The energy source will be a ~6.5V solar panel (circa 5-6 watts) with built-in buck converter (output 5V, max. 1A) that feeds a USB port mounted on the back of the panel. The panel works fine on its own, for charging phones and powering small USB desk fans, but I wanted to make it more dependable as currently the output stops every time a cloud passes overhead or I shade part of the panel with my hand.
The reason for using capacitors instead of batteries is mainly as an experiment, but also because they have significantly greater cycle life. I have 6x 2.7V 500F super-capacitors on order, along with protection circuitry. I also have a variety of DC-DC buck/boost converters I can use (some with adjustable constant-current, which would work great for charging the capacitors I suspect).
What I need help with is coming up with the best way to connect everything together.
Ideally, the build would implement this algorithm:
- If solar panel output is > load, excess energy is stored in capacitors
- If solar panel output is < load, deficit is made up using energy stored in capacitors
- In each case, as much of the solar panel output as possible is driven to the load before the capacitors get involved.
I've considered simply putting a capacitor string (for example, 2 in series to give 5.4V 250F) in parallel with the panel output, but the fraction of capacitor stored energy that is usable in that case would be quite low (maybe the interval between 4.5V and 5V). Also if caps start from a low voltage it would take a while for them to charge up to usable voltage.
I've considered connecting the caps in parallel with the solar panel (before its built-in buck converter) but this again can only utilise the inverval between 6.5V and 5V. Not to mention the need for a blocking diode to stop the caps discharging through the panel in the shade.
I've also considered feeding a boost converter that charges the capacitors, and a buck converter taking capacitor voltage and using it to drive the USB output - this kind of path:
Solar panel output (5V) -> Boost converter (CC 2A, CV 16.2V) -> Capacitor bank (6 in series for 16.2V 83.33F) -> Buck converter (5V) -> USB output.
But the problem with that is all the conversion stages give rather low efficiency from the solar panel to the final output.
Is there a way to implement the "ideal" algorithm electronically? Where panel output is routed directly to the load when possible (to preserve efficiency) but capacitors can be utilised over most of their voltage range (to have useful output in absence of sunlight)?
The reason for using capacitors instead of batteries is mainly as an experiment, but also because they have significantly greater cycle life. I have 6x 2.7V 500F super-capacitors on order, along with protection circuitry. I also have a variety of DC-DC buck/boost converters I can use (some with adjustable constant-current, which would work great for charging the capacitors I suspect).
What I need help with is coming up with the best way to connect everything together.
Ideally, the build would implement this algorithm:
- If solar panel output is > load, excess energy is stored in capacitors
- If solar panel output is < load, deficit is made up using energy stored in capacitors
- In each case, as much of the solar panel output as possible is driven to the load before the capacitors get involved.
I've considered simply putting a capacitor string (for example, 2 in series to give 5.4V 250F) in parallel with the panel output, but the fraction of capacitor stored energy that is usable in that case would be quite low (maybe the interval between 4.5V and 5V). Also if caps start from a low voltage it would take a while for them to charge up to usable voltage.
I've considered connecting the caps in parallel with the solar panel (before its built-in buck converter) but this again can only utilise the inverval between 6.5V and 5V. Not to mention the need for a blocking diode to stop the caps discharging through the panel in the shade.
I've also considered feeding a boost converter that charges the capacitors, and a buck converter taking capacitor voltage and using it to drive the USB output - this kind of path:
Solar panel output (5V) -> Boost converter (CC 2A, CV 16.2V) -> Capacitor bank (6 in series for 16.2V 83.33F) -> Buck converter (5V) -> USB output.
But the problem with that is all the conversion stages give rather low efficiency from the solar panel to the final output.
Is there a way to implement the "ideal" algorithm electronically? Where panel output is routed directly to the load when possible (to preserve efficiency) but capacitors can be utilised over most of their voltage range (to have useful output in absence of sunlight)?
