Hi Bruno28
I think others have given you fair warning of the difficulties of "design by committee". But it sounds like a great project and I've often wished I had the time, tools and skills to do something similar.
You will never make something to satisfy everyone as some want low-powered EDC (AAA) lights and some want Li-Ion for the power. Some will want throw, others flood. High CRI is essential as is maximum output coupled with extremely long runtimes. And of course 3 "moonlight" modes between 0.01 and 1 lumens is imperative. It must light up that threatening shape at 200 yards and also have a pocket clip so I can keep it unobrusively tucked in my jeans.
You can see what I'm getting at.
If you only plan to make the light(s) yourself and sell via CPF then it seems you can make pretty much anything. As long as it is functional and percieved to be of high quality, someone will buy it. If it is "pretty" or made in limited numbers (preferably from titanium) then you can command a very nice price. Add 50% if you track the serial numbers for every light you sell. There are some really nice lights in the custom threads as examples.
If you want to look at something more mass-market with a view to having it produced commercially, then you need to look at what features will make it appealing to the everday user. Spend countless hours coming up with a great product, prototype it, have it reviewed and approved and be lauded for your efforts. Then contract a Chinese company to produce it and watch clones of your light dropped onto DX or KD before your first batch even hits the shelves.
Wow, I think I've just unloaded most of my cynicism..... now to more constructive things.
My 2c - A modular approach would be ideal.
Dimensionally, it would be great if you could design a host (head, tube, tailcap) capable of taking several different battery types. 18650s are essential for CPF buyers and AA is where it's at for the general public. There are a few lights which can accomodate 3AAA or one 18650 but 3*AA/1*18650/1*26650 is probably more flexible. (Spark SX5 for example although that is a headlamp).
The reception of the 4AA lights like the Nitecore EA4 and Sunwayman D40 prove that there is definitely a market for a high-powered AA light and quite a few users in here who still fondly appreciate their eneloops.
A P60 drop-in may not be where you want to compete but the idea of selectable emitters is sound. It should be possible to source two or three reflectors (of the same size) to suit a range of emitters. Nichia 219, XP-G2, XP-E2 and XM-L2 are where it's at presently. (Perhaps leave the triples and quads to those specialising in making very nice modules already.)
Many makers currently offer protruding clickies or flush tailcaps for tailstanding so it shouldn't be difficult to offer this as well.
Drivers? Unless you can be sure of offering a 4*AA rather then 3*AA body, then a buck/boost driver is probably essential or your maximum output will fall rapidly on the NiMH cells. Unfortunately, this would add to the cost. Most people would probably expect to be able to run 2*RCR in an 18650 light so a driver rated for up to nearly 9V or so would be smart.
The trouble here is making mode selections that suit the majority of people. Too few or the wrong choices (that insane H-L-S found in the cheapies for instance) will turn many people off. Custom builders like Vinh/Nailbender can offer your choice of modes at the time of order, so perhaps you could as well.
The idea of making the light programable (Armytek, HDS, etc) sounds nice but I suspect it would only appeal to a small section of the market. I'd love it, but many would find it overly complex.
Potted electronics is a nice quality touch and thermal step-down is sensible if you are offering high-output options.
Keep in mind whether you are producing a light for sale "as is" or whether you are considering that some of your customers may want to do mods. Some modders don't like loctite, potting and thermal glues. Personally, I think if you can offer sensible options at order time there should be no need to plan for modding.
If you decide to start with the single AAA light, then keep in mind some of the suggestions above. At the bare minimum, a choice of emitters to bias towards flood or throw with options of high CRI or neutral tint would be well-recieved.
Options - people want options. Make several features selectable and your customers will feel they can get exactly what they need instead of something that is "almost right."
If you'ver read this wall of text all the way through, congratulations and good luck with your project.
Subscribed!