• State TI Thinks Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 2 answers
  • Subscribers 34 subscribers
  • Views 606 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.

DS90UB935-Q1: Serializer UB935 TI2AB A6SK G4

Houssem Mouelhi
Prodigy 10 points
Part Number: DS90UB935-Q1
Other Parts Discussed in Thread: DS90UB936-Q1

The Camera does not start to stream an image only after 15 seconds (Or more) at cold temperature (-10C or below) , This defect on the camera was detected initially after RPA (Reinforced Product Audit) during the parametric test when the camera was checked at -40C degrees.
The same phenomenon was confirmed on the Bench Test after putting the camera at -40C for 30 minutes or more. After HW analysis, we confirmed that at -10C degrees or below, the lock between the Serializer and Deserializer did not happen only after 12 seconds thorugh the EVM board.
We performed a full cross test and we confirmed that the defect is following the Serializer component:
> Known Good Camera assembled with NOK Serializer becomes behaving exactly as the failed part (Starts to stream image after 15 seconds at -10C degrees or below).
> Camera failed originally and assembled with OK Serializer becomes a good part (Starts to stream an image instantly at -10C degrees or below).
Could you please support us to identify the root cause of this difference between the OK and the NOK parts .

  • 0
    TI__Expert 8030 points

    Hi Houssem,

    The behavior that you have observed while using this device is unexpected. To help us understand why you are encountering this issue with the device, I have the following questions:

    1. What is the part number of the serializer you are using? For the deserializer, are you using the DS90UB935-Q1?

    2. For your serializer and deserializer configuration, what clock mode are you operating in? Synchronous, non-synchronous external, non-synchronous internal, or DVP mode?

    3. For your test setup, are you using EVMs for both the serializer and the deserializer?

    4. How many serializers have you tested? Of the total amount of serializers that you have tested, how many have failed?

    5. Are you following the power sequencing requirements specified by the serializer datasheet?

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    Prodigy 10 points

    Hello 

    You find below in blue my response 

    1. What is the part number of the serializer you are using? For the deserializer, are you using the DS90UB935-Q1?

    DS90UB935TRHBRQ1

    2. For your serializer and deserializer configuration, what clock mode are you operating in? Synchronous, non-synchronous external, non-synchronous internal, or DVP mode?

    Synchronous

    3. For your test setup, are you using EVMs for both the serializer and the deserializer?

    Yes

    4. How many serializers have you tested? Of the total amount of serializers that you have tested, how many have failed?

    We tested 3 parts and we had 1 NOK 

    5. Are you following the power sequencing requirements specified by the serializer datasheet?

    Yes

  • 0 in reply to Houssem Mouelhi
    TI__Guru** 107650 points

    Hello Houssem,

    just to clarify, are you testing with your own camera as stated in the original post, or testing with our EVM as stated in your reply?

    If testing with our EVM, are you de-soldering and resoldering UB935 on the EVM and testing?

    During the first 15 seconds, where you have no LCOK, can you access the SER via local I2C? If yes, can you provide register dump from both, SER and DES, during the failure and after the failure?

    Which DES are you using?

    Can you provide the schematic from the camera?

  • 0 in reply to Hamzeh Jaradat
    Prodigy 10 points

       

    Hello 

    You can find below the videos of the part NOK vs part OK 

  • 0 in reply to Hamzeh Jaradat
    Prodigy 10 points

       

    Hello 

    You can find below the videos of the part NOK vs part OK 

  • 0 in reply to Houssem Mouelhi
    TI__Expert 8030 points

    Hi Houssem,

    I'm currently looking into this and will provide a response within the next 24 hours.

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    TI__Expert 8030 points

    Hi Houssem,

    Thanks for providing both videos. I see that it takes much longer for LOCK to be established for the NOK device. 

    To help solve this issue, we need to gather more information about the configuration of both devices. Therefore, we have the following questions:

    -Are you testing with your own camera as stated in the original post, or are you testing with our EVM as stated in your reply?

    -If testing with our EVM, are you de-soldering and resoldering UB935 on the EVM and then testing?

    -During the first 15 seconds, where you have no LOCK, can you access the SER via local I2C? If yes, can you provide register dump from both, SER and DES, during the failure and after the failure?

    -Which DES are you using?

    -Can you provide the schematic from the camera?

    When testing with your own board, have you tried disconnecting your PoC circuit, and powering each side locally? If not, can you perform this test, so that we can see if your PoC network may be causing this unexpected behavior?

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    Prodigy 10 points

    Hello Arnold 

    You will find in blue my answers :

    -Are you testing with your own camera as stated in the original post, or are you testing with our EVM as stated in your reply?

    We tested with our own camera plug into the EVM 

    -If testing with our EVM, are you de-soldering and resoldering UB935 on the EVM and then testing?

    No we didnot remove the serializer and place on the EVM

    -During the first 15 seconds, where you have no LOCK, can you access the SER via local I2C? If yes, can you provide register dump from both, SER and DES, during the failure and after the failure?

    No access to I2C when no lock

    -Which DES are you using?

    The serializer is ds90ub936-q1

    -Can you provide the schematic from the camera?

    See attached file 

    When testing with your own board, have you tried disconnecting your PoC circuit, and powering each side locally? If not, can you perform this test, so that we can see if the PoC network may be causing this unexpected behavior?

    We did not disconnect the PoC circuit

    b137141C2_Var01.pdf

    Could you please also confirm the reception of the part NOK ?

    Track & Trace Express _ DHL _ Ireland, Republic of.pdf

  • 0 in reply to Houssem Mouelhi
    TI__Expert 8030 points

    Hi Houssem,

    Thanks for sending the schematic file. Please allow me until next Monday to review the schematic file and provide feedback.

    Houssem Mouelhi said:
    No access to I2C when no lock

    Without lock, you should still be able to perform a local I2C read of the device registers. This can be done by connecting your I2C controller locally to the I2C bus pins of the device. Remote I2C read commands will not work whenever there is no link established, but I2C read and write commands sent over the local I2C bus will still work since the signal does not need to travel over the link. 

    Can you use the local I2C bus to provide a register dump of both an OK serializer and a NOK serializer? 

    Houssem Mouelhi said:
    Could you please also confirm the reception of the part NOK ?

    Could you please provide more information about where the NOK part was sent? Was the NOK device sent to a TI office? What is the name of the person that you sent the NOK device to?

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    Prodigy 10 points

    Hello Kenneth 

    Do you have any feedback about the returned part ?

  • 0 in reply to Houssem Mouelhi
    TI__Expert 8030 points

    Hi Houssem,

    I reviewed the DS90UB935-Q1 schematic that you provided. Here is my feedback:

    -The resistor connected to the IDX pin does not match any of the recommend resistor configurations provided by the DS90UB935-Q1 datasheet. If you are trying to configure the I2C 8-bit address of this pin as 0x30 (1.8 V I/O Voltage), I recommend that you switch the 39 kΩ resistor with a 40.2 kΩ resistor, so that the configuration of this pin matches the recommendations on the datasheet.

    -The resistor connected to the MODE pin does not match any of the recommend resistor configurations provided by the DS90UB935-Q1 datasheet. If you are trying to configure the device for CSI-2 synchronous mode, I recommend that you switch the 4.7 kΩ pull down resistor with a 10 kΩ pull down resistor, so that the configuration of this pin matches the recommendations on the datasheet.

    -The voltage rating of the DOUT- AC coupling capacitor is lower than what is recommend. For this component, I recommend that you use a capacitor with a 50 V rating instead of a capacitor with a 25 V rating. 

    -The power supply ferrite beads being used in the design (L11, L3, and L10) do not meet the recommended specifications given by the DS90UB935-Q1 datasheet. As stated by the datasheet, these ferrite beads should have an impedance of 1 kΩ at 100 MHz, and a DC resistance less than 25 mΩ. For the power supply connections of this device, I recommend that you use ferrite beads that meet these specifications. 

    Houssem Mouelhi said:
    Do you have any feedback about the returned part ?

    Could you please provide more information about where the returned part was sent? Was the returned part sent to a TI office? What is the name of the person that you sent the returned part to?

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    Prodigy 10 points

    Hello Kenneth 

    Thank you for your feedback .

    I was waiting for the test results and the analysis on the returned NG part .

    We are manufacturing thousands of cameras every week and only this one has this issue. And with A-B-A analysis it was proven that this specific serializer is causing this problem. 
    Here are my comments:
    -The resistor connected to the IDX pin does not match any of the recommend resistor configurations provided by the DS90UB935-Q1 datasheet. If you are trying to configure the I2C 8-bit address of this pin as 0x30 (1.8 V I/O Voltage), I recommend that you switch the 39 kΩ resistor with a 40.2 kΩ resistor, so that the configuration of this pin matches the recommendations on the datasheet.
    --> The value of the resistor is not important, the voltage at the pin is. According to the datasheet as long as the voltage at the IDX pin is lower than 0.131*VDD18 then correct address will be set. 39K or 40.2k will not make any difference.


    -The resistor connected to the MODE pin does not match any of the recommend resistor configurations provided by the DS90UB935-Q1 datasheet. If you are trying to configure the device for CSI-2 synchronous mode, I recommend that you switch the 4.7 kΩ pull down resistor with a 10 kΩ pull down resistor, so that the configuration of this pin matches the recommendations on the datasheet.
    --> The value of the resistor is not important, the voltage at the pin is. According to the datasheet as long as the voltage at the Mode pin is lower than 0.131*VDD18 then correct address will be set. 10K or 4.7k will not make any difference.

    -The voltage rating of the DOUT- AC coupling capacitor is lower than what is recommend. For this component, I recommend that you use a capacitor with a 50 V rating instead of a capacitor with a 25 V rating.
    --> DOUT- is not connected to the outside world so it won't be exposed to high voltages. Can you explain to me how the voltage rating causes this weird behaviour? 

    -The power supply ferrite beads being used in the design (L11, L3, and L10) do not meet the recommended specifications given by the DS90UB935-Q1 datasheet. As stated by the datasheet, these ferrite beads should have an impedance of 1 kΩ at 100 MHz, and a DC resistance less than 25 mΩ. For the power supply connections of this device, I recommend that you use ferrite beads that meet these specifications. 
    --> PoC filter is perfectly fine for this application. It has been used in multiple projects and millions of cameras are produced with no issues. 

    Could you please share with us the test results and the analysis of the returned NG part .
  • 0 in reply to Houssem Mouelhi
    TI__Expert 8030 points

    Hi Houssem, 

    Houssem Mouelhi said:
    --> DOUT- is not connected to the outside world so it won't be exposed to high voltages. Can you explain to me how the voltage rating causes this weird behaviour? 

    I do not expect the voltage rating of the capacitor to be causing the behavior that you have observed with this device. I mentioned this is my initial review because the voltage rating of this capacitor did not match our datasheet recommendation. 

    Houssem Mouelhi said:
    --> PoC filter is perfectly fine for this application. It has been used in multiple projects and millions of cameras are produced with no issues. 

    In my initial feedback, I was referring to the ferrite beads connected to the VDDD, VDDDRV, and VDDPLL pins of the device. In the DS90UB935-Q1 datasheet, we recommend for the ferrite beads connected to these pins to have an impedance of 1 kΩ at 100 MHz, and a DC resistance less than 25 mΩ. I mentioned the ferrite beads connected to these pins (L11, L3, and L10) in my initial review of the schematic because they do not meet our recommended specifications. Similar to the voltage rating of the capacitor, I do not expect these ferrite beads to be the cause of the behavior that you have observed with this device. 

    Houssem Mouelhi said:
    Could you please share with us the test results and the analysis of the returned NG part .

    I am currently in the process of reaching out to other members of the team to find more information about the current status of the returned device. I will provide an update on the status of this device within the next 2-3 days. 

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    TI__Expert 8030 points

    Hi Houssem, 

    Thank you for your patience. 

    A member of our quality team has been assigned to this issue and should be contacting you over email for updates regarding the status of the returned part. Any future communications relating to the NG part will now take place over email with the assigned quality team member. Since this E2E thread will no longer be used for communications relating to the NG part, I will go ahead and close this thread. 

    If you do not receive an email from a member of our quality team, please let me know. 

    Regards,

    Kenneth

  • 0 in reply to Kenneth Arnold
    Prodigy 10 points

    Hello Kenneth 

    NO i didn't receive any email from a member of your quality team .

  • 0 in reply to Houssem Mouelhi
    TI__Expert 8030 points

    Hi Houssem,

    I spoke to the quality team member today and was told that you responded to the initial email. If this is not the case, please let me know.

    As previously mentioned, all future communications relating to the NG part will now take place over email with the assigned quality team member. Since this E2E thread will no longer be used for communication relating to the NG part, I will go ahead and close this thread. 

    Regards,

    Kenneth