Analog input question?

Status
Not open for further replies.

mfw

Member
Amplifier.jpg

Hello, I am using a teensy 3.5 analog pins (ADC) to read signals coming from my amplifying circuit. This circuit is designed to amplify very small currents from a photodiode. After testing the alpha prototype (figure attached), i have concluded that this circuit does not do a good job of differentiating small differences in current from the photodiode, meaning it is not accurate enough. To fix this I want to increase the supply to the rails of the op amp, in order to increase the range of amplification (Better gain). But i understand I am limited to what I can read on the ADC (3.3 V). How much can I increase the range on the ADC?

note: This is all on a breadboard, which causes noise
 

Attachments

  • Code.pdf
    69.5 KB · Views: 148
Increasing the voltage to the op amp will not increase its gain. You may want to add another op amp stage to get more gain. What are your readings now? I am wondering if your output voltage is in an active region of the op amp or stuck at one of the supply rails, 0.0 or 3.3.
 
Increasing the voltage to the op amp will not increase its gain. You may want to add another op amp stage to get more gain. What are your readings now? I am wondering if your output voltage is in an active region of the op amp or stuck at one of the supply rails, 0.0 or 3.3.

Hi rcarr,

We are using the photodiode to capture fluorescent light. We first excite the sample with an LED, then capture the data. This fluorescent light results in micro amp or smaller, currents, which is why we need big gain. Problem is ambient light is adding to the input, and slamming the output to the 3.3 v rail. In a dark container (no ambient light), we measure 73 (raw ADC value 0-1023) with no sample present, and 74 (raw ADC value) with the sample present. So in the dark we would hope to see a bigger differentiation than 1, between sample and no sample.
 
Depending on what you are doing with the data and the fade rate of the light but in addition to tweaking gain would it work to run illumination at a rate and then high pass filter the result to strip the DC component. Can see that supressing the direct illumination signal might be tricky. Another option might involve having a bias offset first stage, where you use the Teensy DAC to adjust the DC level to the input op amp to give you a high gain centered about 0.2V (or whatever it happens to be on the day) rather than 1.65V.

While chasing low level analog signals, are you already using the external Aref over the default internal one?
 
I read somewhere that the Teensy 3.x has a 16 bit A/D with 13 useful bits. And there is a command AnalogReadRes(n) that will set the desired number of bits. Using 12 bits would give you 4 times the resolution you currently have.

Another idea: could you arrange a 2nd photo diode that sees ambient light but does not see the florescent light? If so, you could wire the two diodes in a common mode rejection circuit and then the output would be only the signal you wish to detect.
 
GremlinWrangler,

We are using the default reference. I believe what you are explaining is Modulation, where you pulse the LED at a high frequency, then sample at that rate, and hopefully the ambient light (very low frequency) will not be picked up at all. I want to test this design, but first I need to understand how to write code to pulse the LED at a selected frequency.

Thanks
 
I read somewhere that the Teensy 3.x has a 16 bit A/D with 13 useful bits. And there is a command AnalogReadRes(n) that will set the desired number of bits. Using 12 bits would give you 4 times the resolution you currently have.

Another idea: could you arrange a 2nd photo diode that sees ambient light but does not see the florescent light? If so, you could wire the two diodes in a common mode rejection circuit and then the output would be only the signal you wish to detect.

Interesting,I will look into that command.

Thanks
 
Status
Not open for further replies.
Back
Top