Hi all,

I will preface that I am not incredibly experienced in electronics and even less so when it comes to RF interference and noise shielding/filtering so any tips or knowledge is greatly appreciated!

I am trying to design a photodiode circuit that will serve to detect the irradiance of a single uLED light source and provide a fairly high resolution measurement of said light source. This is being done mainly for QC and testing purposes, therefore the measurement needs to be accurate and consistent. The device will also be used in a high power HF field operating at 13.56 MHz which will certainly cause interference and noise that will need to be minimized through filtering.

My vision for the design will feature a detection stage in which the photodiode will output a signal proportional to the detected irradiance. That signal will be amplified (most likely with an op-amp) and then interpreted by an MCU (I'm thinking an Arduino but nothing is set in stone). I was thinking of having low pass filters in between the detection and amplification stage, as well as just before the signal input into the MCU to ensure the DC bias provided by the photodiode is robust and meaningful in terms of the small differences in uLED irradiance I need to detect. I am a little unsure as to the design of these filters and the nuances involved, or even the effectiveness of these filters against open-field radiated interference.

Any help would be appreciated and if one of my assumptions is blatantly incorrect, please correct me. I'm just trying to learn all that I can and see if something like this is even feasible!

Thanks,
Tom

 

Does your uLED source have a signal on it? If so, what frequencies do you expect to see on the signal? What sample rate do you expect to use to digitize the signal? Depending on the sample rate and,of course, the signal itself, the 13.56MHz may not have a significant impact on your digitized signal.

Photodiodes detectors output very small currents. A typical transimpedance amplifier will use a very high a very high impedance feedback resistor, typically 10M ohms or greater. This results in the input node of the transimpedance amp being very susceptible to electromagnetic interference. You will want to put the op amp for your transimpedance amp very close to the detector and to keep traces between the detector, the op amp and the feedback resistor very short. The output of the transimpedance amplifier will be much lower impedance allowing you more flexibility regarding part placement and trace length.

 

Does your uLED source have a signal on it? If so, what frequencies do you expect to see on the signal? What sample rate do you expect to use to digitize the signal? Depending on the sample rate and,of course, the signal itself, the 13.56MHz may not have a significant impact on your digitized signal.

Photodiodes detectors output very small currents. A typical transimpedance amplifier will use a very high a very high impedance feedback resistor, typically 10M ohms or greater. This results in the input node of the transimpedance amp being very susceptible to electromagnetic interference. You will want to put the op amp for your transimpedance amp very close to the detector and to keep traces between the detector, the op amp and the feedback resistor very short. The output of the transimpedance amplifier will be much lower impedance allowing you more flexibility regarding part placement and trace length.

Thank you for your response! That is a very good point about the input impedance of the amplifier stage and the trace/component placement, however I do have some questions: Could I get by this EM susceptibility by just using a lower impedance resistor? (If yes, is there benefits of using such high impedance resistors vs. smaller ones?) Is the high impedance resistor due to the fact that the gain needed to amplify the PD output will be very large as well so it is unavoidable?

To hopefully answer your question about the uLED source signal, the uLED is fed by a regulated constant DC voltage as part of a separate system that is powered wirelessly by the 13.56 MHz (which is the reason for the heavy RF presence). I'm not exactly sure about the sample rate, however if the purpose would be solely to measure the irradiance of a constant light source, would the digitization sample rate play a large factor in this? Forgive my ignorance, and again, I appreciate you taking the time to respond!

 

...gain needed to amplify the PD output will be very large as well so it is unavoidable?

That's the key. The PD output current is typically very small. That is the reason for the high value feedback resistor. As I noted, placing the op amp near the detector and keeping traces on the input nodes very short will help to mitigate EM interference issues.

Since you are measuring a nearly constant output, your sample rate can be fairly low. Alternatively, you can average samples over time with a faster sample rate to filter out any high frequency content. For this application, I don't think you'll need much in the way of analog filtering to get good results.