• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 27 subscribers
  • Views 610 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.

DLPNIRNANOEVM: How to achieve fastest scan rate

Salvatore Polito
Prodigy 35 points
Part Number: DLPNIRNANOEVM
Other Parts Discussed in Thread: DLP2010NIR, DLPC150

We would like to use the DLPNANOEVM to acquire 100 spectra per second in the range 1150- 1350nm with 12nm resolution. Looking at the GUI, the minimum exposure time is 0.635ms so it should be feasible.

When trying these settings we realize that the total scan time is much longer than expected (even keeping the lamp on and using  a fixed PGA gain).

What is actually preventing us to achieve the desired scan rate? is that feasible with the current FW/GUI provided with the DLPNANOEVM or do we need to modify the FW/GUI?

Thanks a lot in advance.

Salvatore

  • 0
    TI__Mastermind 22830 points
    Hi Salvatore,
    Welcome to DLP forum! Could you please share detail about GUI /FW version and parameter values you are using?
    regards,
    Vivek
  • 0 in reply to Vivek Thakur
    Prodigy 35 points

    Dear Vivek,

     thanks for your response and welcome. We are currently using GUI v.2.1.0, TIVA SW 2.0.1, DPLC FW 2.0.0, SPECTRUM LIBRARY v2.0.0.

    At the moment, we are using the parameters listed hereby and the GUI returns a Total scan time of 1.075 sec instead of our expected scan time of  0.635[ms] x 125 = 79.37msec. What are we missing here?

    Thanks a lot in advance. 

    Regards

    Salvatore

    - Num scans to Average 1

    - # back to Back Scan 1

    -  Keep Lamp ON (to shorten scan time)

    -  PGA gain 64 (to shorten scan time we do not use PGA auto)

    - Method Column

    - Start nm 900

    - End nm 1700

    - Width nm 12.88

    - Digital resolution 125

    - Exposure time 0.635ms

    - No. of scans to average 1

    - No. of sections 1

  • 0 in reply to Salvatore Polito
    Prodigy 35 points
    Just a quick clarification: the parameters listed in my previous post were just an example to show the inconsistency we experienced with timings. When we tried the spectral range specified in the first post (1150..1350nm), we used the following parameters and we got a total scan time of 0.991sec, which still does not match with our expectation: 0.635[ms] x 16 = 10.16ms (corresponding to ~100Hz scan rate).

    Thanks!

    - Num scans to Average 1

    - # back to Back Scan 1

    - Keep Lamp ON (to shorten scan time)

    - PGA gain 64 (to shorten scan time we do not use PGA auto)

    - Method Column

    - Start nm 1150

    - End nm 1350

    - Width nm 12.88

    - Digital resolution 16

    - Exposure time 0.635ms

    - No. of scans to average 1

    - No. of sections 1
  • 0 in reply to Salvatore Polito
    TI__Mastermind 22830 points
    Hi Savlatore,

    The DLP2010NIR DMD and DLPC150 chipset is capable of meeting your performance requirement i.e 100 spectra per second in the range 1150- 1350nm with 12nm resolution.

    However , the firmware flow in the DLPNIR Scan nano evaluation module is not configured to execute at this speed. You have to modify TIVA firmware and GUI for you application.

    Please refer to the section 5.2.2 (page 51) of "DLP® NIRscan™ Nano EVM User's Guide" (www.ti.com/.../dlpu030g.pdf) for total scan time. As you could see there is significant initialization overhead.
    1. 100 ms delay to avoid lamp power-on surge
    2. 625 ms for lamp to stablize
    3. time to determine PGA gain
    4. actual scan time
    5. ~100 ms overhead for sensor & processing

    This should explain why you are getting such high measure time for a single scan.

    I would suggest running following experiment:
    1. Take single scan measure with required setting. Measure total time (you current measure)
    2. Take 101 back to scan with lamp on and no delay. measure total time

    The difference between these two shows actual time for 100 scans.

    If the actual scan time for 100 scans is close to you requirement then you can just focus on how to make overhead as part of power up or eliminate with known values where applicable.

    If the actual scan time is still high then please also look at TIVA firmware to optimize time between scans.

    I hope this answers your questions and gives you idea how to proceed.

    Regards,
    Vivek