Connect with us

photodiode shunt resistance equation

Discussion in 'Electronic Design' started by Wanderer, Feb 23, 2008.

Scroll to continue with content
  1. Wanderer

    Wanderer Guest

    What is the equation for the photodiode shunt resistance relation with
    temperature? I know it's logarithmic and got kT/q in it but I can't
    seem to find it.

    Thanks
     
  2. MooseFET

    MooseFET Guest

    Look for the (non-photo)diode equation.
     
  3. Wanderer

    Wanderer Guest

    Thanks for the reply. It's probably that simple and staring me in the
    face. I know the diode equation I don't see how to relate it to photo
    diode shunt resistance so I can get the shunt resistance noise. If I
    use ohms law and the bias voltage is zero I'm back to square one.

    At one point Aldert Van der Ziel in his book converts the dark current
    noise into a conductance to simply his model.

    Idark * q/2kT

    Is that 1/Rshunt? Is the shunt resistance just a different way to
    write the dark current noise?

    Thanks
     
  4. Wanderer wrote:
    (snip)
    I don't think so. The incremental shunt resistance is the
    slope of current as voltage changes. This is in addition to
    the approximately constant (but noisy) current that is
    almost independent of voltage but very dependent on temperature.
     
  5. Phil Hobbs

    Phil Hobbs Guest

    Dark current noise in photodiodes is almost never the limiting noise
    source in a measurement. What are you actually trying to measure?

    Cheers,

    Phil Hobbs
     
  6. Wanderer

    Wanderer Guest


    I'm building an incoming inspection tool for one of these

    http://jp.hamamatsu.com/products/division/ssd/pd041/pd064/pd065/S9295-01/index_en.html

    I'm computing the noise through a power spectral density matlab
    program. I see the dark current drop like it is suppose to, but I only
    get 6% change in noise from 25C to 5C. I'm told I should see a much
    greater change because of cooling the shunt resistance. I know it's
    just some bugs in my code, but I'd like to do some calculations to get
    an idea of expected results.

    Thanks
     
  7. Wanderer

    Wanderer Guest


    I don't know. Karl Spangenberg has an Appendix on shot noise in
    Fundementals of Electron Devices 1957(I have a bad habit of collecting
    old electronic books, though sometimes their better than the
    internet). He takes the derivative of the diode equation and gets a
    conductance qI/kT and calculates the rms noise to get 2qIB. Though he
    adds the disclaimer:

    "The above demonstration is not really a derivation but rather a
    rationalization. It assumes that the phenomenon giving rise to
    resistance noise is the same as that giving rise to emission noise. It
    will at least make the formula seem reasonable and dimensionally
    correct."
     
  8. Hello,

    How do you calculate the noise?

    Is it the integrated noise voltage over a certain bandwidth?

    What's the lower and upper frequency limit of your
    measurement and noise calculation?

    I remember that Rshunt will fall by a factor of two
    for every temperature rise of 6 to 10 degree.

    Best regards,
    Helmut
     
  9. Phil Hobbs

    Phil Hobbs Guest

    Judging from the very high feedback resistor and the rapidly rising
    noise curve on Page 2 of the datasheet, that gizmo is dominated by op
    amp noise above about 10 Hz.

    The photodiode capacitance and the feedback resistor form an RC
    differentiator that multiplies the voltage noise of the op amp, and
    (with this device) once you get as far as the audio region, the PD noise
    hardly enters the picture.

    Cheers,

    Phil Hobbs
     
  10. Wanderer

    Wanderer Guest

    I take the FFT over 100hz bandwidth convert it to a Power Spectral
    Density by taking the absolute value, throwing away half , multiplying
    by 2, etc. I then have a function which breaks up the data into
    different bandwidth regions, sums the points in the different regions,
    divides by the number of points in the region, by the bandwidth of the
    region,divides by 2 and takes the square root. I've been experimenting
    with different size regions.
     
  11. Wanderer

    Wanderer Guest

    It's a low frequency dc application. I'll try focusing my analysis
    below ten Hz. I've been focusing around 50Hz because I don't trust the
    edges of the window.

    Thanks
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-