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.

TFP410: Display signal lost within minutes of power up

Part Number: TFP410

I encountered some problems when using the TFP410. When using some chips, there will be a signal loss on some displays within 3 minutes after power-on. The screen will be black, then the signal is normal, and the signal will be normal in the future.

I measured some signals and found that the output data(TX2+) and line synchronization (HSYNC)would shift(About 40ns) at the moment of signal loss.

The following are oscilloscope pictures. The yellow one is HSYNC and the purple one is TX2 +.

I suspect that this displacement caused the display to lose the signal. Which conditions will cause this signal displacement, how to avoid this signal displacement?
Could there be other reasons for this phenomenon?

Thank you.

  • Hi Hao,

    Is HSYNC shifting on the input of the TFP410 from your video processor? If so then that is probably what is causing the issue. The TFP410 is just a bridge, so if there's shifting on the input then there will be shifting on the output.

    Regards,

    I.K. 

  • Hi I.K.

    I am sure HSYNC is not shift. 

    I did this experiment on my motherboard. After swapping two HDMI TFP410s, one of them is not lost, one is lost, and the other signals are unchanged. Finally, after the two TFP410s are swapped, the original normal HDMI signal starts to lose signals, and the other one Lost signal HDMI becomes normal. After only swapping the TFP410 chip, the phenomenon still follows the TFP410.

    So I wonder if there is any inconsistency in this chip, or if there are any settings that are not set properly, which causes the chip to work in an unstable state.

    Thank you

    Regards,

    Yin hao

  • Hi Hao,

    How are you measuring HSYNC on the output of the TFP410? Do you have a way of decoding the TMDS lanes? 

    Also, just to clarify, did you see HSYNC shifting 40ns on the input of the TFP410 when the signal was lost, or just the output? 

    Regards,

    I.K.  

  • Hi I.K

    I didn't look at TMDS lanes. I saw that there was a shift in the set of data signals on the oscilloscope, just like the oscilloscope picture I sent before. This moment corresponds to the moment when the image is black.

    The HSYNC signal I give to the TFP410 has no transitions.
    In general. I saw above the data output by TFP410 that a signal has shifted. This moment of signal shift corresponds to the moment when the display is blank. I am not sure whether the output synchronization code has moved. I just think the signal movement is a skeptic point.

    Regards,

    Yin hao

  • Hi Hao,

    Okay I think I understand. You don't see HSYNC on the input shifting, but when the issue happens you see output data (TX2) shifting in relation to the input HSYNC, correct? 

    Have you checked to see if there's any difference in the VSYNC and DE signals when the display blanks?

    Regards,

    I.K. 

  • Hi I.K.

    Yes,you got it.

    I checked it,when display blanks,Hsync Vsync and DE are not shifting,the output data shifting in relation on the input Hsync(and Vsync,and DE).

    Some chips will cause black screen. I replaced the chip and found that the black screen of the display is consistent with the TFP410.

    And I found that it has something to do with temperature. Each time these chips reach a certain temperature, the display will be black (and then normal). There is a chip whose display is blank at 35 degrees. I have tried to change the speed of temperature rise. At about 35 degrees, the display is blank.

    Regards,

    Yin hao

  • Hi Hao,

    This is starting to sound like a quality issue with those chips. Can you confirm the following for me?

    1. What is the operating frequency? Are you able to change the frequency to see if this also affects the issue?

    2. What is the length of the cables that you are using? Do you see the issue more/less frequently if you use shorter cable lengths? 

    3. Can you share your schematic and layout with me at i-anyiam@ti.com?

    4. When you swap the units, does the good unit always keep the signal even after varying the temperature? 

    Regards,

    I.K. 

  • Hi

    I.K. thank you for your reply.

    1. 148.5MHz , 1080P60. I dont change the frequency.

    2. I tried a short, good quality HDMI cable and this issue still happens.

    3.I wiil send some picture to you.

    4.Yes,the good unit always keep the signal even after varying the temperature.

    Regards,

    Yin hao

  • Hi Hao,

    Thanks, I received your files. I have some additional follow-up questions:

    1. It looks like you're using the I2C interface. Can you provide a register dump of the device for a good unit and a unit when the display blanks? 

    2. Can you confirm that the HDMI is video only (no audio)?

    3. Are you providing a proper reset to the device on power-up (ISEL/RST# low during power-up and then asserted high after power is stable)? 

    4. When the display blanks, do you notice any differences in the input clock? What about the output clock? 

    Regards,

    I.K. 

  • Hi I.K.

    Thank you for reply.

    1.I just send 0Xbf to 0X08.

    2.HDMI is video only.

    3.I provide a reset 3second.

    4.The input clock and the output clk have not changed.

    Regards,

    Yin hao

  • Hi

    I.K.

    Today I did a test. When the reset is grounded and the SCL is connected high, the display does not appear black when IIC mode is not used.

    Is there something wrong with our configuration of IIC?

    Regards,

    Yinhao

  • Hi Yinhao,

    It is probably because you are not providing a proper reset to the device with ISEL/RST# in relation to I2C (see section 7.4.4 of datasheet).

    However, I don't see the point of you using I2C for this device. The main reason people use the I2C is to make use of the DE generator when their source does not provide a DE signal. Since your source has a DE signal, you don't need the DE generator, and you don't need I2C since everything else is pin configurable. I recommend you keep ISEL grounded to disable I2C.

    Regards,

    I.K.