Post Go back to editing

LTC3109 shunt harvesting

Dear all,

I made a pcb design to harvest energy from a voltage drop (>20 mV) that occurs in a trace i have in my PCB. This trace carries 2 A and runs on the bottom layer.

However, it does not work. Has enyone tried this already? Could it be that the LTC3109 is afected by the magnetic field generated by the high current underneath? I didn't find any information about that in tha datasheet ... Thanks!

  • Hi,

    Since the startup input voltage is so low usually input source doesn’t have enough power to startup. Have you checked the design using a DC supply? How does the startup waveforms look like?  Does VAUX/VLDO gets charged? Pg 11 of datasheet has fig 1 showing the output voltage sequencing. 

    If the circuit powers up with a power supply means the circuit is working fine but the input source is very weak to power up the circuit. 


  • Hi,

    Thank you very much for your reply.

    I found the startup voltage to be, arround 50 mV. It may be that the PCB is quite warm (~50ºC), because of the high current underneath. I think it has more to do with the temperature, because the source is a voltage drop in a trace carrying > 2-4 Amps. So the current this source can deliver sould be more than enough.

    I could't try it with another supply as the inputs are shorted by this trace i was talking about...

    50 mV is quite good anyway.


    Joan Marc

  • Hi ,

    Using a DC power supply as the input is a good way to check if the part is not working due to weak input source. Voltage drop in a trace can be used as a source with matched impedance for maximum power transfer.  

    Can you share the startup waveforms of VSTORE,VAUX,VLDO,VOUT and PGOOD while powering to look for any further issues.  


  • Hello Anisha,

    In the first place, thank you very much for your time and intrest.

    I attach 4 
    waveform that i recorded in previous tests, but i don't have the Vaux waveform... is it important? i can get it if it is.

    In the "mPV_Vin_Vout3.pdf" waveform, i can see that the Vout voltage starts climing at an input voltage of arround 50 mV. (maybe because i am using AC?)

    Also, i am not using any 
    Vstore capacitor, i only have 100uF on Vout and another 100uF on Vout2.


    P.S. I apologize in advance for the noise that appears on the waveforms, but i think they are usable anyway.


  • Hi,

    All the Waveforms on the same time scale would help you see the sequencing that VAUX gets charged first Once it reaches 2.3V VLDO gets active and  starts getting charged. And VOUT start getting charged as soon as VAUX reaches 2.5V. 

    I could see that VOUT reaches regulation voltage 3.3V. The circuit is working as expected. So what is the issue you are having? 

  • Hi,

    It is just that i am not quite getting the 20 mV startup voltage. I am gettig arround 50 mV which is Fine for my aplication. However, if i could find out how to get those 20mV it would be a lot better. Right now i have two hypotesis:

    • The trace on the bottom layer is at arround 60 ºC, maybe this is afecting the startup voltage.
    • The magnetig field created by this trace carring arround 4 - 10 Amps is afecting the LTC3109.