Photodiode current measurements

[DX400]

I am trying to build transconductance amplifier circuit to obtain values of the reverse biased photodiode (PD) current in order to estimate some of the LED characteristics.


At first, I have used the scheme named “Recommended Zero Biased Circuit” from page **broken link removed** (placed at the bottom of the page, the fourth one figure from the bottom ). The photodiode is Vishay BPW21R, OpAmp is LM358n, the value of the Rf control resistor of the negative feedback is 2 kOhm. The output voltage is measured by the multimeter. To obtain some signal, I am using flash light with 3W cool white LED like this https://www.sparkfun.com/products/13105. The flashlight is placed near in front of the photodiode to obtain the same luminous flux at each try.

When flashlight is off (no light) the output voltage (Uo) is about 0.

I have found that, the variation of the Rf (the gain of the amplifier) from 300 Ohm to 500 kOhm leads to not quite proportional change in output voltage (Uo). So, the values of the PD current are slightly different for the different Rf. What is the cause of it?

For example for the Rf of 2 kOhm I have obtain Vout of about 3.6 V and the PD current of 1.8 mA.


Then, I have used the scheme named “Positive bias circuit” from page **broken link removed** (placed at the bottom of the page, the second one figure from the bottom ). I have used 10 V which supply the OpAmp and to make photodiode reverse biased.

The value of Vout for the Rf of 2 kOhm decreases to 0.7 V for the same illumination.


So the problems:

1. What is the reason of output voltage drop.

2. I have something like overflow. When flashlight is off (no light) the Uo is about 0. Under flash The variation of the Rf, which defines the amplification gain of the circuit, from 1 to 100 kOhm does not change the value of the output voltage Uout of about 0.7.

3. The increase of the Rf up to 0.5-1 MOhm leads to much decrease of the Uo.

4. The variation of the Rf also leads to not proportional change of the output voltage. So the Rf does’t work as gain.

5. I don’t understand how to be certain about obtained results? How to calibrate the results of the PD current? How to make a calibrated light source for such measurements to be sure of correct PD current?


I should be very obliged for help to solve this problem.

 

[MikeMl]

Post your circuits. Your links dont work.

What is the supply voltage for the opamp?

 

[DX400]

Sorry. Corrected links:
https://home.sandiego.edu/~ekim/photodiode/pdtech.html

“Recommended Zero Biased Circuit”
https://home.sandiego.edu/~ekim/photodiode/fig9.gif

“Positive bias circuit”
https://home.sandiego.edu/~ekim/photodiode/fig11.gif

 

[DX400]

The supply voltage of the OppAmp is 10 volts unipolar in all cases.

 

[audioguru]

You used a single positive supply then how in earth can the output go negative as shown on the diagrams?? Their opamps probably have a dual polarity supply that is stupidly not shown.

 

[MikeMl]

V(out) for Rf= 3KΩ, 6KΩ, 12KΩ,..., 96KΩ

D1 is needed only for simulation. I1 represents the photodiode current:





When you understand the plot, you might have some insight into what your circuit is doing...

Questions you should be asking?

What is the input bias current of the opamp?
What is the output swing of the opamp?
What is the input common-mode range of the opamp?
What is the noise floor of the opamp?

 

[KeepItSimpleStupid]

I have found that, the variation of the Rf (the gain of the amplifier) from 300 Ohm to 500 kOhm leads to not quite proportional change in output voltage (Uo). So, the values of the PD current are slightly different for the different Rf. What is the cause of it?

RC needs to be the same for your tests, otherwise your changing the response of the amplifier.

Be careful, if the OP amp cannot drive a capacitive load, sometimes the OP amp will oscillate. One range of an I-V converter I designed was very unstable at say 20 mA from a solar cell. That result scared me to death.

The company I worked for designed an I-V converter for in house use and the IC selection is problematic. Their initial design used the CA3140. I changed it to an LF351 which had better performance.

My "Holy grail" design did have some bugs, but it outperformed anything else we ever had for low level AC measurements as the front-end to a lock-in amplifier.
I had intended to be able to zero the amplifier automagically and I incorporated no mechanical zero adjustments. When I set a D/A to zero, it;s output was non-zero. In order for the algorithm to work, I would need a "hard zero", but 40 pA of offset current was no bg deal for us in the DC mode. i was told the project was "done".

The "Holy grail" was also a true 4-terminal device with 4 ranges, biasing and suppression, For giggles it had an open circuit voltage mode and a 2T/4T mode.
Zero Check and Zero Correct didn't work.

Ib, Vos and the temperature dependence of Vos. With high values of the FB resistor, Vos is is very important without offset-nulling,

0.7 V is a magic number.