Measuring irradiance

Discussion in 'The Projects Forum' started by yardleydobon, Jun 23, 2009.

  1. yardleydobon

    Thread Starter Member

    Apr 19, 2009
    19
    0
    I would like to measure irradiance. The sensor needs to be able to measure from around 350 nm up to 450 nm.

    I know there are photodiodes that can do this, however I am not able to find one that has a linear output versus wavelength, i.e. a flat spectral response. Can someone point me to one, please? :confused:

    Also can thermopiles measure at these wavelengths?

    Suggestions on other ways to get irradiance, such as measuring illuminance and converting, would be appreciated.
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    You may or may not find a device with a linear output. Fortunately, methods exist to linearize data.

    See if this company's stuff is interesting - http://www.taosinc.com/
     
  3. steveb

    Senior Member

    Jul 3, 2008
    2,433
    469
    I believe that wavelength range is very difficult to have flat response from any photodiode material and design. The simplest approach may be to use a silicon photodiode with a custom made optical filter to flatten the response.

    Do you have other specificatons such as power level, linearity required and frequency response (electrical bandwidth)? Are you trying to measure optical pulses, or average power?

    I think you can find thermopiles that work at this wavelength, but they are slow and easily corrupted by stray light at other wavelengths - hence filters may be needed here too.
     
  4. yardleydobon

    Thread Starter Member

    Apr 19, 2009
    19
    0
    I want to measure how much power an "ultraviolet" LED puts out. I don't have a spec sheet for the LED. All I know is that its peak wavelength is around 390 to 410.
    I want to measure the total irradiance.

    When an LEDs description says "wavelength :" and then lists a range like 390-410, that is only referring to the peak wavelength and not the entire bandwidth right?
     
  5. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    The structure of a LED confines the output to some strict limits. That is the greater part of the output bandwidth (probably 99%).

    Have you followed that link yet?
     
  6. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,672
    899
    Have you considered chemical actinometry? (see: http://en.wikipedia.org/wiki/Actinometry )

    There are several other systems not mentioned in Wikipedia. The amount of conversion is quite linear with the right system. That said, I would still use a photodiode or photomultiplier tube and correct for the response, if I had to do a lot of them.

    John
     
  7. yardleydobon

    Thread Starter Member

    Apr 19, 2009
    19
    0
    I have a TSL230R. It's will detect wavelengths as short as 300 nm. How can I correct for it's responsitivity?

    I've thought of two ways:
    1) Assume all the power is centered at the peak wavelength.
    2) Assume the LEDs spectrum is gaussian shaped. Linearize the responsivity curve around 400nm, then multiply that line times my gaussian spectrum using graph. Then I could calculate the power that would result from an ideal flat response and the power from the nonideal response and compare the two numbers to get my correction factor. It looks like the responsivity curve for this datasheet is not in A/W vs nm, so I'm not sure how I'd get this method to work out.

    Are there any other correction factors (in the datasheet or not) that I need to worry about?

    Thanks for the lead on actinometry.
     
  8. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,672
    899
    How accurate do you need to be? We had datasheets for our photomultipliers, but we still calibrated them using chemical actinometry to get corrected response curves when we needed to know quantum efficiencies accurately. It was a PITA to do.

    As you point out, LEDs almost monochromatic. You might be able to calibrate at just one wavelength. I would still do a calibration rather than rely on the manufacturer's data as a way to control all the other variables in the measurement system.

    John
     
  9. yardleydobon

    Thread Starter Member

    Apr 19, 2009
    19
    0
    I'm not very certain how accurate I need it to be. Probably as high as 20% would be OK. I've been calculating theoretical radiance by converting candela to lumens to watts to irradiance. That should be a decent check on actual measurements.
    I think I'll buy a blue-enhance photodiode and solar cell as well. I need to look into how to convert luminance (which can be measure with a digital camera using CHDK) to illuminance to irradiance.

    I really came here because I couldn't understand why a flat-response photodiode was so hard to find.

    Are there light sources that put out a known accurate amout of watts that I could use as a reference?

    Thanks for your help.
     
Loading...