• State TI Thinks Resolved
  • Locked Locked
  • Replies 18 replies
  • Answers 5 answers
  • Subscribers 33 subscribers
  • Views 1816 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.

Original question:

DLPNIRNANOEVM: NirScanNano: Scan or reference data interpret failed

DLPNIRNANOEVM: Scan or interpret data filed after GUI software upgrade

Xing Zhou1
Prodigy 150 points
Part Number: DLPNIRNANOEVM
Other Parts Discussed in Thread: DLPC150

Hello there,

The device worked fine with GUI v1.1.9 and reported a "scan or interpret data failed" error after I upgraded to GUI v2.1.0. . How can a fix this issue?

What I have done:

- Installed the GUI 2.1.0

- Download and installed the DLPC150 firmware and Tiva firmware updates per instruction of the user guide section 2.3.

During the upgrade process I did not see the dialog "EEPROM cal data does not match..." as described in the user guide at the end of section 2.3. Therefore I don't think I have erased anything. (BTW, if it is relevant, I purchased this device directly from TI website.)

Thanks.

Xing

  • 0
    Prodigy 150 points

    BTW, can I do a "factory reset" somehow to downgrade the device to the v1.1.9 compatible status? 

  • 0 in reply to Xing Zhou1
    TI__Genius 9971 points

    Hi Xing,

    Welcome to TI E2E forums and thanks for showing interest in the DLP technology. Our expert will look into your query and get back to you soon.

    Thanks and regards,

    Mayank

  • 0 in reply to Mayank Khandelwal
    TI__Genius 10925 points

    Hi Xing, could you kindly perform the following:

    Kindly make sure you have flashed the latest DLPC150 FW (v2.2.0) and the latest TIVA FW (v2.1.0). Then, could you kindly take a new reference scan and then use this reference to take another scan (scan reference select - previous) and see if it still fails? if you are still getting the error, kindly let us know your OS that you are using..

    Thanks & Regards,
    Hirak. 

  • 0 in reply to Hirak Biswas
    Prodigy 150 points

    Thanks for the response. Yes, the versions of both FWs are as stated. For the reference scan, I read from another post that it requires a material coated with spectralon which I do not have and don't know how to get one. Is it possible for TI to mail me a piece of such material? Thanks.

  • 0 in reply to Xing Zhou1
    TI__Genius 10925 points

    Hi Xing, Yes that is correct. The way it works is:

    There is one Factory Reference Scan data stored on the flash memory, and there is another reference scan data slot that you can use to store your own reference scan data. Even if you scan another reference, the factory reference scan data should be unmodified.

    If somehow your factory reference scan data gets corrupted, then the scan interpretation may fail. This is why I requested you to scan a reference (even if you do not have spectralon material) and then use the reference scan to interpret. If it does not fail, this will mean that your factory reference data is corrupted. If it still fails, then the problem is elsewhere.

    On a separate note, the scan results may shift with changing humidity, temperature, and lamp age. It is suggested to take new reference scans regularly to maintain system accuracy.

    Thanks & Regards,
    Hirak. 

  • 0 in reply to Hirak Biswas
    Prodigy 150 points

    Thanks for the clarification. I can confirm that it worked after I took a reference scan. Therefore, it is almost certain the factory reference is corrupted. What is the best way to get this fixed? Thanks.

  • 0 in reply to Xing Zhou1
    Prodigy 150 points

    Just note that this thread is marked at "TI Thinks Resolved". My question is what is the best way to restore the factory reference? My reference is just taken by a shining object, not spectralon coated. What is the consequence of using a low-quality reference scan?

  • 0 in reply to Xing Zhou1
    TI__Genius 10925 points

    Hi Xing, the only way to get a proper reference is to use a 99%+ reflective material like spectralon material. Otherwise, the results will be skewed with the skew varying wavelengths. If you can get this material, then only periodic reference scan and using this material and using this reference scan for new scans should give accurate results. 

    Thanks & Regards,
    Hirak.

  • 0 in reply to Hirak Biswas
    Prodigy 150 points

    Understand. But by looking at Amazon, such material is very expensive. Given the factory data was corrupted during software update per official instruction, would it be possible for TI to help repair/restore factory reference? The device was purchased directly from TI website. Thanks.

  • 0 in reply to Xing Zhou1
    TI__Mastermind 28005 points

    Xing,

    I believe that each individual system is calibrated individually so you cannot restore it simply in this manner. Here is some discussion on the topic:

    https://e2e.ti.com/support/dlp/f/94/t/870563

    Regards,

    Philippe Dollo

  • 0 in reply to Philippe Dollo
    Prodigy 150 points

    Thanks for the response. This is why I was asking for repair or replacement. To me, it is something relatively easy or at least doable for TI to replace the factory reference (I see GUI has this option.) It is sounds a bit unreasonable to ask a customer to discard a $1000 device because some factory data was corrupted during an official software upgrade.

    BTW, does TI have a warranty policy?

  • 0 in reply to Xing Zhou1
    Prodigy 150 points

    Is there any update on this issue? I am in touch with your customer service team and they are waiting for engineer's recommendation before being able to proceed (e.g. repair? replace? etc) Thanks. Xing

  • 0 in reply to Xing Zhou1
    TI__Genius 10925 points

    Hi Xing, Let me discuss with our internal team and get back to you soon. Thanks for your patience!

    Thanks & Regards,

    Hirak.

  • 0 in reply to Xing Zhou1
    TI__Genius 10925 points

    No problem!

    Thanks & Regards,

    Hirak.

  • 0 in reply to Hirak Biswas
    TI__Mastermind 23080 points

    Hello Xing,

    Thank you for your patience. There are two types of the data stored in the EVM's non volatile memory:

    1. Calibration data - This has information about optical system, alignment etc. This can be calibrated only in factory.

    2. White  reference data - This data represents  lamp's intensity. It is captured by placing a NIR reflective standard (99% reflective) against  windows and taking a scan. This data is used to calculate absorvance when you scan a sample. The white point reference can be stored in one of the two ways - as factory reference in EWM's non-volatile memory or in PC's local memory. When you select "new" button in "GUI->Scan Reference Select", the GUI scan the sample as white point reference and stores in PC's local memory. Next time when  you scan a sample and select "previous" button , the software uses locally stored white point reference to calculate absorvance.

    Based on previous communication, it seems the EVM is working and only white reference data is corrupted. The "factory" white reference is intended to be overwritten by user. In fact , user should periodically take factory reference and update to reflect current behavior of lamp. The lamp ages with time  and its performance changes with time, humidity, temperature.

    Any spectral sensing application should incorporate periodically taking white reference otherwise your results will not be accurate.  

    The EVM  seems to functional and the "factory" white reference has got corrupted or overwritten accidentally. It is user programmable data and  correct use of the EVM will require periodic overwrite of factory reference by user.

     I hope this explains.

    regards,

    Vivek

  • 0 in reply to Vivek Thakur
    Prodigy 150 points

    Thanks for the clarification. Do you know any less expensive 99% reflective material? It appears that spectrulon costs $500+ on Amazon which does not make a lot of sense to spend just to restore some reference data. To me, a working product comprises both working "hardware" and required "software/reference data".

    Is it possible for me to send the device to TI and get the reference data scanned/stored there?

  • 0 in reply to Xing Zhou1
    TI__Mastermind 23080 points

    Hello Xing,

    TI does not offer any repair services for EVMs. Your request to reprogram  the 'white reference " can not not be supported.  Please note that it is an "Evaluation module". The intended purpose of evaluation module is to facilitate  evaluation of TI technology for an end application. 

    The capability to take  "white reference" is  fundamental to any reflection based spectral sensing application  because the performance profile of illumination source changes. Without a periodic white reference , performance of any application will  deteriorate with time and will cause unit to unit variation.

    You can use any reflective surface for reference scan. The issue is as you deviate from 99% reflection across wavelength of interest, you compromise on accuracy of your result.

    Some of the options for reference :

     1. 99% NIR reflective standards ex . spectrulon - Best option

    2. Fluorion 99- a n excellent alternative to standard - Please contact them directly - https://aviantechnologies.com/product/gray-scale-standards/

    3. Metallic reflective   surface coated with "white out" - Only good for proof of concept; Will not use it for any final application.

    4. NIR reflective material ex. light diffuser used in office lights etc - same as #3.

    Hope this answers your questions.

    regards,

    Vivek