Okay I took apart a simple bench light (the kind with tons of super bright LED's) and tried to reverse engineer the tiny little circuit board.
First, all the tiny SMD components (diodes and transistors) have no markings so I used LTSpice defaults. No big whoop. I think.
It's got two, paralleled 4V sealed lead acid pieces of crap, and they're charged directly via the 120v wall outlet thru a full bridge (8-10 hours to charge, and only
5 hours usage on HI

; which means maybe 2-4 AHr batteries? No markings on anything). It has a switch to select HI, LOW, and OFF, and I think all those do is switch in some different biasing resistors for a couple of transistors to vary the current thru the LED array. BTW, I only show the HI configuration in the LTSpice circuit .
I measured the actual and it's about 360mA thru the LED array when ON HI.
If you look at the attached LTSpice I made, the simulation results do pretty much match what I measured on the real unit. But I'm trying to figure a couple things:
- WTF is the D5, and why is the bridge output circuit hooked up like that? Maybe I goofed? Makes no sense, but seems to work and I triple checked the connections on the board. Makes my brain explode.
- The RC frontend (200k) I suppose blocks DC and limits the battery charging current? I don't understand how the output of the 120V rectifier matches the required 4+volts to charge the battery. DOH!!!
- And one other PITA about this piece of garbage is that you can't have the light ON when it's plugged into the wall (it doesn't charge the battery). But I don't see how that is reflected in the circuit.
Anyway, thanks for any help.