• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 63 subscribers
  • Views 1486 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.

LM4132A - Solder Heat Drift?

Bond - James Bond
Expert 4910 points
Other Parts Discussed in Thread: REF5025, LM4132, REF2033

Hello Reference Experts,

REF5025 has mentioned in the data sheet information about Parameter Measurement Information and Solder Heat Shift. 

Does something similar apply to the LM4132A.. There seems to be mentioned nothing in the data sheet?

Reason for question is LM4132A after mounted on PCB show  Vref with tolerances larger than expected max tolerance of +/-0.05%..  measured deviation from 3v3 is -0.302%

Parts are mounted close to edge of the PCB so we don't expect mechanical stress to be the issue leading to the unexpected tolerance of -0.302%

  • 0 in reply to Bond - James Bond
    TI__Intellectual 1745 points
    Hello,

    The LM4132-3.3 A grade has initial accuracy of 0.05%, but once the device has been exposed to temperature excursions the temperature coefficient must be taken into account in the guaranteed accuracy calculation. In this case, adding initial accuracy, temperature coefficient and thermal hysteresis, the maximum deviation is +/-0.38%. This puts the units within the datasheet tolerance.

    Let me know if you have more questions.

    Chris
  • 0 in reply to Christopher Dean
    Prodigy 60 points
    Hi

    Could you elaborate how you calculate a maximum deviation of ±0,38% of FS?

    We have a maximum temperature of roughly 240°C during the reflow process, so could this be the root cause for the shift in output voltage? According to the datasheet a maximum temperature of 260°C are stated, but it doesn't explain the effects of the soldering temperature.

    Thanks in advance for feedback.

    Best regards

    Carsten
  • 0 in reply to Carsten Hyldal
    TI__Intellectual 1745 points

    Hi Carsten,

    Certainly.  Soldering the part to the board is a significant thermal excursion that puts stresses on the part.  Therefore, the temperature coefficient must be included in the allowable error bounds of the part.  The total error is the sum of the initial accuracy, maximum temperature drift and thermal hysteresis. 

    Initial accuracy: 0.05% -> 500 ppm

    Temperature Coefficient: 20 ppm/C over -40C to 125C. -> 3300 ppm

    Thermal hysteresis: -> 75 ppm

    Total: 3875 ppm -> 0.3875%

    The initial accuracy is guaranteed at 25C prior to any temperature excursions.  Once a unit sees a temperature excursion, such as soldering or other event, the over temperature spec is appropriate.  It is possible that many units will remain close to the initial accuracy after soldering, or have tighter over temperature performance than the spec, but the numbers above are what is guaranteed from manufacturing lot to lot and is based on extensive characterization over temperature.

    Regards,

    Chris

  • 0 in reply to Christopher Dean
    Prodigy 60 points
    Hi Chris

    Thank you for the feedback. I wasn't aware of this phenomena, so I'm a bit surprised, because the part was selected based on initial accuracy of ±0,05%.

    We have 6 x REF5025 in SO8 packages on the same PCB and we haven't seen any issues related to these items. They are actually spot-on every time. But it is also a more expensive part, so perhaps this is one of the benefits of the REF5025 part. Additionally, I don't know, if the LM4132 and REF5025 parts are based on the same principles and techniques.

    Based on your feedback, we will most likely shift to another part in order to achieve better performance. I have looked at the REF2033 part and it seems to me, that it should be possible to mount this part using the same footprint. Based on your formulas I calculate an initial accuracy of ±0,188%.

    By the way does this mean, that even though a precision reference has been selected for a design, it could be beneficial to save the actual output voltage in e.g. EEPROM in order to compensate for output voltage shift related to soldering process?

    Thanks in advance for feedback.

    Best regards

    Carsten
  • 0 in reply to Carsten Hyldal
    TI__Intellectual 1745 points

    Hi Carsten,

    I understand.  The REFxxxx parts are in a different product line so I am not as familiar with their design, but parts like the REF50xx do have lower temperature coefficient (tempco) than the LM4132.  

    Yes, you could save the voltage value in EEPROM if that is available in your system/application.  This can be used for system calibration or offset correction to help reduce the effects of the soldering process or the normal shift from ambient temperature changes in operation.  

    Regards,

    Chris