How can I design a Photodiode Amplifier Circuit optimized for 470nm wavelength?

Hi David, What is the application of this photodiode? Do you have a bandwidth requirement? What range of photocurrents do you expect to measure (i.e. 1n -1uA)? Are you trying to make your design run off of a single supply? Do you have an A/D picked out and are you going to be performing any signal processing? These questions should provide a starting point for us to be able to recommend a photodiode amplifier for your application. Matt

Dear Matt,

I have a similar issue, i need to design a photodiode amplfier with low bandwidth requirement ~ 100k  samples per second. My photocurrents are of the order of several tens of pico amps. I intend to run it off a single supply (perhaps using button cells or lithium batteries). Right now, we do not have an A/D picked up and are looking for a 16 bit A/D. The problem i am facing is about noise. How do i design a circuit that does not pick up ambient electromagnetic interference? Are there eval boards available from TI to do the same? Please let me know; it would be good help

Varun

Varun,

Creating EMI/RFI immunity in analog circuitry is always very challenging.  Aside from good layout and shielding, sometimes your only recourse is to filter the ambient noise with a HP filter.  Depending on your application, another thing that you can do is to sense differentially across the photodiode as this makes helps reject common mode noise.  This means, however, that you will put a zero bias across the photodiode which will inherently increase your junction capacitance and may inhibit your rise time such that this will not work, in which case you will have to apply a reverse bias to the photodiode.

With that said, measuring 10pA at 100ks/sec on a single supply is not a trivial task.  If your photocurrents are pulses this makes it even more challenging especially if you are targeting high resolution.  What is your application?  Have you picked a photodiode yet?  Are you open to creating a -supply voltage?

Matt

Matt,

 

Thank you for the information on designing.

We are trying to measure light levels in different spectra using photodiode, for an application similar to spectrum analysis. We intend to use the photodiode response and we have designed an amplifier board with ADCs, however, the noise being picked up is severe and we are not able to use the system. Currently the photodiode being used is a Hamamatsu one. Does TI have any modules such as these that we can purchase initially to try out?

 

Varun

Varun,

The only device that TI offers that integrates the photodiode is the OPT101; although, I don't know if this will resolve your noise issue.  Noise in a photodiode circuit is a function of the diode shot noise, johnson noise, OPA voltage and current noise, and the thermal noise of the feedback resistor.  Ambient noise, however, is a function of layout, shielding, and environment.  What level of noise are you measuring and at what current level?

Matt

Matt,

I am currently checking with OPT 101, to receive an SNR of close to 70db. The hamamatsu diode gives a current of less than a nano-amp. We are looking at close to 100db; is there any quick fix way to try out to acheive said perfomance or some material for ready reference to get started with. Are there any TI amplifiers you suggest?

Varun,

There really is no "quick fix" for this type of performance requirement .  One good amplifier to consider for your application would be the OPA380; however, achieving an SNR of 100dB with a signal of < 1nA will be very difficult if you are pulsing your photodiode at 100kHz as you are up against small signal bandwidth, OPA voltage and current noise, the thermal noise of your feedback resistor, and settling time and accuracy.  Ultimately you will need a good notch filter to reduce the total integrated noise at your bandwidth of interest.  In fact, you may have to go to a higher speed amplifier to achieve the bandwidth and performance you are trying to achieve.  Check out the following applications notes that discuss the tradeoffs of different high speed transimpedance applications: 

http://focus.ti.com/general/docs/litabsmultiplefilelist.tsp?literatureNumber=sboa117 

http://focus.ti.com/lit/an/sboa035/sboa035.pdf. 

You might also check the following thread for some additional inputs from our higher speed applications team: http://e2e.ti.com/forums/t/9230.aspx.

Matt

Hi Matt,

I would like to design an transimpedance amplifier with my target bandwidth is 0-20MHz. Is that OPA846 is a good choice?

Yong,

What amplifier you pick requires many other factors aside from bandwidth.  What levels of current are you trying to measure?  What level of SNR do you want to target?  Is your application pulsed? 

Hello,

I should build an amplifier front-end to detect a small current from a photo detector.

In my case i have chosen a pin photo diode. This diode delivers a current of approximately 1nA (peak) - with a frequency of about 100MHz to 1 GHz. So i need a pretty fast pre-amplifier stage.

I decided to choose a current feedback amplifier in a trans-impedance configuration. Because of my little signal i think i need a CFA with very low Input Bias Current. Can you please recommend me one or give me an idea for other possibilities to solve my problem.

Thank you in advance.

Andreas

PS.: After pre-amplification and further amplification i like to hand the signal to an ADC with differential input.

I would suugest using may be OPA111. By the way, I am working on similar idea on medical project. Where do you take care of your noise? I mean do you use a LPF on the output of your amplifier?

 

Hello,

sorry, my answer comes a little bit late. Your OPA111 is not matching my boundary conditions, but thank you anyhow. About noise I haven´t thought about yet. First I need to find an OPAmp (I think THS3202 is the right one) and then I can think about SNR stuff. In my case - I have a pretty small input signal - it will be quite difficult.

Regards

Andreas

Take a look at this presentation:

4370.PhotodiodeAmplifers.pdf

hi dear

my sugesstion  go through  OPA656/ OPA657/ OPA695

select any one   according to  your  optical pulse Rise time   & Bandwidth

Slew rate & settling time

good result with PIN
 photodiode  in Transimpedance mode

 

I have done several low noise, wide dynamic range designs with photodiodes and sometimes the only solution to external noise is shielding.  You need to seperate the source of the "noise" you are seeing.  There is optical "noise" or interference (that can be due to ambient light, reflections) and electromagnetic "noise" or interference.  The best solution for optical noise is optical filtering (using coated optics, dyed plastics etc, common in IR applications) and optical aperaturing that can also limit the amount of stray light, coating any cavity surfaces with flat black paint etc can also limit reflections.  If your light originates from a controlled source, like in optical proximity sensors or in instruments that measure opacity or transmission you can modulate the source and then use a bandpass filter in your reciever design.  If the dominant noise source is electrical fields (common mode pickup typical in high impedance circuits) then transparent optical conductive (indium tine oxide) coatings thin film metallic coatings or wire mesh can give electrical field shielding and only minor optical degradation.

Also, the amount of photons captured by the photodiode is a squared relationship to the diameter of the photodiode active area, use as large a photodiode as you can but then be careful that the capacitance doesnt become too large to limit the frequency reponse in pulsed applications.  

Varun,  We have a similar problem.  Which Hammamatsu photodiode did you pick?

 

We're looking at 405nm about, and are trying to get a stable measurement of 1 part in 10^6 at low light levels....  I'd love any feedback you may have.  glennfish at gmail.com

 Andreas;

 

Your application requires very high bandwidth (100MHz to 1GHz) and high sensitivity at the same time-- this puts it out of range of a PIN photodiode. To achieve what you want, requires a detector that provides internal gain, such as an avalanche photodiode (APD) or even a photomultiplier tube (PMT) . An APD can provide gain of about 100 while a PMT provides about 1,000,000. Both are high BW devices with low noise.

The disadvantage of both detectors is that they require a high voltage power supply. Fortunately, not much current is required. An APD will require 80 to 250V depending on its type and a PMT will require 1-2kV. These voltages can be provided by tiny modular supplies made by companies such as Ultravolt, EMCO, and Pico Electronics.

Regards,

Neil Albaugh   (formerly Burr- Brown & TI)

Diamond Bell Technology

Grant;

Your recommendations are all excellent but one-- that one is a trap that almost everyone falls into when first designing photodiode amplifiers.

:...use as large a photodiode as you can but then be careful that the capacitance doesnt become too large to limit the frequency...."

That qualifier at the end is all-important. Generally, one needs to select a photodiode with as low capacitance as possible, given the active area requirements of the optics. This means selecting a small PIN photodiode instead of a much higher capacitance diffused junction photodiode. Also a high-resistivity silicon PIN diode is preferable but usually its resistivity is not shown on a photodiode data sheet, only a choice of "photovoltaic" or "photoconductive" photodiodes. A photovoltaic photodiode is fabricated on low resistivity silicon and has better linearity at high output current. A photoconductive photodiode is processed on higher resistivity silicon and is meant to be operated with reverse bias. These types are faster and have lower capacitance but, in reality, they also work just fine in a zero-bias mode but they are slower and not as linear at high current output. Also, in zero-bias mode their dark current is extremely low.

The critical parameter for photodiode transimpedance amplifiers is input capacitance. This is the sum of the photodiode shunt capacitance and the op amp input capacitance. This capacitance together with the feedback resistor causes the amplifier's noise gain to increase with frequency, so noise becomes more of a problem with capacitance and this can also cause enough gain peaking to cause oscillation.

Another thing regarding feedback resistors-- make the resistor as high as possible consistent with your BW (as you pointed out) and input current requirements. It is not intuitively obvious but the noise current of a resistor decreases as its resistance increases-- just the opposite of voltage noise. Remember that the signal gain of a transimpedance amplifier increases directly as its feedback resistor increases but its noise current increases only as the square root. Therefore, SNR is improved by the square root of the feedback resistor increase. 

Well, I've pontificated here far too long; just open some of the Example Circuits in TINA-TI to see a few photodiode circuits.

Regards,

Neil Albaugh   (formerly Burr- Brown & TI)

Diamond Bell Technology

David;

The best photodiode for your application depends on your signal level and bandwidth requirements as well as on the wavelength of your optical signal. High speed photodiodes are generally optimized for longer wavelengths while general purpose photodiodes peak at a shorter wavelength, even "blue-enhanced" and UV-enhanced photodiodes are available. Look at your overall system requirements before choosing a photodiode or op amp. As a rule of thumb, an OPA827 is great for high sensitivity low bandwidth applications while an OPA656 might be optimum for faster applications.

Hamamatsu makes a broad line of nice photodetectors but they are not the only vendors-- try OSI, etc. See my other posts on photodiodes and transimpedance amplifiers for more information.

Regards,

Neil Albaugh   (formerly Burr-Brown & TI)

Diamond Bell Technology

hiii..

i am using opt101 photodiode amplifier for absorbance. i am getting fluctuatuions in output. Should i use filter. if yes thn wht filter shud i used.

or wht can i do to get a constant output.

Tia,

This is an unrelated topic. Please make a separate posting. To provide any meaningful reply, we would need more detail on your application:

Provide a circuit diagram or refer to one in the product data sheet.

Provide information on power supply voltage and feedback resistor value.

Provide information on the DC output voltage and AC variation in output voltage (amplitude and frequency content).

Provide information on the light source. Do you have assurance that the light source is constant?

Regards, Bruce