i made a graph (i like graphs)
the following is the relative efficiency increase from a given starting current if you double the number of chips and cut the current to each in half
View attachment 3798079
this was with citi 1212
i did some ultra-low current measurements lately with some multiple chip designs to get the low current data. it seems to fit with the other tests but as always take it for what its worth. overall it looks pretty linear to me.
so starting at 3.2A if you double the number of chips and run them at 1.6A you will gain a whopping 26% in efficiency
but say youre at 1.6A and do the same, you gain a little over 15%
once youre at 0.8A, if you try to double you only gain about 9%
obviously the law of diminishing returns is in effect, and the steps are cumulative
from 3.2 to 0.8A is 1.26*1.15=1.45 (+45% efficiency@ 4X the chips)
from 3.2 to 0.4A is 1.26*1.15*1.09=1.58 (+58% efficiency@ 8X the chips)
its all relative..... so taking 30% to 30*1.45= 51% (in other words not adding 45% to 30% to get 75%)
you save on heat abatement as efficiency goes up as well.
now you see why 700-1750 mA is popular