Photodiode variable-gain TI amp design

Discussion in 'Analog & Mixed-Signal Design' started by TLS3, Nov 10, 2018.

  1. TLS3

    Thread Starter New Member

    Nov 10, 2018
    5
    0
    Dear all,
    For research purposes, I need to acquire near-infrared light emission over several decades, starting at very low values.
    I plan to use a FCI-InGaAs-120 photodiode in photovoltaic mode. Light in the 1100nm field focused on the diode should range from 5nW to 50mW. The design of a TI-amp eventually followed by a non inverting amp present no difficulties (TI notes and circuit examples), only a big (variable) resistor is needed. OP-Amp requires extremely low input current (~pA) and low noise (<10nV/sqrt(Hz)).
    Considering the acquisition circuit will be in a remote place, the gain of the transimpedance amplifier has to be set by some means. So far I considered motorized potentiometers but their lack of feedback and 20% tolerance make the design weak though easy. A set of precision resistances shorted by relays may induce noise in my very weak signal.
    Could someone help me with a solution asides VCR2N (too low range) and LT1228 (too much input current)?
    Thanks a lot.
     
  2. danadak

    Well-Known Member

    Mar 10, 2018
    1,750
    359
    What kind of BW are you looking for ?


    There are V OpAmps available in the Femtoamp bias current region.


    Page 41 and page 45 might be of help -

    https://www.ti.com/ww/en/bobpease/assets/www-national-com_rap.pdf


    If you go to digikey.com you can filter OpAmp selections by a number of parameters
    to help choosing the right one.


    Regards, Dana.
     
  3. TLS3

    Thread Starter New Member

    Nov 10, 2018
    5
    0
    Hello Dana.
    So far, no bandwidth; I'm looking of the de-excitation of a phosphor, with a rate of several billions emitted photons/seconds. I just care about the light-to-intensity output the photodiode. If the signal was indeed too weak to be monitored online, an capacitor working as an integrator may be flushed after some milliseconds of so.
    The OP-amp selected so far were the OPA602 (my own), LMC6001 (application note OSI) or OPA128 (application note TI).
    I'm afraid a JFET would induce a significant capacitance + Johnson noise, rather than a precision 20M resistor.
     
  4. danadak

    Well-Known Member

    Mar 10, 2018
    1,750
    359
  5. TLS3

    Thread Starter New Member

    Nov 10, 2018
    5
    0
    Thanks for the tip, but the bandwidth or the low capacitance of the photodiode is not (yet) an issue.
    I focus more on changing remotely the gain of the trans-impedance amplifier to ensure signal strength over 4 decades of illumination (and prevent signal saturation).
     
  6. danadak

    Well-Known Member

    Mar 10, 2018
    1,750
    359
    AGC Loop driving a JFET acting as a variable R in the fdbk loop ?


    Regards, Dana.
     
  7. crutschow

    Expert

    Mar 14, 2008
    20,067
    5,652
    What is the minimum and maximum amplifier output voltage that you want?
    What is the dynamic range (resolution) of the acquisition system?
     
  8. TLS3

    Thread Starter New Member

    Nov 10, 2018
    5
    0
    Hello Crutschow.
    The desired output voltage should range from 5mV to 13V (rail to rail 15V).
    In steady state condition, the impacting light should have 5nW of power at 1300nm. The high gain let me see light intensity fluctuation within 10% (qualitatively).
    Eventually, a pulse rising 300 000nW on the same photodiode/circuit with a 10ms FWHM should be detected, and the gain of the TI-amp modified to see the Gaussian deposited energy curve.

    Solution 1: automatic gain control with a JFET as stated by Dana.
    Solution 2: AGC with a resistive photocoupler (LED + CdS)
    Solution 3: Two channels : one on the TI-amp for high-pulse acquisition and the other at the end a non-inverting amp for low power.
     
  9. KeepItSimpleStupid

    Distinguished Member

    Mar 4, 2014
    2,874
    505
    Keithley at one time had a Logarithmic picoammeter. Maybe this is the sort of circuit you need?

    Various fixed gains are possible. Without dealing with solid state solutions, mercury wetted relays do work. I used them.
     
  10. TLS3

    Thread Starter New Member

    Nov 10, 2018
    5
    0
    Alright, instead of a remotely variable gain TI-amp, a logarithmic amplifier looks like being a solution, as in sun-radiation meters of fiberoptics power monitors.
    I came across the LOG114 with a range from ~100pA to 10mA.
    I'll now order and modify a LOG114 development board (limited range from 10nA to 2mA)
     
Loading...