DS90CR286AT-Q1: Request to Check Screen Color Issue

Hi Yoon,

1. Issue frequency: Is this issue only on a single board? Could you tell us how many units were tested and how many show this issue? Does the issue chip show this issue on every power-cycle continuously or is there any variance like it changes color intensity during runtime or flickers?

Is this a mass produced system or only being tested in the design/validation phase?

2. ABA swap: Is this issue following the specific DS90C286AT-Q1 chip? 

• If you place the issue chip on a working board, does it show the issue?
• If you place another chip on the issue board is the issue resolved?

3. On the issue chip, could you check whether other colors are working as expected? You could try red, green, blue, white and black, and check whether there is always a reddish tint.

You could also output solid colors on full screen and check the RGB TTL/CMOS input pins to check whether there is a specific faulty pin/line.

The device simply takes in the RGB + HS/VS/DE signals and serializes them to LVDS output. So if there is only a certain color showing more on the screen, it could be due to a certain faulty pin, or it could also be due to skew margins and the bits shifting. 

To check for the skew issue, although it may not be likely at lower frequency RGB input, could you try lowering the PCLK rate and see if the issue goes away? You could also increase PCLK and check if it still has the issue.

Best regards,
Ikram

​Please find below our responses to the questions in the email.

​

1. Issue frequency: Is this issue only on a single board? Could you tell us how many units were tested and how many show this issue? Does the issue chip show this issue on every power-cycle continuously or is there any variance like it changes color intensity during runtime or flickers?

​

Is this a mass produced system or only being tested in the design/validation phase?

​

Answer : 

  1. A defect occurred on a single board, with 1 defective unit found out of 100 units.

  2. The same defect persists even after power cycling the device.

  3. No flickering was observed.

  4. This is a product currently in mass production.

2. ABA swap: Is this issue following the specific DS90C286AT-Q1 chip? 

If you place the issue chip on a working board, does it show the issue?

If you place another chip on the issue board is the issue resolved?

Answer : Replacing the DS90CR286AT-Q1 component resolved the defect.

3. On the issue chip, could you check whether other colors are working as expected? You could try red, green, blue, white and black, and check whether there is always a reddish tint.

You could also output solid colors on full screen and check the RGB TTL/CMOS input pins to check whether there is a specific faulty pin/line.

The device simply takes in the RGB + HS/VS/DE signals and serializes them to LVDS output. So if there is only a certain color showing more on the screen, it could be due to a certain faulty pin, or it could also be due to skew margins and the bits shifting. 

To check for the skew issue, although it may not be likely at lower frequency RGB input, could you try lowering the PCLK rate and see if the issue goes away? You could also increase PCLK and check if it still has the issue.

Answer : Other colors operate normally.

Best regards,
Duck won

dwyoon@cotstech.com 으로 답변 부탁 합니다.

Hi Duck won,

Yoon Duckwon said:

Answer : Other colors operate normally

1. Test color patterns: Did you test this on board? If you have the SoC (video source) send full screen color bars pattern, solid colors (yellow, red, blue, green), or checkboard black and white pattern, is the screen always more tinted red or brighter red?

2. Lower PCLK rate: If possible, also check with lower pixel clock rate and see if the issue persists. If there is a skew issue, where the RGB patterns are marginally being misread, then a lower pixel clock rate might not show the issue. This would not be a workaround (because the panel must have exact PCLK needed) but it might indicate a skew margin issue.

If the RGB input, or LVDS output has significant trace length mismatch, or more delay on certain lanes, then the receiver may strobe and read the data at the wrong bit.

3. Probe input and output: With a logic analyzer if you probe the input and output pins while transmitting an exact color pattern where all the pixels are the same color, then the waveforms could show how if the there is significant channel to channel skew.

Please also check the layout for whether the traces are length-matched.

Best regards,
Ikram