Most of those I’ve ever seen, regulation is zero. As in, no resistors. It just relies on the internal resistance of crappy hateful little AAA alkaleaks to “regulate” current. That, and hair-thin springs, oxidised contacts, other resistances that add up a little more to limit current.
Figure 9 LEDs in parallel should draw maybe 200mA to be “safe”. So, okay, 4.5V - 3.5V or so leaves 1.0V to drop. In theory.
Realistically, the LEDs get hit harder with fresh cells, but once they start to “wear”, internal resistance increases, so at first you might get 300mA or more through all 9, which’ll drop to 250mA, then 200mA, then progressively less and less as you spend the cells. By the time they’re pitifully dim that you want to swap in new cells, terminal voltage from the 3 cells will be about 3.0V, with the LEDs just barely glowing. So the cells are down to about 1.0V each, which sounds about right.
Naw, most of those LED boards are just a quick swirl of peecee board traces of the 9 LEDs in parallel, no resistor included.
Can’t believe I wouldn’t let this go…
Goggled around, found https://d2ei442zrkqy2u.cloudfront.net/wp-content/uploads/2016/03/MN2400_US_CT1.pdf , so resistance starts out at around 0.25Ω and rises to about 0.40Ω near end-of-life. So figure for a trio of AAAs, that’s 0.75Ω up to 1.20Ω.
A’ight, so that 1.0V drop from above, add in another fractional-ohm to maybe 1Ω total… about 1A total for 9 LEDs? Sounds awfully high, but then again, those lights are impressively bright with fresh cells. So maybe.
Never bothered to measure current on those critters, but it doesn’t really change the fact that most of those cheapcrap lights don’t even bother with a ballast resistor to limit current.