETITIA BARRAU
    TI__Guru 53846 points

    Hello Laetitia,

    Perfect! I have passed the photos on to the product engineers.

    "C397" is the correct code for a LM397MF. There still should be a three digit date code on the *bottom* of the device.

    What is the failure rate they are seeing? All devices or just a few? From one reel or several? Does the VOL change if the device is changed?

  • 0 in reply to LAETITIA BARRAU
    TI__Guru 53846 points

    Hello Pol,

    I have heard back from the Product Engineers.

    The fab lot the devices came from had slightly lower NPN beta than normal. Lower betas reduce the output sinking ability. However, these devices are still within fab limits and datasheet specification and would not be considered failures. The short-circuit current is not a trimmed or controlled parameter (but it is limit tested), so there can be wide variations across lots.

    The customer is running a fairly high sinking current (~10mA), near the average current limit, and is outside the specified output values in the table. Their 10mA current is right at the average short circuit current for that particular lot - which is why they are seeing changes in VOL.

    I would recommend they change their pull-up resistor to 1k or more to reduce the current by half and move safely away from the current limit area. The typically used pull-up resistor value for the LM397/LM339 family is 1k to 100k (4.7k typ), and 500 ohms is a bit heavy. There will be a small penalty in rise-time, but not much depending on Cload.

  • 0 in reply to Paul Grohe
    Prodigy 30 points

    Thank you very much for your support. We will correct the problem as you suggest.