2
\$\begingroup\$

I am using PIN photodiode BPW34 (1).

I am then using an opamp (LM324) (2) to amplify and get an output voltage from the current source generated through the photodiode.

Following is the circuit I am using to realise this.

enter image description here

Following is the output I am getting at particular frequencies. I am actually giving square wave input light signal to photodiode.Following image is the output we got from the digital oscilloscope(Left-400kHz,Right-2kHz).

I also tried changing parameters for Cf in the diagram and also putting direct square wave voltage source(Left image) instead of photodiode with equivalent resistance.

Following is the output I got.

enter image description here

After all the experimentations, I feel that frequency response is sluggish(Right image) due to RC limitations of the circuits and that too due to capacitance of photodiode.

Someone please suggest me something for the same. I have read a lot of documents related to the same but couldn't find anything that is concrete.

\$\endgroup\$
2
\$\begingroup\$

I have used bootstrapping on photodiode amplifiers like this: -

enter image description here

The idea behind it is that any AC the capacitance of the photodiode is largely muted by applying the same voltage at its cathode thus you can use a much smaller feedback capacitor Cf.

The problem comes down to reducing the noise gain of the op-amp when used as TIA (trans impedance amplifier) - because the JFET reduces the capacitance seen at the inverting input, high frequency noise (due to the op-amp) is not amplified as much therefore Cf can be lower and bandwidth can be significantly bigger.

I am then using an opamp (LM324)

The voltage noise of that op-amp is terrible for applications like this. Choose a device with much better performance and much improved lower bias currents.

\$\endgroup\$
  • \$\begingroup\$ I love the LM324. Mainly because I have 30-40 in the loft ( aka free and if you have to buy them they are so cheap) . They seem to be clumsy engineer proof (indefinite short circuit to ground). However they suck near the rails as well. \$\endgroup\$ – Spoon Dec 21 '16 at 12:27
  • \$\begingroup\$ @Spoon LOL I used to love LM324s back in the 80s when Spandau Ballet were strutting their stuff but I find them boring now. \$\endgroup\$ – Andy aka Dec 21 '16 at 18:02
  • \$\begingroup\$ ...it's all about context :-) when the boss is a cheap skate you use what you can beg or "borrow" ... until the day you realize that the dude in accounts can help you build a case for production ATE that saves you and the company money and time. Then the budget magically appears ... now you buy the right part for the job which is also way more fun... and buy the dude in accounts a drink. \$\endgroup\$ – Spoon Dec 22 '16 at 0:07
0
\$\begingroup\$

Are you certain that the problem isn't with your light source? LEDs aren't fast when driven by a 50ohm sig generator, and certainly not fast when compared to even a $1 laser-pointer. The onset of nonlinear laser action will greatly speed up the edges of your opto square wave. Cheap laser diode, or get yourself a $10 used fiber-optic Agilent transceiver module on eBay. Those have IR laser output at tens of MHz, up into GHz for the more expensive modules, with correspondingly sharp edges to their light signal.

I haven't messed with PDs recently, but I was always hoping to play with the circuit in Phil Hobbs' paper on Cascode bootstrapped front end for photodiode amplifiers. There's also the Bob Pease version. Note that these are late-1990s. Extremely fast op amps are now available.

And obviously search the datasheet schematics on goog using keywords: photodiode transimpedance MHz. Also try keywords: large area photodiode transimpedance.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.