• State
  • Locked Locked
  • Replies 11 replies
  • Subscribers 29 subscribers
  • Views 265 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DLP6500FYE: Nonlinearity of the display's amplitude characteristics

Remigiusz Gajda
Prodigy 10 points
Part Number: DLP6500FYE
Other Parts Discussed in Thread: DLP6500FLQ

Tool/software:

Hi,

we use DLP6500 to expose 8-bit patterns on photoresist using a 405nm laser.

DLP 6500 works in Video Pattern mode and the laser is controlled by the LED R output.
Unfortunately, the characteristics of each RGB channel are non-linear as shown in the figure below:

 


I made the measurement by displaying uniform gray boards, the integration time for each board was 100ms.
Pattern Mode settings: depth 8bit, start bit R0, exposure 4046us, dark time 0, Clear: RED, Frame change checked.
The visible depression and peaks in the characteristics are clearly visible in the exposures:

 

Displayed image (not exactly the same)

 

Irregular "rings" (light grey) can be seen in the exposure (the irregularity results from the additional application of a bitmap to equalize the lighting), 
which result from the non-linear characteristics. They should not be visible.
We tried different Video Pattern Mode settings on different channels without any positive results. 
What determines this behavior of the DLP6500 and how can it be fixed?
Regards,
Remigiusz 
 
 
  • 0
    TI__Guru 51395 points

    Hello Remiqiusz,

    What are you using to determine the levels?  Are you using a camera or a photodetector?  Since you are using laser do you observe any constructive effects (i.e. speckle) at the target surface.  There is some non-linearity, but it should not be this large.

    Even with the nonlinearity, I would expect the edges to move in or out radially with non-linearities, but not to have irregular edges as shown in your displayed image.

    Fizix

  • 0 in reply to Fizix
    Prodigy 10 points

    Hello.

    I use a photodetector with an adjustable integration time. 

    The 100ms integration time should smooth out the ghosting and noise from switching, which it does to a great extent. However, the decisive factor is the exposure, which is definitely non-linear. 

    Below is a cross-section of the illuminated structure. The structure was exposed in Video mode, so you can see different noise.

    We can compensate for smooth non-linearity by displaying appropriately modified grays at the expense of dynamics, but such "jumps" are more difficult to remove.

    Note to all post followers: non-linearity refers to deviation from a straight line.  The cross-section should look like the red line below. The curvature is the result of exposure to the Fresnel lens.

    Regards,

    Remigiusz

  • 0 in reply to Remigiusz Gajda
    TI__Guru 51395 points

    Dear Remigiusz,

    I would like to talk to our optics team, but I am concerned about the effect of speckle at the photodiode.  What is the area of the photodiode?

    If you happen to know the model of the photodetector that would be helpful (like a ThorLabs one or other).

    Fizix

  • 0 in reply to Fizix
    TI__Guru 51395 points

    Dear Remigiusz,

    Also, often we will defocus the image to mix the values together more.  If I understood correctly, you are setting the whole image to one gray level at a time.  Is this correct?

    Fizix

  • 0 in reply to Fizix
    Prodigy 10 points

    Photodiode Hamamatsu, surface area up to approx. 1 cm2. The detector is our design. All the light from the DLP is focused onto the detector using a lens.

    We measure the response of the entire panel, all pixels have the same shade.

  • 0 in reply to Fizix
    Prodigy 10 points

    I found similiar measurements: e2e.ti.com/.../1349204

  • 0 in reply to Remigiusz Gajda
    TI__Guru 51395 points

    Dear Remigiusz,

    Thank you for your answers.  This will help me as I consult with the optical team.  It sounds like you are doing things the way we would.  Please allow a day or two to consult with the optical team.

    Please verify that you are indeed using the Type A DLP6500FLQ DMD and not the WLP DLP6500FYE DMD.

    Fizix

  • 0 in reply to Fizix
    Prodigy 10 points

    Thank you for drawing attention to the display model. It is DLP6500FYE not FLQ. Sorry for the confusion.
    I hope this doesn't significantly change our discussion.

  • 0 in reply to Remigiusz Gajda
    TI__Guru 51395 points

    Dear Remigiusz,

    It does not change it significantly because I had not yet consulted the optics team yet, but it will be helpful to talk to them about the actual DMD.

    I did change the title of the thread to FYE.

    Fizix

  • 0 in reply to Fizix
    Prodigy 10 points

    Ok, thanks Fizix!

    For my part, on Wednesday I will conduct experiments in Pattern mode, exposing individual bit planes separately to simulate an 8-bit image.

    Best regards,

    Remigiusz

  • 0 in reply to Remigiusz Gajda
    TI__Guru 51395 points

    Dear Remigiusz,

    When you complete the experiments please let us know the results.  In the meantime, I will consult with the optics team to see if they have suggestions.

    Fizix