It’ll be a while until I can start actually building something, but that’s no reason to don’t collect ideas in the tea breaks, right? 
I think I want something small and floody in an interesting form factor. Can’t be to complex mechanically, I don’t have a bench vice or something like that. (Yes, it’s going to be somewhat ugly. :P)
I wanted to try out these Nicha NS2L757AT-V1 for quite some time, why not now? Not the typical flashlight emitter, but all those small mid-power LEDs have a special place in my heart. 
Some specs: (is it possible to force line breaks somehow?)
3x3x.5mm package size;
2700k color temp;
83Ra;
120° angle of radiation;
At 180mA, they will deliver 60lm @25°C (107lm/w)
Or at least the datasheet says so. 
I think 5 or 6 in series would be nice.
I have some of those 50x31x7mm 850mAh Lithium polymer batteries here. Including driver (maybe LM3410+µC(tiny10?)+ maybe battery controller) it should all fit in a match box.
I am actually thinking about a box shaped light. I always liked that, but they stopped making them some time before I was born.
Maybe I can make a housing out of brass sheets, have to look if I can handle that with my very basic tools.
I’m looking forward to start building! 
Update: 28th June 2013:
Afte a long and annoying week of exams, I rewarded myself with a few hours of fun.
So I built the power part of the driver:
That LM3410 is really nice. Few parts needed and takes not much space even when built on veroboard. Btw, that’s my index finger, not thumb. (Yeah, I know, not exactly a good design practice for a smps, but it’s fast, and this is a quick&dirty after all.)
It works quite nice, you can see it drive 5 LEDs in series @~125mA from 3.2V here
Maybe I find the time to build the microcontroller part later this Weekend.
I still have not found my brass sheets. But since I messed up the LEDs a little (thankfully, they are to bright in the picture to see it ;-)), I may have to come up with a different idea anyway.
I think I’ll go with the ‘big’ battery
Nice weekend, everyone!
30th June 2013
Now, time to share my failures.
Last time I used the glue, I had quite some time to position the parts.
Problem is, we have a very high air humidity at the moment, significantly reducing the curing time of the cyanoacrylate glue.
But see for yourself:
The emitter boards stuck while I was positioning them, leaving them glued all over the place.
I guess I have to change from polished brass reflector to diffusor screen then.
(And please excuse the crude soldering job, next time I better get some sleep before working. ;-))
I thougt, adding an emitter with a small optic would be nice as an optional light for a little throw.
I found this beautiful 9mm optic in my parts box:
I ordered it as an 20° beam angle, but instead it is this:
Some kind of a bright bar with oval spill around it. Thank you RS components, exactly what I wanted.
I still have no idea what this optic might be useful for.
More as a note to myself:
A rotary encoder would be much nicer for dimming than buttons.
I would need a bigger controller though, and the casing would have to e a little thicker.
I have to think about that . . .
3rd July 2013
I was searching for something to feel better about the oversized controller I have to use . . . and I think I have just found it.
It is a LTC2942, a ‘battery gas gauge’. It can measure the battery voltage, charge/discharge current and temperature. (And calculates the charge in the cell)
The 2mm by 3mm DFN package may not look that hobbyist friendly, but since it has only 6 pins I got away with a sharp knife and a piece of PCB.
It looks a little wild, but after sanding with fine paper and tinning with enough flux, the edges will be nice and clean
The nice thing about small pieces of 0.5mm RF4: you can reflow the parts by pressing the soldering iron tip on the back.
I like the little guy, looks even small next to the current shunt resistor.
I hope that I can get startet with the casing on the weekend.
4th July
I just quickly put the Attiny44a on a board.
Now I actually need to invest a little thought in how I will put it all together.
I found my brass sheets, but that stuff (0.5mm) is actually a little more tough than I remembered. (I hear you metal guys already laughing. :P)
We’ll see if I can actually get it in the shape I want.
6th July
So, I tried to build the first part of the housing.
As promised, it is not very pretty. If you cannot stand seeing crooked constructs and unclean cuts and edges, please look away.
The brass sheet metal is really unpleasant to cut without decent tinsnips or something like that, so I traded a couple of cutting disks for my rotary tool.
Still difficult to bend into shape without a bench vice or at least some strong clamps. So I bent it between two tables.
I guess claping to a piece of wood in the centre would have prevented the buckle it now has.
It also got a little bigger than what I was aiming for.
The edges are soldered with a standard soldering iron btw. Brazing would have been nicer, but it actually feels quite sturdy.
I hope I can polish the soldering somehow to look a little nicer.
That is more or less how I want the stuff inside. I don’t want a knob for the encoder but will try to make a wheel stick out on the longer side.
7th July
I tried for quite some time to make a good enough working reflector, but it doesn’t work like I wanted.
Since I will barely have the time to write te software next week, I decided to use a glass lens I had lying around instead. The beam looks interesting with multiple emitters, but not as bad as I feared.
I probably will have to thange the LiPo for this Li-ion battery since it’s format works better now.
I whish I had more time to make my original idea work, but I simply don’t.
11th July
I got a little time to work on the light today. Most mechanical work is done now. I still have to cut out the back, I think I will do that on sunday.
I soldered two of those threaded tubes (don’t know how they are called) to mount the LEDs on the lens. (And I should have brought my center punch. )
Thermal contact is not great that way, but seems to work better than I feared. The battery will last on full power only 3/4h anyway.
There you can see the emitters throug the lens. Looks a little strange I have to admit, and the soldering doesn’t gain beauty magnified.
But it works ok. Nice and floody, I can barely see the artefacts.
Near the wall the artefacts are clearly visible.
I think I am ok with them since I feared they would be much worse, and they don’t limit the usability that much.
The battery monitor got wired up and taped down on the battery to measure more or less accurate it’s temperature.
(This is the good Chinese high temperature PTFE tape. One of the most useful utilities for soldering these sellers offer, seriously. )
Besides cutting the backside and mounting all the stuff, I have a little software work to do on sunday. But I am quite confident now that I can meet the deadline.
(BTW, the beam looks really beautiful with a single emitter. I will order another one for that purpose with my next FT order. )
16th July
Sadly I didn’t got the time I needed today.
I implemented the basic functions and the light works, the only mechanical work not done is the leather strap I planned.
The stuff fits barely in if you pack it correctly.
But it doesn’t anymore when you let fall it on the ground and have only 10 minutes left to take pictures and post results.
LEDs are also misaligned now. 
But al least it still works.
