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.

DLP9000: about repeating only one pattern

Part Number: DLP9000

Hi, I have question about repeating just one all-on pattern. I know keeping the mirrors at one state for too long is not good, but sometimes I do need long time exposure. My question is if I only have one pattern, and I keep repeating it, is there any dark loading time between the repeats during which the mirrors can have a break. If so, I can just set a small on-time to avoid long time staying at one state.

I am not sure I am clear, here is an example:

I can set the on-time to be 10.5 ms, and there will be 105 us after that during which the mirror can relax and have a rest. And I got 100:1 duty cycle, the frequency will be 1/ (10500+105). This will be different to the case If I set on-time to be 525 ms. the duty cycle will be 5000:1, frequency will be 1/(525000+105). 

But if there is not such a 105us between the repeats, the time length doesn't matter anymore, because the mirror will always be at one state. Duty cycles will be 100:0 and no frequency anymore. 

Please leave some inputs, greatly appreciated. 

  • Hello Yanfei,

    Yes, it would be best to have a small dark time to allow the mirrors to perform a crossover transition before then loading the all-on pattern. This approach will help to minimize hinge memory for all the mirrors in the array. There is no specific time that you need to allow, but a duty cycle of 1/99 (or 99/1 in the case of 99 On and 1 Off) is the best approach.

    Hope this helps. Please let us know if you have further questions.


  • Hi, Matt, thanks for you reply. Just want to confirm, you answer indicates the following:

    If I set the dark time to 0, the mirror will always be on one state. There will not be any dark time between each cycle added automatically at the end of the cycle, right? 

    Thanks again

  • Hi Yanfei,

    Let us double check for you. We need to review the chipset timing parameters. Should have an answer today or tomorrow.



  • Hello Yanfei,

    If you set a dark time it must be 105 us or greater.  If you set 0 then it will still receive a mirror clocking pulse every n us where n is the exposure time.  This is better than having very long on periods like 11 sec.  Setting n to 105 us with no dark time is better than every 11 sec with no dark time.   

    Are there periods where you do not need to have the exposure time.  For example if you ran at 105 us with no dark time for 5 minutes while taking readings and then set 50/50 duty cycle for 10 minutes or even 105 us on with 10.5 ms dark time for 5 minutes when no readings are being taken.

    Is there a particular requirement on the duration that you need to run the high on/off duty cycle?


  • Hi, Fizix, Thanks for your help

    I am working on a microscope built by previous engineer who have left. There is a problem when rolling shutter is used for the camera as aliasing was observed. Both the camera and the DLP are not triggering each other, and therefore aren't synchronized. I know it's a bad design. But at this point, I am trying to see where exactly the problem is. Could be different clock times. But before I conclude that, I want to rule out that possibility that there is a hidden time after each cycle. If I repeat one pattern with 105us + 0us times, does it take 10500 us for 100 cycles. 

    According to your reply, I assume this: There is not additional time after each repeat. So it does take 10500 us for 100 cycles. BUT, within the 105us, the mirror undergoes some changes, like going to parking first then go back on. That's why it's better than longer exposure times. But they all happens within the 105 us. Am I right about this? Thanks again, to both Matt and you

  • Hello Yanfel,

    If there is no dark time then the DMD only recieves a "mirror clocking pulse"  (MCP - reset waveform).  This wiggles the mirrors slightly, but does not "park" them in any way.  This is better than a truly static image with no MCP.

    A rolling shutter is not really compatible with the PWM used by the DMD if you are attempting to use 8-bit grayscale values.

    However, if you can take one image per binary pattern on the DMD then you could potentially synchronize the camera exposures to those patterns.

    I think you are going to have to have some kind of synchronization.


  • Thanks, Fizix. That answers my question.I agree, synchronization is need definitely.

  • Thank you Yanfei. 

    Hope you have a good weekend.