# [Solved] Transimpedance Op-amp decreasing small signals amplifying larger ones?

Having a real issues with my transimpedance amplifier circuit. I'm running an LTC1050 op-amp and a BPW-34 (for testing) photodiode to try and amplify the output of the photodiode. I can draw a circuit diagram if that will help however, I'm not thinking the issue is that my circuit is wrong, more-so that I need to use a different kind of circuit such as a two-stage amplification to get what I want. What is happening is this:

Without any amplification the BPW-34 reads about 0.25V with ambient light. When I put the BPW-34 into the live op-amp circuit I get about 0.09V with ambient light. Without any amplification the BPW-34 reads about 0.6V when a flashlight is aimed at it's surface When I put the BPW-34 into the live op-amp circuit and shine the same light I get ~5.0V (with a 100kOhm load resistance)

So there is amplification with my circuit when an intense light is acting on the diode. Yet negative-amplification at ambient light? However, I really only want to amplify say the values from 0-0.250V to either 0-2.5V or 0-5V(for arduino analog). Interestingly, with an LM358N I did not observe this effect, and was observing a surprisingly decent amplification on the desired range with an awful amount of noise.

My circuit is that of a typical Transimpedance Op-Amp I have an RL of 10kOhm and a Rg of 2.2kOhm well under the spec for maximum gain in this type of set-up. Using a (5.12V) DC wall-wart supply that is decoupled for noise.

So I guess the question is, how can I amplify the lower end range rather then the higher end range of the photodiode? I've seen some people using MOhm resistance for RL but when I tried this I recieved no response from the circuit no matter what light was acting on the photodiode. Is it something specific with the LTC1050 (meaning do I need an op-amp that caters to the signal level I want)? If so how do I determine this?

I apologize for not really knowing much about electronics but any pointers, ideas, or key-words to research would be immensely appreciated. I didn't create a circuit diagram yet because it literally is a cook-book diagram for a photodiode transimpedance op-amp (assuming I didn't screw it up).

This is a hardware issue. Standard Operating Procedure for such issues is post a photo of the circuit and a photo of a hand drawn schematic. We don't like getting links to Fritzing's done by some expert who knew what they were doing. We want YOU to draw YOUR circuit so we KNOW that you were forced to look at it at least twice. Like it or not, that's what we want. After you post the photos we'll walk through the debug , one measurement at a time if necessary. We really need a schematic of your circuit and a photo to help you.

I’m not thinking the issue is that my circuit is wrong,

Who knows, engineers speak schematic.
However, it does sound like you have some sort of threshold going on with the circuit.
But we need to see the circuit.

However why use a transimpedance Op-amp at all?
I would have thought a normal op amp would be better suited.

Alright here’s my circuit schematic. I’m really sorry if the photodiode is drawn backwards. I have it such that when read with the multimeter the positive potential (pin) of the diode goes to the positive of the op-amp and similarly for the negative pin. I always get confused about which is the anode or cathode with photodiodes when they are used for recieving light verses emitting. Kind of dyslexic, won’t go into it. I have tried it the other way around at risk of the IC, and it read flat 0.00V with any light source I tried.

Sorry about not posting the schematic I’m just a little rusty with them and haven’t drawn many since physics 2(two years ago). Hopefully, I did this in an acceptable manner. The curved loop leading to the resistive load and filtering capacitor is not an inductor it is just to signify that that lead is not connected(jumps over) to the V+ ‘rail’.

Also I created a voltage divider with a ‘virtual ground’(1/2 VCC) but am not using it in the circuit. It was something I built for the single supply op-amp circuit (LM358N) and decided to keep. As it already has the decoupling capacitors in place, and bread boarding with jumper wires is tedious. I have tested the circuit without it and had no change in result, so I don’t believe the issue is there. Although, I do know that he type of voltage divider I am using can become unbalanced in amplification experiments.

As for transimpedance verses just an op-amp, I was reading a book about photosignal detection and all they seemed to talk about were transimpedance for photodiodes. So I assumed(perhaps incorrectly)that this would be the way to go. Any input would be really appreciated though. I’m really not seasoned with electrical engineering at all.

For clarity and not having to scroll up again the op-amp is a LTC1050, and the photodiode is a BPW-34.

LTC1050 Datasheet - http://cds.linear.com/docs/en/datasheet/1050fb.pdf
BPW34 Datasheet - http://www.vishay.com/docs/81521/bpw34.pdf

Edit -
Okay fritzing was too confusing for me to use. Also attached to this post is the bread board layout, cable for cable, hand-drawn. Editted again. Resized the images they were too big and combined posts.

Hi,
What are you trying to achieve? My guess is that you’re trying to increase the voltage range so that you’ll have more sensitivity on the ADC. Using an amplifier is not the best solution since any noise will be amplified too, unless you use a low-noise opamp.
Typically the light intensity changes very slowly (unless you put it under a strobe light :D)

See http://www.vishay.com/docs/84154/appnotesensors.pdf and http://www.fairchildsemi.com/an/AN/AN-3005.pdf for schematics and background information

If you just want to measure the light intensity for small period of the day e.g. transition from dusk till night , you don’t need an amplifier at all. You just need to tune the resistor in series with the phototransistor, so that you have maximum sensitivity for those light conditions.
For instance, I also use a phototransistor in one of my circuits, I have 1.5 MOhm resistor in series and that gives me a good resolution during dusk and dawn. During the day, it’s simply maxed out, during the night it’s zero.
I also noticed that light intensity follows a logarithmic scale, by taking the log10 of the analogread and multiplying it by 100 (digital equivalent of the opamp). I get an almost linear response.

The light source will actually be a laser pointer, I am testing with a flashlight and ambient lighting. Right now I am doing basic tests to get the circuit working with a cheap photo-diode before using a more expensive one in the actual circuit(which will require some tuning). I want to monitor it's intensity which does varry in my application. I was under the impression that photodiodes in general have quicker response times then phototransistors. Response time is also important for my application. Although I am sure a phototransistor could be used here, I believe that the problem I am running into is solvable and see no reason to turn back now. Thanks for sharing your results though, I may consider it if I get stuck here.

I am trying to get a better resolution in the 'low' end of my signal response from the photodiode to my operational amplifier circuit as stated in the original post. for output voltages(guesstimating) of about 0-0.25V my op-amp circuit gives readings of about 0-0.09V(decreased signal out-put after amplification?). While for readings of about 0.35-0.60V I get an amplified analog potential of about 1-5V? I am not sure what is wrong or how to improve my amplification in the lower end voltage range. As my application will only actually recieve 0-0.25V input voltages and I would like to see it amplified too 0-2.5V or 0-5.0V for the arduino analog in. Hopefully that clarifies what I am asking.

Read the notes starting from page 6, it gives you a lot of background information. I was wrong in my assumption that photodiodes and phototransistors work alike.

APPLICATION SCHEMATICS FOR PHOTODIODE-BASED AMBIENT LIGHT SENSORS
A sensitive, FET-input type of photodiode preamplifier, with a very low input noise, should be used, such as:
• Linear Technology LTC6240
• Texas Instruments (OPA827)
• Burr-Brown (OPA128)

Nice bit of information thanks. Again this isn't for ambient light. There will be a light source directly acting on the photodiode. Is there something in the application notes that I am missing that's pertaining to my problem? I see all of the example circuits are using large load resistances (10^9 - 10^12 ohm) that 'approach unity'.

On a side note, I am also noticing that the circuit application for the photodiode amplifier in the LTC1050 data sheet uses 500kOhm of resistance for both resistors. So I tried a combination of a few of the bigger resistors I have available on both of the spots where resistors are in my original circuit. I found that 540Kohm on both locations my so called RL and Rg helped the issue a bit. Well now when I read 0.210V off of the photodiode without any source it appears as 0.260V when the op-amp circuit is running. The lower signals are amplified as well or so it seems.

So maybe I just need to boost both the resistors on both locations? It seems like if I do an off-set such as 540kOhm on RL and say 50kOhm on RG I get a decrease in voltage increase. Probably need to go out and buy some bigger resistors? I had no idea the resistances would need to be this high. the circuit I built with the LM385N only had I believe 100kOhm and I believe a 10kOhm. Worried about how the noise is going to look because of all of this load?

Would really appreciate it if someone checked my circuit.

I have the impression that the circuit depicted with the photodiode in the app notes is for high frequency applications, like fibre optics and such. You stated in your post you need this quick responsiveness, but how quick is quick?

If you're talking about milliseconds I wouldn't go with this type of circuitry and a photodiode. The people who wrote the app notes are experts and those transimpendance amplifiers look quite nifty. The components need to be picked carefully (low-noise, FET inputs opamps), proper shielding needs to be foreseen, PCB layout,... All come into play to get a stable circuit.

If you have a look in the same document at the circuits depicted with a phototransistor, they are much simpler. Basically it's a voltage divider you directly connect to the ADC. You can put an amplifier in between, but it's just a regular non-inverting opamp.

In regard to the difference in signal amplification. This is taken from wikipedia, it's the current characteristic that is linear and not the voltage.

In electronics, a transimpedance amplifier, (TIA) is a current to voltage converter, most often implemented using an operational amplifier. The TIA can be used to amplify the current output of Geiger–Müller tubes, photo multiplier tubes, accelerometers, photo detectors and other types of sensors. Current to voltage converters are used with sensors that have a current response that is more linear

I don't see why you're so against the op-amp set up for photodiodes. People use them for really high-speed circuits all the time. Photomultipliers are crazy when it comes to low signal to high signal leaps(literally on the single photon/electron level), granted they traditionally also run off of high voltage sources... Yes op-amps and TIA circuits do have to be tuned very well to perform as desired. I am not an expert. I also understand fine tuning op-amps can be deep water for non-experts. At the same time, I want to learn about them because I frankly think they are genious. Yes the current for a PIN photodiode is the linear characteristic verses irradiance as seen in almost all of the data sheets. I also figured that was the point of using a transimpedance op-amp? Maybe I am wrong, or missing the point again. I can put it on PCB, and shield it but I do not believe that I am getting the behaviour I want from the circuit yet, so I don't want to 'finalize' anything until I do. Meaning I have no problem with instability for now so long as I am getting a 'good enough' gain in the region I want.

Checking the datasheet for the op-amp that I have selected, it has a slew rate of 4V/uS(50pF capacitor and 10kOhm load resistor present). That's plenty fast not taking into account the capacitance of the photodiode. 1-3ms if I hit the 'rails' which I don't really plan too and would prefer not too. The capacitance of the photodiode I will be using eventually is 400pF at 5V reverse. Also it has a sensitivity of 9nA/lux. Which should be way lower then I need. I should also mention that my changes in light intensity or irradiance are pretty gradual. Meaning it's not an "on/off" type jump, more like 0.110V(t=1ms) to 0.113V(t=2ms), then to 0.117V(t=3ms), etc. ?s there something that I am missing here, or is the LTC1050 not suitable for this application? Like I said I had pretty reasonable results with an LM358N(which to my understanding is essentially junk for these types of applications), but the over-all noise in my circuit was awful. Not noise limitted too the op-amp itself, to everything else going on(back EMF from a 12V 1A motor, PWM signals on a common ground, etc).

Anyways. I still haven't been really clearly suggested any way to amplify the 0-0.250V signal I am getting to say 0-2.5V. Is there a key-word I am missing? Should I do a dual stage amplification? Is there a way I can clip the signal above roughly 0.250V and scale that input into an op-amp(would even that help)?

So I worked out a way to do what I wanted.

What I did was, used a transimpedance stage for current -> voltage. Then a non-inverting voltage amplifier second stage. Works pretty well, still some noise. Figured I would atleast share how I worked out the problem for any future readers.

Why use a transimpedance op-amp at all ?

Well the relevant output of the photodiode is a current proportional to the light level, but to eventually read this as an analog input, you need to create a voltage proportional to that current.