Audioguru, AK, and Martin,
Here is my circuit as described earlier, powered by a Lithium Ion battery at 3.7v nominal and dropping volts over time.
To sum up - I am trying to design a long-lasting solar light for my patio that will last in full brightness from dusk to dawn, and a bonus if it goes on for days. I will add a low voltage drop-out later and add then a solar charger circuit. I don't like the cheap dusk-to-2am solar lights that are in the stores.
This is the exact circuit drawn with the EveryCircuit app. I pulled the three parallel LEDs from a flashlight. Their little led holder board had no electronics but did have a 10 ohm resistor that I kept. The upper resistor in this simulation represents the CdS light sensor. This particular image represents dark conditions and shows the circuit at 24.3 ma, 8.07 ma for each led. I have the LEDs set 120 degrees apart on a horizontal plane in my test pickle jar and the light is fine for my purposes. The CsD sensor changes all the way down to 400 ohms in high sun and up to 400k ohms in the full dark.
The above image resents the ohms in normal daylight. The "sleeping" circuit draws about 0.15 ma (0.16 in simulation).
This image is a mystery to me. It represents the simulator's math in that I should get 92 ma draw above 400k ohms. On the bench, and not connected to anything, the CsD sensor easily reaches 400k ohms when I cover it with a dark, dense, black towel. In fact, it can reach 1 meg. But, if I cover my actual circuit's CsD sensor, the Fluke measurement does not exceed 29 ma, in extreme darkness. A head-scratcher. But, for me, that is a relief for I don't want more than 29 ma at the battery. So, the extreme simulated condition shown above with LEDs at 30.8 ma can not be duplicated in the actual circuit.
Now, for your comments about my battery. I did not give the details and I confess I was not too educated on Li-Ion cells. The Li-Ion cell I am using was salvaged out of a Asus Chromebook. It is one-half of what the specs say is a 7.4 volt, 4400 maH battery. I separated the two halves and have now two Li-Ion flat batteries. I assume the longevity of operation I am experiencing is due to that massive maH spec of 4400. Previously, above in this thread, I thought all such cells were 1800 maH. (I checked on eBay, and very similar flat cells are going for about $15 each.) It is about 3 1/3" square and 3/16" thick. It is bendable slightly and indicates by bend lines that there are 4 cells in each half, obviously wired in parallel and brought out to two wires providing 3.7 volts for each half. Each cell seems to be about the size of a stick of gum.
on edit - the transistor is a 2N2907 PNP.