DLP4710EVM-LC: PatternData generation/display trouble with DLPC-API V1.5

Prodigy 175 points

Replies: 5

Views: 76

Part Number: DLP4710EVM-LC

Hello,

we want to use the DLP 4710 EVM-LC to display 8 bit patterns in internal pattern mode.

We started using the DLPC-API1.5, but already have trouble with the sample code.

When we want multiple 8 bit patters, for example NUM_EIGHT_BIT_HORIZONTAL_PATTERNS = 6 , the projector display 6 patterns, but not the ones we expected. We know how the genereted patters are supposed to look like, since we can see the code. This way we figured out, that the projector doesn't display the patterns 0 ,1, 2, 3, 4, 5. It displays 3, 3, 4, 4, 5, 5 (figured out, by counting the bars in the images). We tried to find the cause of this behaviour, the Patters are populated correct, the pixel data in the arrays is as it should be. We even had a look at the function WritePatternDate and inspected the binary data send for each pattern, this also sends the data for the patters 0 ,1, 2, 3, 4, 5 to the projector.

When we upload the generated .bin file with the GUI, we also get the same wrong patterns.

Do you know any other possible cause for this error? We just downloaded the API, set NUM_EIGHT_BIT_HORIZONTAL_PATTERNS = 6 (Error also accures with NUM_EIGHT_BIT_HORIZONTAL_PATTERNS = 4 (out of the Box Value)) and run the code.

Kind regards

Lennart

5 Replies

  • Hello Lennart,

    Can you please let me know the firmware version that you are using currently?

    Also, kindly share the pattern binary with us.

    Regards,

    Mayank 

    If a post answers your question, please click on "Verify Answer" button

    Have any questions about DLP Pico Chipsets? Check out the DLP Pico Universal FAQ for quick answers: https://e2e.ti.com/support/dlp/f/94/t/946506 

  • In reply to Mayank Khandelwal:

    Hello Mayank,

    I am currently using a version 8.1.0 adapted by TI with all LEDs disabled.

    I couldn't attached the binary data, I generated a link to the file. As I said, this also includes the Error.

    https://cloud.imr.uni-hannover.de/index.php/s/GpLBHBQNND4yZKJ

    Following I send a link to the adapted API. It includes only small changes to the TI version.

    -6 instead of 4 8-bits patters

    -Trigger IN and Trigger OUT configuration

    -new pattern population funtion for a SinWave Pattern

    https://cloud.imr.uni-hannover.de/index.php/s/rRit4L5bLPH26JY

    Regards,

    Lennart

  • In reply to Lennart Denecke:

    Lennart,

    Thanks for the uploads. We will investigate the files attached and provide feedback in the next couple of days.

    Regards,

    Philippe Dollo

    --------------------------------------------------------------------------------------------------------

    If a post answers your question, please click the "Verify Answer" button.

    Have questions about DLP Pico Chipsets? Check out the DLP Pico Universal FAQ for quick answers: https://e2e.ti.com/support/dlp/f/94/t/946506 

  • In reply to Philippe Dollo:

    Hello Philippe.

    We downloaded the API one more time and used CLion to run it.

    We ran into antoher problem. We are still using the internal pattern mode.

    Setting: 2 Patternsets,  6  8-Bit Vertical Patterns, 6  8-Bit Vertical Patterns (But problems occure with all setting, also the default setting in the newly downloaded API)

    I attached some pictures, which shows the projected images. As you can see, the images are divided excatly in the midle and both halfs show different patterns. We think this is due to some problems with the master and slave controler.

    It is hard to explain, but there are some paralles to the earlier decribed problem. We want to show image 1 -> 2 -> 3 -> 4 -> 5 -> 6. What we get are images [Left Half / Reight Half]    [4/1] -> [5/1] -> [5/2] -> [6/2] -> [6/3] -> [7/3] -> [7/4]. When only looking at the left half, this is the same problem as discribed earlier. But earlier the right half displayed the same as the left half. Do you have any idea why those problems? Maybe this further discribtion halps solving the new prblem and the one discribed earlier. How is this even possible? Differenc between left and right half? From our point of view this can only be explained in a problem with the data tranfer to the controlers. But we can't find any passages in the code, which explains us why this happens.

    When we are using the GUI we get normal results. When we save a file from the GUI and load it with the API we get the same results as with teh GUI. Only when we generate the pattern data in the API we are having those trouble.

    Regards Lennart

    Examples:

  • In reply to Lennart Denecke:

    Lennart,

    Apologies for the late reply. We were able to reproduce the first problem which you had reported at our end. The fix for this problem will be available in the next maintenance release of the firmware. 

    Regarding the second problem, are you sure that the logic which you are using to generate the patterns is correct? Can you attach the modified code here so that we can further look into it?

    Regards,

    Mayank

    If a post answers your question, please click on "Verify Answer" button

    Have any questions about DLP Pico Chipsets? Check out the DLP Pico Universal FAQ for quick answers: https://e2e.ti.com/support/dlp/f/94/t/946506