Here's what I know about the recent 110s (with green LED)
After a few hours of reverse engineering, I've realized 2 things. One, my version of the GPL110 is different than that shown around here, the LED is green , not red, the layout is different - but the circuit looks to be similiar. Two, whoever designed this is either a complete idiot - or a total genius.
Here's my schematic, derived from some digital macro photos, head-scratching, in-circuit measurements, and educated guesses.
The ceramic capacitor values are not relevant, but they are probably 100nF or so. Anyway it's pretty straightforward, but after reading several posts that indicate tweaking the pot does nothing, I wondered why. Looking at the circuit a little closer, it soon dawned on me that the voltage drops of the series string LD, Q1, and Q2 would pretty much total 3V. Besides, with Q3 in series with the 1.35V base drive, and Q1 above the negative rail by the saturation voltage of Q2,there is near zero margin to turn ON Q1 - let alone headroom to maintain the LD current with <3V.
Curiosity got the best of me, and by applying about 3.5V from a partly discharged Li-Ion cell, the output woke up noticably. And the green LED was summarily toasted (I replaced it with a better one after the picture was taken). OK then! The design depends on a solid 3V, either by clever intent, misapplication, or just poor design practice.
So using NiMH batteries (their fully charged voltage is lower than alkaline) with this design is doomed from the start, and I really wanted to be able to use NiMH cells.
Since Q1 is the "controlled" pass element, and Q2 is always ON and basically is there to share the dissipation with Q1 (and that is the Good Idea part), with a 3V supply Q2 can be jumpered without any adverse affect (worst-case dissipation in Q1 is under 250mW at ~200mA through the LD). In fact, removing the forward voltage drop of Q2 will allow the pointer to work deeper in the battery discharge curve. So if you want to get a little more battery life out of your pointer, or use rechargables, have a look at the photo and note where the yellow bar is.
If the pot is left untouched, the green output power should remain exactly where is was before, and in fact will be more stable over battery life - because the APC circuit is still doing its job. Note that I do not condone tweaking the pot as this violates the FDC/CRH classification of these pointers! This information is intended solely as an academic study of the circuit.
Here's the rearview of the PCB. This is apparently a new PCB design that AtlasNova is now shipping (black barrelled APC units). The schematic above is valid ONLY for units with this version PCB layout...
Hope this helps, but if you really don't understand it at LEAST understand that the so-called 106 capacitor mod (which defeats the current limit and APC entirely) and any resistor mod that extends the range of the pot to increase the pump diode current beyond 450mA WILL result in damage to your pointer.