Transimpedance Op-amp, little gain for small signals, large gain for large signals

Discussion in 'The Projects Forum' started by AFool, May 21, 2014.

  1. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    Hello, first of all I am brand new to the forum. I did ask this question on another forum and was not given any advice, useful information, or really even a relevant avenue to look down. Hoping this place will be a bigger help. I'll start off by saying I am pretty new to electronics but am not asking to be spoon-fed answers.

    Essentially here is what I have going on. I am shining a laser pointer on a photodiode(BPW21). By checking the voltage coming off of the diode I get a nice 0.00V(no light) and with the light(0.250V). I will spare you the detail but the intensity of the light acting on the photodiode does change for my application. Meaning I can get 'intensity'/'irradiance' measures of 0-0.250V.

    Now, the goal is to turn this range into a larger but still accurate analog signal so it can be read by a microcontroller. Ideally I would like to see 0.0-5.0V but at this point will happily settle for 0.0-2.5V. Now I am aware, it is the current of photodiodes which has a linear response to light intensity not the voltage. So I selected a transimpedance operational amplifier to achieve this goal(current->voltage+gain).

    I have attached an image of the circuit diagram I am currently using(hopefully the photo diode isn't backwards in the diagram it's not in the actual circuit :(). I will create a diagram of what is exactly on the bread board by request.

    The problem. Without the transimpedance op-amp I get potentials of 0-0.250V. With amplification I get 0-0.330V even after a megaohm load resistor. Although, for the range of around 0.400-0.500V(not amplified) when amplified I see 4.00-5.00V after amplification. In my application I will only be seeing the 0-0.250V (as well as the associated current response), so having the gain so large after the 0.400V is not useful to me. I think there is something wrong with my circuit, but am not sure what. Any input, advice, things to look into, etc, would be greatly appreciated. IE should I be doing two stage amplification, is something blatantly wrong with my circuit, should I be doing more of an instrumentation amplifier circuit, should I use a bigger supply voltage on the op-amp(the circuit will be shielded from light when operated), etc. Noise tuning, all the finalized stuff I can handle on my own, right now though, getting an appropriate/useful gain is what I am stuck on.

    Interestingly I was obtaining more useable results with a junky LM358N op-amp on a single supply and the same photodiode. I only bought the LTC1050's because they have significantly better specifications, but I must be doing something wrong here.

    Here are the data sheets for the components I am using:
    LTC1050 Datasheet - http://cds.linear.com/docs/en/datasheet/1050fb.pdf
    BPW21 Datasheet - http://www.vishay.com/docs/81519/bpw21r.pdf

    Sometimes I test the circuit with a BPW34 but it's not the ultimate photodiode that I want - http://www.vishay.com/docs/81521/bpw34.pdf
     
    Last edited: May 21, 2014
  2. crutschow

    Expert

    Mar 14, 2008
    12,977
    3,221
    I assume that the + and - terminals are shown incorrectly since the schematic has positive feedback.
     
    AFool likes this.
  3. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    Dang yes you are right. Circuit diagram is now drawn how it is operating.
     
    Last edited: May 21, 2014
  4. alfacliff

    Well-Known Member

    Dec 13, 2013
    2,449
    428
    with negative feedback and identicalo resistors, the gain is unity. you shouldnt expect a lot of gain that way.
     
    AFool likes this.
  5. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    I have tried changing the resistor which goes to ground(-5V) to lower values and did not see any improvement. Which I thought was very strange given all the reading I have done and how my circuit with the LM358N worked. Actually the out-put signal decreases?

    For example I just swapped that resistor(Not the load resistance) to a 10kOhm (from 540kOhm) and am getting a signal of 0.330V. Now with 540kOhm I get 0.56V. Unamplified the signal voltage is about 0.25V. This is using the BPW34 not the BPW21.
     
    Last edited: May 21, 2014
  6. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    AFool likes this.
  7. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    WMODavis I have visited this web-page before and it was how my original circuit was constructed.

    The first circuit in the first link:

    The only difference between their circuit and mine, is that they do not have a resistor going to ground and instead they shorted it to ground with a jumper cable. Unless I somehow mixed up the order of my wires?

    Anyways when I remove that resistor and short it as is done on their web-page, I get a decreased signal. Meaning what would be 0.250V without the op-amp I get say 0.100V. For the higher signals there is still amplification. Granted I am amplifying current, not voltage with the transimpedance. It's substantially worse. Unless I am supposed to use a megaohm load resistance? The only issue is when I have tried a 5 megaohm load resistor(perhaps too high of a wattage rating) I was getting no signal at all on VOut.

    The circuit in the second link:
    It looks as though they have temperature compensation and a 5V voltage regulator. Not sure what the difference is other then they are using megaohm resistances in that circuit as well. Could the issue be that I simply do not have enough resistance in the load resistor to amplify this region of current?

    I will try and scoop up some megaohm resistors by tommorrow of a few varieties and see if this solves the problem.
     
  8. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    Update - Just tried 3.7MOhm of resistance and shorting the V- pin to the - Input. Now the 0.250V signal is resting at 0.750V. This is almost useable! So I guess the question is do I continue increasing the load resistance or should I be not shorting the V- to the - Input and instead be putting a balanced resistor load there as well? Should I be thinking two stages or will noise *probably* not be too bad doing this all single stage??
     
  9. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    Basically the transimpedance amplifier converts the photodiode current into a voltage at the amp output. They mention in first link "With the small solar cell shown, this circuit produces almost 6V in hazy, partly cloudy summer sunlight." I know you are not using same device but with a 9V supply they are getting current to voltage conversion and he mentions changing the feedback resistor from 10K to perhaps 10MOhm when using a LED which needs more gain to get more output voltage.

    Not sure if that is what you are calling the load resistor but in terms of OP Amp topology I'm use to it is called a feedback resistor and used to define the gain.

    I have a piece of Dr Brooks equipment. Not the photometer however. I have the pyrometer which is just a photodiode with a load resistor. The output voltage is calibrated so I read solar radiation (W/m2) and have it connected to a data logger.

    Is your application for the purpose of detecting the laser pointer output only?
     
  10. atferrari

    AAC Fanatic!

    Jan 6, 2004
    2,644
    759
    Please post the actual/current circuit.
     
  11. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    The BPW21 seems to peak at ~550nm whereas I think a red laser is around 650nm so the response might suffer partly for that reason.
    The BPW34 peaks @ almost 950nm.
    That's why I asked "Is your application for the purpose of detecting the laser pointer output only?" or just testing at this point. In the long run may be best to match light source and photodiode spectral response for best output.
     
  12. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    @WMODavis- Yes the circuit is designed to only detect the laser pointer light. I believe the laser pointer is a 5mW laser pointer but it was cheap so it's impossible to say without calorimetry or another form of measurement. That is a good suggestion though. I also know that the BPW-21 has a green colored 'lens' which may be removed (very carefully) to improve the response if necessary. For now I want to keep it in tact and use this photodiode so I can understand the basics then perhaps work from there. The laser pointer may be removed in the future and another color used instead.

    @atferrari - Attached is the current circuit diagram. On the supply line there is also a 100uF and a 0.1uF capacitor in parallel for decoupling purposes.

    I did get my first results with the new op-amp. It doesn't appear noisey. It does appear to be oscillating wildly(AC signal coming out?). Is the AC signal likely a correct assumption? I posted results of the output, keep in mind it should resemble one period of a sine wave. Instead it's almost like there are two sine waves(of different amplitudes) and my signal of interest is in the middle? I tried fiddling with the 'filtering' capacitor and it did change the pattern. I'll have to look into ways to really clean this up once the amplification is completely worked out. Noise is a big reason why I was afraid to use such high (RL resistors) or feedback resistors. My terminology is infantile, my apologies.
     
    Last edited: May 21, 2014
  13. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    Guys, the diode is in backwards. Check the schematics at the bottom of the first link page in post #6.

    To confirm, see the schematic on page 13 of the LTC1050 datasheet.

    Check that, the new sch went up while I was typing.

    ak
     
    AFool likes this.
  14. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    Well before I say that the issue could be the photodiodes sensitivity I should also mention that the LM358N amplification circuit was hitting nearly 1.0V. Ridiculously noisy however. Attached is an image of the results from that experiment(250 intensity counts ~= 1V). Note that those results are two periods of the 'sine wave'.


    Did I draw it in backwards again? I'll try and flip the diode to be double sure it isn't actually in backwards. I keep putting them in backwards in the schematics. Edit - checked with the multimeter the photodiode is in the correct way in the circuit I will flip it in the diagram. Diagram should be updated and correct now. I'm really sorry I always get diodes backwards for some reason :(
     
    Last edited: May 21, 2014
  15. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    Another thought is that based on Brooks use of that opamp is to use it with a single supply - actually a 9 V battery. The output does not need to go negative it seems to me. Just a power supply simplification that might be worth a try not that it would fix the real issue you're trying to overcome.

    I've seen these cheap laser diodes and have thought they'd be fun to mess with sometime but am off on other interesting stuff now. Well hope you find the right way to do what you're trying.
    http://www.ebay.com/itm/10pcs-mini-...357?pt=LH_DefaultDomain_0&hash=item3a8b9c67b5
     
  16. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    I actually was originally trying to run the op-amp as a single supply. I didn't have enough load resistance at the time, so it may be worth revisiting.
     
  17. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    AFool likes this.
  18. DickCappels

    Moderator

    Aug 21, 2008
    2,647
    632
    For the oscillation, you can increase your feedback capacitor.

    Also, notice that iwth the very high transimpedance even a little bit current capacitively coupled from the power lines through the air to your photo diode can cause the output to thrash about wildly. You are entering the region in which you must either roll off the frequency response of your amplifier so the transimpeance for the interfering signal is low, or start adding shielding.

    For further reference:
    Notice the solution to hum pickup here was putting a 22 uf capacitor across the photodiode.
    [​IMG]
    http://www.cappels.org/dproj/photoamp/photoamp.html
     
    AFool likes this.
  19. AFool

    Thread Starter New Member

    May 21, 2014
    11
    0
    Thanks DickCappels this definitely has me thinking in another direction. I'm realizing it's essentially impossible to calculate the filtering/feed-back capacitor. So I have been experimenting a bit. Quick question though, sorry if this is really elementary, in your photodiode circuit, what is the dashed circle representing?

    I found removing the 0.1uF parallel capacitor from the feed-back/load resistor and putting it in parrallel with the diode prongs gave me a beautiful collection of signals(but clipped)! Only I lost amplification in a drastic way(by about 1/4)(Left most picture). I'm assuming the clipping is from the excessive capacitance(?). Found that a 270 pF capacitor in parallel with the diode didn't appear to decrease the oscillation at all and lost amplification again(by about 1/2) (right most picture). So what I am looking for is something inbetween 0.1uF and 270pF hah... I'm thinking I need to get this thing off of a bread board, but with the low amplification I'm not sure if that's too quick of a judgement call. Do you have any tips for this sort of 'balancing act'.

    I think in the end I might need a small capacitor such in the 100-1000's of pF's parallel with the photodiode, and a larger cap parallel with the feed-back resistor. So I don't lost too much signal but am not getting so much oscillation overloading the bandwidth of the op-amp, but am still smoothing the feed-back resistor noise. Not sure if this is true, I'm mainly going off of intuition for now.

    Also, I'm not sure what the small value resistor(330Ω) is doing in series with the feed-back loop? Is it for heat dissipation or to compensate for the capacitance which is running to ground?

    WMODavis - Thank you. That circuit looks very nice. Definitely a bit eye opening.
     
    Last edited: May 22, 2014
  20. wmodavis

    Well-Known Member

    Oct 23, 2010
    737
    150
    Of course Dick can speak for himself and correct me if I'm wrong.

    1. the "dashed circle" indicates that a shielded cable is used and that the shield encloses the sensitive signal wire from the sensor. Shield is connected to circuit common.
    2. The cap across the feedback resistor is there to increase negative feedback at higher frequencies where oscillation occurs.
    3. The 3.3k resistor in series with the feedback resistor just sets a minimum feedback resistance and thereby a minimum amplifier gain when the pot is turned all the way in one direction.
    4. I had found the circuits in that OPT301 datasheet helpful when I was doing something similar.
    5. Keep on it!
     
Loading...