• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 3 answers
  • Subscribers 32 subscribers
  • Views 621 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

DLPC150: DLP210NIR supporting boards and programming

matthew garrett
Prodigy 20 points
Part Number: DLPC150
Other Parts Discussed in Thread: DLPA200, DLP2010NIR, CCSTUDIO, DLPNIRNANOEVM, DLPDLCR230NPEVM

We're beginning to design a system using DMDs for spectroscopy. The DLP2010NIR, DLPC150, and DLPA200/DLPA 2005, from my research, seem to be what we'd want to incorporate into the system.

We do not wish to use the DMDs in the typical manner described in the sample documentation, rasterizing a vertical line across the incident spectrum. Instead, we wish to use it to select portions of a sample to be measured, by blacking out and selecting portions of a sample, and feeding that into a point spectrometer. So, we'd do this by programming blocks of mirrors to be on or off, based on which portion of the sample we want to image, as the DMD reflects imaging light back into the spectrometer.

So my question is about the best products to order to set this up initially.  Obviously, we would need the DMD, PMIC, and DLP Controller. Are there demo boards available that would let us install the DLPA200 and DLPC150 onto them, connect them to the DMD, and then program the mirrors according to our wishes (presumably with CCSTUDIO)?

I see there's the DLPNIRNANOEVM NIRscan evaluation module, which seems to include parts similar to what I'm looking for. But, many of the parts on that we don't need, and we'd want to program the DMD in a way that doesn't seem to be immediately supported the way that the NIRscan module comes set up for, so we would have to completely disassemble (and reprogram?) that to get the function we want to demonstrate, which we'd rather not do. But, if we could acquire similar parts to the ones contained in that, especially the boards supporting the PMIC and controller, that seems to me like it would be a better option.

  • 0
    TI__Expert 3825 points

    Hi Matthew

    DLPNIRNANOEVM has the TIVA MCu as the host controller which generates the patterns(vertical/hazamard bv default).

    The TIVA firmware is open source and available on ti.com : https://www.ti.com/lit/zip/tidcc48 

    The instructions to build and debug the code is available in the user's guide: https://www.ti.com/lit/pdf/dlpu030

    The patterns are generated using the DLP Spectrum library: https://www.ti.com/lit/zip/tidcc49

    Spectrum library supports generating your own patterns by providing corresponding frame buffer as input. Refer to dlpspec_scan_genPatterns() api. I highly recommend you to go through spectrum library for proper input parameter structures and usage.

    Since the firmware is opensource. You can reprogram it to whatever pattern you want to generate.

    We only have this EVM available for both demo and evaluation. 

    What do you mean by many of the parts are not needed by you. Could you please elaborate which parts are those. So that we can confirm the same if those parts are essential or not for operatoin. 

    Regards,

    Akhil

  • 0 in reply to Akhil Guntumadugu
    Prodigy 20 points

    I may not have been completely clear; in our application, we want to put the DMD between the spectrometer and the sample we're imaging, rather than as part of a spectrometer.

    So, by parts on the EVM that we don't need, I mean that for the design we're currently pursuing (currently as a proof-of-concept), we want to use our own point spectrometer, pointed at the DMD array, imaging a reflection of light from the sample under test. That means we don't need the spectrometer parts of the EVM: the grating, the array, etc. What we need is the DMD array, and the hardware/software to control the mirrors on it. We can then mount the DMD like we need it, and prove that we can get data like this.

    It seems like you're saying that we can control the mirrors in any pattern we want with the TIVA firmware, which is good. That's one of the things we needed to know, and I did just download that and the spectrum library also you linked to.

    So for acquiring the hardware we need, it seems like we have three choices:
    (1) Buy the DLPNIRNANOEVM and disassemble it, removing the parts we don't need and rearranging the ones we do, then command the mirrors into our desired configurations with the TI software options.
    (2) Buy the DLPA200, DLPC150, and DLP2010NIR, design a custom board to connect them to each other, and connecting that assembly to a computer running the TIVA firmware.
    (3) Buy the three parts in (2), but also purchase some other hardware from TI that allows us to easily connect the three parts into a working, prototype-able assembly, and command the mirrors from there, with the TIVA firmware.

    My initial thoughts was that (3) seemed like the most convenient way to go about this. However, from your reply, it sounds like (3) may not be an option. It sounds like you don't make/sell that necessary hardware on its own?
    In that case, I'm thinking (1) could be the best for now, until we're further along in the design process, in which case we'd switch to pursuing option (2). Does that sound reasonable? Or is there a reason you know of why (1) would not work, or would otherwise be a bad idea? Or is there something else I'm missing that we could get that would make our process easier?

  • 0 in reply to matthew garrett
    TI__Genius 12785 points

    Hello Matthew,

    The parts that you have mentioned may be necessary for the EVM to function. Please allow us until next week to consider if this is feasible.

    Regards,

    Austin

  • 0 in reply to Austin Snyder
    Prodigy 20 points

    Thank you for continuing to look into this.

    For reference, we already did a similar experiment in the visible spectrum. We took an DLPDLCR230NPEVM, disconnected the LEDs, and put fiber optic cables in their place. Then we were able to run light through the system in the opposite direction, and have indirect control over the DMD mirrors by loading different color images. Some "unnecessary" things had to stay intact, like for example the fan had to be attached for the EVM to function. But we could just move the fan out of the way, which was ok enough for then. Something similar is what we'd be initially pursuing with the 2010NIR EVM, if we go that route.

    But, out of curiosity, what do people normally do with they begin to design a system around the DMD, PMIC, and controller? Do they typically just build their own boards around these products every time? There is no demo board or some similar product available just to connect those those three things, and design a mechanical and optical system around that?

  • 0 in reply to matthew garrett
    TI__Expert 7035 points

    Hello Mathew,

    Usually designers will have an EVM to get an idea of what they can do with the chipset. TI has reference designs for the EVMs that can be used. There are third parties that have designs based on the EVMs as well.

    It is also up to the designer on who to contact for optical engines and how they want to build around that part of the system.

    Best,

    John

  • 0 in reply to matthew garrett
    TI__Expert 3825 points

    Hi Matthew,

    You have rightly pointed out the available choices.

    All three of the options are possible, I have listed down the challenges and pre-cautions for each of the option

    For option 1: Disassemble EVM

    • TIVA MCU Firmware needs to be modified. To ensure the system boots up even if some modules are not connected.
    • Connections within the chipsets i.e DLPA200, DLPC150 and DLP2010NIR should be intact.
    • Since you are planning to remove the grating and the photodetector in your case. Care must be taken with respect to optical engine because every EVM has been factory calibrated for the slit, distortions and wavelength. You may have to calibrate by yourself again.

    For option 2: Custom board with chipset and Computer

    • Firstly you can not use the same TIVA firmware directly from computer. Rather you may have to develop your python scripts or code to communicate with DLPC150 via I2C interface. You can refer to TIVA Firmware on the sequence of commands to do so.
    • ou may need to calibrate the optical engine for distortions

    For option 3: Custom board with chipset and TIVA MCU

    • The custom board schematic should be meeting our reference design and you may need to calibrate the optical engine for distortions.

    Option 3 would be the best as you mentioned. We do have the chipset and TIVA MCU available. It is just we do not have a combined solution other than the EVM. So As per option 2 you may have to design your own custom board. 


    Regards,

    Akhil

  • 0 in reply to matthew garrett
    TI__Expert 3825 points

    Hi Matthew,

    Most of the customers usually by the optical engine from manufactures which comes with the DMD itself.
    Then they design their own custom board with Host MCU of their choice, PMIC and DMD Controller.

    We do not have any demo boards other than EVMs.

    You can search for third party OMM(optical module manufactures): https://www.ti.com/tool/DLP-OMM-SEARCH

    Also information about third party EVM manufactures is available in the DLP2010BIR page as shown below.


    Regards,

    Akhil