ADPD103 - crosstalk between photodiodes

Hi community,

I have set up a ADPD103 Evaluation Board with two photodiodes (PD1 and PD2) and one LED. I want to measure the intensity of the LED with both photodiodes. When i bring the LED close to PD1, the signal rises to high LSBs, but at the same time the signal of PD2 decreases for a certain amount.
But when I disconnect PD2 from the Evaluation board, and bring the LED close to PD1, the signal of PD2 rises for a fraction of the signal of PD1. 

This crosstalk gives me wrong results in my measurements. Is there a meaningfull reason for that, and does anyone have a solution to get rid of it?

  • 0
    •  Analog Employees 
    on May 22, 2017 10:04 AM

    Hello,

    If you are trying to measure both photodiodes in the same timeslot then it is possible that you are getting some feed through via the common cathode connection. I have been able to minimize this in the past by placing a capacitor from the PDC pin to ground. You can try 0.01uF, and if that doesn't help enough you could try up to 0.1uF. You can also set the reverse bias on the photodiodes to ~250mV by setting register 0x54 = 0x0AA0. This could help by reducing the capacitance of the PD.

    Regards,

    Kevin

  • Hi Kevin,

    Thank you, that helped me a lot. I only had two capacitors here, 0.01uF and 1uF. The one with 0.01uF did not show any effect, the other one canceled the feed through completely. I am not sure now if there is any side effect when using a capacitor with a capacitance that is much higher than neccessary?

    Regards,

    Josef

  • 0
    •  Analog Employees 
    on May 23, 2017 12:30 PM

    Hello Josef,

    The only real issue is with the settling time of the PDC node. Once the sampling period is over, the device goes to sleep until the next sampling period. At this point, the PDC pin just goes to VDD. When the amplifiers and reference voltages get powered up during wakeup time. The PDC pin is then driven to the programmed voltage (in your case this should be set to the PD anode voltage plus 250mV by setting register 0x54=0x0AA0. Please make sure you do this). By hanging a big capacitor on that node it will increase the settling time of that voltage. It's probably not going to be an issue, we've had other users use this fix to address common cathode crosstalk. However, I think you will get the same benefit from a 0.1uF as you do with a 1.0uF based on data I've seen in the past. The 0.1uF will obviously allow the node to settle much faster.

    Regards,

    Kevin