I just got an idea from a very another group, (man this guy didn't have a clue), but it gave me an idea a wicked one at that.
Li-ion's have a tendency to gobble up all the current they can get until they are about 70% full, reaching max voltage usually 4.2V per cell. then most chargers then switch to a constant voltage and let the cells eat whatever current they like which starts being less and less near full charge (full charge .03 of C in A).
Now the li-ion chemistry will also let you charge them to 4.05V per cell or even less and then start taking constant voltage and the current will follow the same curve as well as if you had charges it to 4.2V and then switched to constant voltage. 4.05V will conserve cell life BTW harmful reactions (to the cell life) Start over 4.05V.
Now 12V solar panels tend to have a mpp of around 17V add a diode and you got 16.2V-16.5V. Over this voltage the current delivered by the solar panel will drop dramatically, Now once a bank of four cells (or multiple banks of fours) reaches 16.2V it will attempt to rise higher, but since the current will drop-off so will the charging voltage (now this will only work if the li-ion bank size is several times the capacity in C of the max current of your solar array. If you put a simple monitor that will cut off the power if a cell reaches 4.2V (max safe voltage) or total 16.8V the cells which are not fully charges will drop back down to around 4.05V per cell total 16.2 V which is your diode modified MPP. If you have a bank that is 10X greater in C than the max current of your array I doubt that the voltage will even rise past the MPP until the cells are as full as they will get at a 4.05V charging curve which is an estimated 85% of C. (Mind you if you go the 4.05/cell route even if the voltage then rises at reduced current you will only get like 86% to 90% of capacity of the cell max.)
This would have 2 advantages
#1 it would be safe as long as two conditions are met (maximum cut-off 16.8V) and a modified MPP around 16.2V (A regular diode should do the trick as long as you don't make that diodes capacity too big - higher capacity less voltage drop :S) A safety temp cut-off would be a bonus)
#2 since the voltage from the get-go is higher you would get more power out of your panels since 80% of the discharge power is above 3.5V in graphite based cells(most cells these days are graphite based) 60% in coke but until 80% even coke will stay over 3V so even if you discharged your bank by 60% you'd have a starting charge voltage of 14V. Lead acids discharged by 40% drop down to like below 12V.
So if you had 6A of charging current that would be an additional 12 Wh well you can do the math.
Question: what do you guys think? Viable? Take a close look at mmp curves.
And also another thought If you have to deal with depressions in the mmp due to temperature a latching relay system that puts the power through diodes with different voltage drops may be the solution. Should be simple with a temperature probe circuit.
