• State TI Thinks Resolved
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 66 subscribers
  • Views 113 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

Original question:

BQ4050: SN8765 chip temperature display data is abnormal

BQ3050: Read the IFIB value of chip SN8765 is all of 0xff ff

zehu chi
Prodigy 240 points
Part Number: BQ3050

Tool/software:

Dear Ti engineer

Dear Ti engineer

In our products using the SN8765 chip, the result of temperature sampling is the maximum，the phenomenon occurs in 3% of cases

Specific details are as follows:

1.In the abnormal NTC value products ,We tested the voltage between the NTC and found it to be normal

2.Replace the IC with the abnormal NTC value to the good product , and the good product also has the NTC abnormal situation

3.Replace the IC with the abnormal NTC value to the normal NTC function IC , and the abnormal NTC function product will be normal

4.We read the IFIB(0x12~0x1F) register for the abnormal IC and find that the values in the IFIB (0x12~0x1F) register are 0xff ff

Do you know why the IFIB value of chip SN8765 is all of 0xff ff ？

Looking forward to your reply!

  • 0
    TI__Expert 5970 points

    Hello Zehu,

    This question has been assigned and we will get back to you.

    Regards,
    Alan 

  • 0
    TI__Mastermind 25991 points

    Hello,

    We are looking into this question. 

  • 0 in reply to Jose Couso
    TI__Mastermind 25991 points

    Hello,

    If those values are all 0xFF then the IC data was lost, I don't know why, that area should have some protection against accidental changes but it fundamentally is just flash.

    I assume other addresses in the IFIB are likewise erased, which means the part is not properly trimmed. It also means any lot trace information is lost.

    It's rather concerning that it could happen in "3% of cases", I would suggest to track down how such high data loss in a flash sector that shouldn't be getting any modifications once it leaves our factory.
    This would require setting specific extra flags/tokens to protect from erasing it by mistake.

  • 0 in reply to Jose Couso
    Prodigy 240 points

    Thanks for you answer!

    We use library function "ifibReadWord" to read the IFIB value and When read ifib register in our SN8765 chip just has the following two phenomenon

    Phenomenon1: CC Offset with CAL=1 is 0xff ff and other ifib register is other normal values

    Phenomenon2: All the ifib register value is 0xff ff

    Then i have three questions

    Question1: About the Phenomenon1 is not normal either,It's just that we ran a current calibration when we programming it.Is the CC Offset with CAL=1 is 0xff ff noraml phenomenon?

    Question2: About the Phenomenon2, what actions does we have that might lead to this?

    Question3:We developed the programming board according to the programming manual. We would like to ask what is the absolute address of the IFIB?

    Best wishes

    Zehu

  • 0 in reply to Jose Couso
    Prodigy 240 points

    Hi Couso

    Do you have time to reply the three questions ? Thank you very much！

    Best wishes

    Zehu

  • 0 in reply to zehu chi
    TI__Mastermind 25991 points

    Hello,

    1. Q1: CC Offset = 0xffff with CAL=1  could be a sign of an overflow or error. Verify against the manual’s calibration output specs.

    2. Q2: Phenomenon 2 (all 0xffff) suggests a communication, power, or initialization issue. Check bus integrity, power stability, and initialization code.

    3. Q3: Why are you accessing IFIB? this is the integrity section of the device. It should never be read. 

      Do you re-program the firmware?, if you do, could you send me the general programming sequence or if you use one of our tools exactly what tool and version? A beagle trace of all the commands you send in the production flow would be pretty great to see if there is anything suspicious where an error could end up corrupting the IFIB contents.

      The most likely place it happens is during programming in ROM mode.