• State TI Thinks Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 2 answers
  • Subscribers 63 subscribers
  • Views 379 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

BQ27Z561: Data cannot be written after the communication method is switched to HDQ

wen xiao
Prodigy 140 points
Part Number: BQ27Z561
Other Parts Discussed in Thread: EV2400, BQSTUDIO

Hello,  expert,

When I use the BQ27Z561, I found that sometimes the data cannot be written after converting to HDQ:
For example, if I send an open lifetime command to the battery, after reset, the lifetime will be closed again;
Or when I write the Manufacturer Info to the battery in the data memory, the Manufacturer Info returns to the default value after re-reading.

Can you help me analyze what is causing this and is there a way to avoid such a problem.

  • 0
    TI__Guru 61126 points

    What do you mean with "open lifetime command"? If you refer to unseal (to get access) as "open lifetime", this will auto-reseal (close) after a reset, regardless of HDQ or I2C interface.

    Please make sure that the gauge is unsealed before writing anything to data memory.

  • 0 in reply to Dominik Hartl11
    Prodigy 140 points

    Hello,Dominnik,

    As shown in the figure below, the IC is already unseal. After sending the LT_EN command, LIfeTime has not been opened.

    When I try to modify the parameters in the Data Memory, the lower left corner of BQ studio will display A read of data written failed comparison.

  • 0 in reply to wen xiao
    Prodigy 140 points

    I found that this state is very similar to the unsealed state, but the BQ studio interface shows that the IC is full access. I tried to encrypt and then decrypt the IC, and found that the IC can be encrypted or unsealed, but the full access failed, and we did not change the password.

    It's been three days since the question was asked, is there anyone here who can help me analyze it?

  • 0 in reply to wen xiao
    Prodigy 140 points

    Additional modification:
    Most of the batteries cannot be decrypted after we encrypt, of course, we have not changed the default password

  • 0 in reply to wen xiao
    TI__Guru 61126 points

    There are several potential issues:

    1. When you get the error "A read of data written failed comparison", then the voltage may have (temporarily) dropped below the minimum required for flash updates. If you use a lab power supply, make sure that the current limiter isn't dropping the voltage temporarily during flash writes. If you use a battery, make sure it's well charged.

    2. If you can't "unseal full access" then you may not wait long enough for the two-word authentication mechanism to be ready. The state machine inside the gauge that controls the unseal process must be in idle state when you start issuing the two unseal full access words. One way to guarantee this is to not communicate with the gauge for 4 seconds (this let's it time out).

    When you say that the batteries can't be decrypted, do you mean unsealed? If so, please make sure that you wait 4 seconds without any communication with the gauge and then you issue both unseal words in two separate write commands. If you do this, the process will work 100%.

  • 0 in reply to Dominik Hartl11
    Prodigy 140 points

    Hello，

    The battery voltage is 3.7V, and I checked the peripheral circuit, it works, so I confirmed that there is no problem with the power supply;
    Regarding decryption, I'm sure we waited more than 4s.

    And, the strangest problem is that BQ studio shows that the IC has Full accsee, but in fact, it cannot write data into the IC

  • 0 in reply to wen xiao
    TI__Guru 61126 points

    Are you using the EVM or is the EV2x00 patched into your own board?

  • 0 in reply to Dominik Hartl11
    Prodigy 140 points

    Use Ev2400

    This problem is found in our production, the defect rate is about 0.5%，and the sample did not find this problem。

    In fact, I would like to know what factors can cause such a problem, the IC itself or outside influences?
    If there are some external factors that can cause such a problem, maybe I can adjust it in production

  • 0 in reply to wen xiao
    TI__Guru 61126 points

    The most likely problem is that the supply voltage to the gauge drops during a flash write. That's when bqStudio won't be able to read back the value that was written because the gauge will not write the value that bqStudio sends to it and therefore the read verification step fails.

    About unsealing: Please make sure that you disable auto-refresh in bqStudio.

    And: "And, the strangest problem is that BQ studio shows that the IC has Full accsee, but in fact, it cannot write data into the IC". If the gauge SEC bits show that it's unsealed or unsealed full access then that is the true state. If you can't write in the unsealed state then the cause isn't that the gauge is sealed (because it isn't), the cause is that the flash write failed and that only happens if either the gauge is defective (=bad part) or, in the vast majority of cases, it's because the supply voltage drops (temporarily). If you can rule out that the supply, even small glitches (please check with a scope), dropped below 2800mV during the write operation, then chances are that the part is defective. It's not normal for the part to reject flash writes when the voltage is correct and when you use bqStudio.

  • 0 in reply to Dominik Hartl11
    Prodigy 140 points

    Thanks for your answer, I will try to measure the voltage when i send the write command.


    In other words, if the oscilloscope reads the voltage is OK, does it mean the IC is damaged?

  • 0 in reply to wen xiao
    TI__Guru 61126 points

    If the voltage is always above approx. 2800mV (without any glitches) and you can guarantee that no other devices interfere with the communications bus and you can't write data to the gauge, then this points to a defective gauge.

  • 0 in reply to Dominik Hartl11
    Prodigy 140 points

    Hello,

    After testing, we found that this problem can be solved by powering off the IC and then powering it on again.

    Can you help us analyze what's causing it

  • 0 in reply to wen xiao
    TI__Guru 61126 points

    The gauge doesn't store anything in volatile memory that would prevent it from writing to flash. When you power cycle the gauge, do you power cycle the gauge only or your whole system?

    The main cause of flash write problems is the power supply not being stable when the gauge actually starts the write to flash process.

    Are you able to reproduce this on the EVM?