Because of the Thanksgiving holiday in the U.S., TI E2E™ design support forum responses may be delayed the week of November 23.

SN65DSI83: SN65DSI83 Display Issue

Part Number: SN65DSI83

Hi Team,

My customer is using SN65DSI83 and met a display issue.

Issue: SN65DSI83 exhibits a probabilistic no‑display issue after system reboot. Only backlight is lit while the panel shows black.
Observation:
1. When the failure reproduces, the chip’s built‑in test pattern mode works correctly.
2. Black screen occurs after switching back to normal video mode.
3. Initialization sequence is believed to be correct. Register 0xE5 reads 0 at failure time.
Could you help provide debug guidance for this failure?

Configuration and Log are attached. Also loop cusotmer Mr. Lu here for disucssion.

MipiToLvds.c journal-A2026091401-black-level7-20261104.log 

Thanks a lot.

Best regards,

Yang

 

  • Hi Yang,

    Sorry to hear you are having black screen issues, I will assist you with this issue.

    SN65DSI83 exhibits a probabilistic no‑display issue after system reboot. Only backlight is lit while the panel shows black.

    When you say probabilistic, do you mean 100% of the time or a different reproduction percentage? If so, what is the reproduction rate?

    1. When the failure reproduces, the chip’s built‑in test pattern mode works correctly.
    2. Black screen occurs after switching back to normal video mode.

    This helps us isolate the issue to the DSI interface, since the LVDS datapath works correctly when internal pattern generation is being used. Is the device being re-initialized when the switch to normal video happens? The datasheet stresses that the complete initialization sequence must be followed on every startup.

    3. Initialization sequence is believed to be correct. Register 0xE5 reads 0 at failure time.

    Are you able to verify the LP11 state at initialization on a scope? We have seen a similar issue in the past where a customer had an intermittent start up failure that was traced to the DSI data lanes not being in LP11 state at initialization step 2. 

    Additionally, register 0xE5 = 0 isn't 100% conclusive since the datasheet requires writing 0xFF to 0xE5 after DSI lanes are in HS and video has started (wait 5 ms), waiting 1 ms, then re-reading it; if nonzero, reinitialize from step 2. If 0xE5 was read before the clear-then-reread step, stale/empty status is expected. 

    Could you help provide debug guidance for this failure?

    Could you please provide / do the following?:

    • Provide a full register dump in a normal case where video is working AND in a case where the black screen issue is occurring?
    • (as mentioned above) Verify the LP11 state at initialization step 2 on a scope?
    • Verify the line time (time from HSYNC to HSYNC) on the DSI input matches the line time on the LVDS output using a scope.
    • Verify the register programming with DSI Tuner tool since this is the recommended way to program this device. 

    Additionally, does the failure occur only on cold reboot, or also after sleep/wake — and is the host re-running the full initialization sequence in both cases?

    BR,

    Cameron