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.

DLP4710EVM-LC: Defective displays?

Part Number: DLP4710EVM-LC
Other Parts Discussed in Thread: DLP4710

Hello TI engineers,

We just bought 2 more DLP4710EVM-LCs. We had bought one 6 months ago, and it is working just fine. However, these two both have (different) problems.

1. One of the units does not display what we expect in 1440 Hz mode. Pixel value 0 is displayed at an intermediate luminance level, instead of being very dark as it is in the other two DLP4710EVM-LCs that we have. Pixel value 0x001 is somewhat more luminous. All the other bits (0x002, 0x004, ..., 0x800), when set individually, yield the highest luminance. This is in contrast to our other DLP4710EVM-LCs, where 0 is very dark, and all the other 24 bits (0x001, 0x002, 0x004, ..., 0x800) yield equal luminance values. It's not a problem with the video signal that we're sending, because we're using the same machine and the same software for testing all three units. I've re-flashed the firmware just in case, but that doesn't change anything. Do you have any idea what could be wrong?

2. The second unit doesn't have this problem, but it does have another one: two dead pixels on arrival. Do you consider that enough of a defect to ask the distributor (Mouser) for an exchange?

Thank you for your help,

Mark Wexler

  • Hello Mark,

    1. Could you please clarify what you mean by pixel value 0,2, 4, etc? How are you sending these values to the EVM and in what mode.

    2. Please let us check into this. 

    Thanks,
    Kyle

  • Sorry, my description wasn't very clear. I'm sending a video signal in 24-bit color mode. In 1440 Hz mode, the least-significant blue bit plane is the first of the 24 images, the next-most-significant blue bit plane is the second image, etc. As a quick check, I display an image with 24 vertical stripes, each stripe in one of the 24 bit planes, i.e., the (R,G,B) colors of the stripes being

    (0,0,1) (0,0,2) (0,0,4) (0,0,8) (0,0,16) (0,0,32) (0,0,64) (0,0,128) (0,1,0) (0,2,0) (0,4,0) (0,8,0) (0,16,0) (0,32,0) (0,64,0) (0,128,0) (1,0,0) (2,0,0) (4,0,0) (8,0,0) (16,0,0) (32,0,0) (64,0,0) (128,0,0)

    If everything's working correctly (as it is for the other two projectors), then the 24 stripes should all have the same luminance. However, on the unit in question, the (0,0,1) stripe is dimmer than all the others, while the remaining 23 all have the same luminance. Moreover, the (0,0,0) value doesn't yield black (or nearly black) pixels, but some intermediate luminosity.

    I hope that this is clearer. Please let me know if I need to clarify anything else.

    Mark

  • ... And about the defective pixels: it turns out there are more than 2. There's actually a blob of several contiguous defective pixels, and another isolated defective pixel.

  • Hi Mark,

    There is image quality specifcatuino for the EVM. However DMD datasheet  has specification for allowed defective pixel. Please refer to the " 6.11 Micromirror Array Optical Characteristics" section of DLP4710 datasheet. 

    http://www.ti.com/lit/ds/symlink/dlp4710.pdf

    If the DMD does not meet specification  of the DMD and the EVM is with-in warranty period (90 days), please contact source of purchase for RMA.

    The other issue

    (0,0,1) stripe is dimmer  - Could you please a similar experiment in 8 Bit mode and let the image have a gray value strip of zero, 1 and few other value?  Lets us know what do you observe?

    regards,

    Vivek

  • Hello Vivek,

    1. Unit with broken pixels

    Your document says that maximum adjacent out-of-specification micromirrors should be 0. Does that mean that if there are any bad pixels immediately touching any other bad pixels, then the DMD does not meet specification? I think this is the case: I've now found 1 isolated bad pixel, and two blobs of 2-3 bad pixels.

    2. Unit with weird luminosities

    I've put it into 8-bit mode (180 Hz monochrome), and display (R,G,B) values of:

    top row: (0,0,0) (0,0,1) (0,0,2) (0,0,3) ...
    mid row: (0,0,0) (0,1,0) (0,2,0) (0,3,0) ...
    bot row: (0,0,0) (1,0,0) (2,0,0) (3,0,0) ...

    This is what it looks like (with some artifacts from my camera's CCD):

    All three rows should have the same luminance profiles. But, as you can see, in the bottom two rows ("red" and "green"), you can see the difference between 0, 1, 2, 3, 4, etc. In the top ("blue") row, on the other hand, 0 and 1 look the same (and they both look like 0 in the bottom two rows), 2 and 3 look the same (they look like 2 in the bottom two rows), etc. So it looks like, basically, the second-least-significant blue bit isn't being read (as if its value were always 0).

    In 1-bit mode (1440 Hz monochrome), on the other hand, I have the impression that the least-significant blue bit is always on (because 0 is too bright, 1 is the same as 0, and 2, 4, 8, etc. are all the same and about twice as bright as 0 and 1).

    Mark

  • One more thing: I just noticed that you updated the firmware a couple of days ago. I flashed the projector with the luminance problem with it, but it didn't correct the problem.

  • Hi Mark,

    Could you please also run the same experiment on good unit and share the picture?

    regards,

    Vivek

  • Hello Vivek,

    Here's the same image on our good unit:

    It's not the same lighting conditions as the other photo (I apologize for their quality), but you can clearly see that, for this unit, the stripes in the upper row have the same width as those in the lower two rows -- whereas, on the bad unit, the stripes in the upper row are twice as wide as those in the lower two rows. Since the columns in the upper row have colors (0,0,0), (0,0,1), (0,0,2), etc., I believe this means that the least-significant bit of the blue component isn't working.

    Mark

  • Hi Mark,

    Thank you for running test on good unit. I assume you are using same PC and setting on both EVMs. If that is true then it is very likely excursion in product parameters and can not be resolved by user.

    Please give me couple of days, I want to run few tests in lab and will get back. 

    regards,

    Vivek

  • Hello Vivek,

    Yes of course, it's the very same equipment and software in the two cases.

    In the meantime, could you please answer my other question, about the other unit: your specification document says that maximum number of adjacent out-of-specification micromirrors should be 0. Does that mean that if there are any bad pixels immediately touching any other bad pixels, then the DMD does not meet specification? I think this is the case: I've now found 1 isolated bad pixel, and two blobs of 2-3 bad pixels.

    Thank you,

    Mark

  • Hi Mark,

    EVM with bad pixels - Could you please post a all black and all white picture from the EVM?  When was this EVM purchased?

    regards,

    Vivek

  • Hello Vivek,

    Here's an all-black picture:

    The areas with the bad pixels are circled. It's hard to see the individual pixels, but you can see that the one on top is larger than the one on the bottom. I think that the one on top is two bad pixels next to each other, whereas the one on the bottom is one bad pixel. Here are the two bad areas magnified (to the same scale). First the top one:

    and the bottom one:

    Here you can easily see that the top one looks like its twice as big. It's quite noticeable in the dark, and pretty devastating for our application.

    There is also a third blob of (I think) 3 bad black pixels next to each other, but I couldn't capture them with my low-quality camera.

    The units (this one, with the bad pixels, and the other one we were talking about before) were bought from Mouser a few weeks ago, but I was only able to check them last week. Two additional units that we bought (one 6 months ago, another one last week) are fine.

    Could you please let me know as soon as possible:

    - If the bad pixels in this unit are out of specification (because they're adjacent to one another)?

    - What to do with the other unit, the one with the bad bit?

    Thank you,

    Mark

  • Hi Mark,

    The TI warranty for EVMs is 90 days. Please see the section 8 of the terms of sales

    http://www.ti.com/legal/terms-of-sale/standard-evaluation-terms.html

    1. Unit with bad pixel - DMD is out of specification. Please reach out to Mouser and request for RMA if this unit is with-in warranty period

    2. Unit with 1 bit  image issue -  This sees to a system issue and can not be fixed by user. Please reach out Mouser and request for RMA if this unit is with-in warranty period.

    It is unfortunate that two EVM units had problem. I am sorry for inconvenience.

    regards,

    Vivek

  • Thank you, Vivek. I will ask Mouser to exchange the two units, and will refer them to this thread if they have any questions.

    Mark