Evening forum, hope you are all well.

I have recently taken my first ever foray into solar charging - very small scale, small 6V panel charging 4xAA NiMh cells.

I originally had a 1N5819 schottky diode as reverse-drain barrier and the delta-V between VSol and VBat was quite high - expectedly the same as the Vf of the diode. At 300-400mV, that is quite a proportion of the available 6-5.4 ish volts optimal charge on the cells. So I changed it for an "Ideal Diode" circuit using a balanced pair of PNP transistors to drive a PFET when the two voltages differed, effetively shorting VSol to VBat. Bench tests showed I might get 50mV Vf, tests of the ideal diode showed 10mV, real-world gives 60mV-70mV and sometimes as high as 100mV - all very acceptable. So I am making a lot more of the panel's output available to charge the cells and it shows.

But I have noticed something unexpected that was not there with the previous schottky setup... it's not a problem but has really got me wondering. See the following graphic. This is a little over two days of charging (Y-axis is 10's of mV). Everytime the solar panel output comes off charge and VSol dips below VBat, I see a "rebound" in VBat.



There is another in VSol but as we are moving into indeterminate functionality, I am less curious.

I know there are brains here that have been productive with solar for decades and you must have seen something like this before(?)

I asked google AI but I can't really square it's answer (like who am I to question it when I literally had never done anything with solar charging until a month ago?  ):"When a solar panel stops charging a battery, you might see a slight voltage rise because the battery is no longer being actively discharged by the panel's current. The battery voltage will rebound to its open-circuit voltage, which is the voltage it would have when not connected to any load or charging source."

So charging a battery represents a load to it? I get that charging is effectively a negative current flow from the cells, but I am connecting a lower voltage (VBat) to a higher (VSol) so I would expect a relaxation if anything.

h

---

Only a guess, perhaps the ideal diode doesn't turn off until there is some reverse current flowing. The rebound would then be when that reverse current stops.

Or it could be due to a change in temperature. The terminal voltage of the cells is affected by temperature, falling as the temperature rises.
When the cells are almost full much of the charging current can no longer be absorbed so goes into breaking down water in the electrolyte, making hydrogen and oxygen. As they recombine the energy is dissipated as heat.
Note how the battery voltage is falling faster than the solar voltage toward the end. Perhaps that indicates the cells are full and heating up. When charging stops the cells cool down and the voltage rises.

Some high rate chargers use the drop in cell voltage to switch to trickle charge.
Edited 2025-06-18 09:00 by phil99

  Quote  

So charging a battery represents a load to it? I get that charging is effectively a negative current flow from the cells, but I am connecting a lower voltage (VBat) to a higher (VSol) so I would expect a relaxation if anything.

I guess you can measure the voltage dip to see if power is consumed it that direction.


Cant say I have seen it, I haven't really been looking and the solar array is fairly large I probably couldn't see it with all the other strings contributing.
I have seen some of the solar shed lights, the panels are matched to battery voltage, not sure if they had diodes or not, some are amorphous others are monocrystalline.
The larger solar shed lights have 2 battery cells and not sure what else they have inside, but on a long hot sunny day I have on occasion seen the lights turn on with a dull glow, I assumed this was excess power, cover the panel a little and its off again, uncover it and shortly after the lights are glowing again, the rest of the year this doesn't happen.

cheers for your input guys, It is appreciated.

  Revlac said  

  Quote  

So charging a battery represents a load to it? I get that charging is effectively a negative current flow from the cells, but I am connecting a lower voltage (VBat) to a higher (VSol) so I would expect a relaxation if anything.

...
Cant say I have seen it

No, and it sounds a bit mad to me.

  Revlac said  

The larger solar shed lights have 2 battery cells and not sure what else they have inside, but on a long hot sunny day I have on occasion seen the lights turn on with a dull glow, I assumed this was excess power, cover the panel a little and its off again, uncover it and shortly after the lights are glowing again, the rest of the year this doesn't happen.

Perhaps it is an excellent charge circuit dumping excess energy when the cells are fully charged - I would design that in if were a thinking man  

I have had a day or two more to watch these traces and I agree the odd drop/rebound is almost certainly down to the switching of the ideal diode.

There is another, less pronounced, wrinkle at the begining of the day where VSol approaches VBat and starts to level out before getting a little kick back into the voltage climb. I think this might be down to the old schottky diode starting to conduct right before the FET gets switched on.

  phil99 said  

...
Note how the battery voltage is falling faster than the solar voltage toward the end. Perhaps that indicates the cells are full and heating up. When charging stops the cells cool down and the voltage rises.

There is a fall in VSol which could be Temperature related or b/c the sun has moved round and is no longer perpendicular (but it is still very strong). I have another theory on the clearly related fall in VBat; Panel open circuit voltage (VOC) is inversly affected by temperature (T). Short circuit current (ISC) increases with T but not by as much ratio-wise, Pmax drops (manifests as a percieved increase in panel resistance) and dV increases. A whole charge cycle starts with a dV of around 50mV and ends with as much as 110mV because the panel is roasting in the June sunshine - If this is the cause, I expect this to reduce/disappear in the cooler months.

The sudden drop in VSol is due to the shadow of a 300-yr old Oak tree moving round and hitting the panel around 16:30 ish. The rebound in VSol is likely because it is no-longer under load as the ideal diode has switched it out(?)

This has been a very enjoyable canter working out why I see what is happening and the good thing is it is repeatable - I like repeatable.

All the above is largely guess work based on a little research - what I have is a theory that fits the observations. I am really open to other explanations if anyone has them.

cheers

h
Edited 2025-06-20 22:22 by CaptainBoing