The Brigadier Project

[tetsujin]

I've been working on this lately and just kinda want to talk about it a bit I guess, since I haven't got anything else to show for it yet...

Basically, Bandai has experimented with including LED lighting with some of their kits in the past, usually by providing the parts and making electrical connections with screws or crimp-beads. In the last few years they've adopted a much simpler approach: lightable kits include space for their "LED unit", a self-contained package (19mm wide, 8mm tall cylinder) containing a LED, a battery, and a power switch, and clear light piping is used to guide the light from the LED unit to where it is to be seen.

I've generally had mixed feelings about this approach, I usually prefer to just put LEDs exactly where light is supposed to appear and run wires through the model. Light piping means more opportunities for light leaks, and potentially more compromises in the design of the model... But if nothing else I've become interested in exploiting the potential of the LED unit cavities that are becoming common on new Gundam kits, as a place to keep batteries and circuits. Plus more generally I've been wanting to design a small microcontroller board with a battery holder on it anyway (seriously, it's like the only compact battery holders out there have to be mounted to a PCB.) So I've been working on creating my own LED unit.

There's a few basic problems with Bandai's LED units... I can't solve all of 'em but the ones that bug me most are:

• Poor quality switches: The power switches are unreliable and so sometimes I gotta tap 'em or something to get them to work. But a lot of kits at least provide access to the switch even when the LED unit is sealed inside the model.
• Small batteries: Although the LED unit is an 18mm cylinder, it uses rather small batteries: two LR41 cells, about 5mm in diameter.
• Limited color selection: So far the kits that use the LED units have needed red, green, yellow, and pink LED units, but at present only the green ones are commonly available. I've worked around this in the past by cracking LED units open and soldering on a different LED.
• Cost: They're not too expensive... In Japan. About $5 each. In the US they're around twice as much.

So the basic idea is something like this:

• Activation via Hall Effect (magnet) sensor - So no need to physically access the LED unit to turn it on...
• 18mm diameter PCB with a battery clip for a 16mm coin cell. Also gonna use a voltage booster so I can bleed the coin cell dry
• RGB LED (they're pretty cheap these days! And great for replicating lots of uncommon colors)
• Cheaper (for me at least) than a Bandai LED unit, and with greater capabilities.

It's kind of a dual-purpose design, in that it's for me but I may choose to sell it too, so I want to make it pretty easy to work with. So for people who don't want to get too complicated there's two pushbuttons - a "mode" button and an "adjust" button, used to control basic things like color and blinking patterns. For fancy guys like me, it can be hooked up to a PC and reconfigured or reprogrammed that way. I thought about different methods of connecting the module to the PC but there's really not a lot of space for a pin header or something like that. The simplest solution, apparently (and almost bizarrely) seemed to be to put a USB connector right on the board and use that to communicate with the PC. Micro-USB connectors are pretty small, actually, and everybody can connect to that without any special extra gear. But that does add a bit of cost to the project: around $2-$3 per board extra for the USB microcontroller and connector. So I'm also making a second, "economy" version that omits USB.

I'm using kicad for the design: it's got a 3-D feature, so I used it to help visualize the board I'm designing:





There are still some issues I need to work out with the design: There are traces and pads close to the edges of the board, and I think the PCB fabrication service might reject the design as a result... Also the pushbuttons are too close to the edge... I don't want them getting mashed against the side of the LED unit cavity. I've reworked the design several times already and there's always something that needs correcting. I really wish I had it sorted already so I could order a few prototypes and try 'em out.

In smallish quantities, the cost for me is about $8 for the USB version and $5 for the non-USB version. Less in higher quantities obviously. I'm planning to make the two versions work together: there's a pair of data lines set aside for communicating between LED units: when these are tied together on multiple LED units they'll form a little network so they can do things like activate/deactivate together and pass instructions to one another. It's a bit overkill, having one microcontroller per LED just for that, but going that route could make it very easy for people... But not -too- easy I guess, since there's still soldering involved.

Anyway, that's what I've been up to lately.

[Madman Lighting]

Hi George,

Sounds like an interesting project. I've used KiCAD for several projects and its a good beginner Schematic / PCB package. I find that at its current level of development its functional, but clunky. However, for a free tool, its pretty good.

What you want to do with your microcontroller / LED widget is familiar to me, but you will need to write very much code to make all those things happen. I'm also familiar with putting a USB port on the micro so it can talk to the PC. If you choose this route, you want a micro that supports USB directly with ports specifically for that purpose, you will need an external crystal, to get the timing of the USB signals accurate enough to meet spec (for most micros) and unless you are already familiar with the complexities of USB enumeration (and they are complex!) get some kind of off the shelf thing to do it. Very many people make the USB widget show up as a virtual COM (serial) port and send text commands to the micro.

Good luck.... its definitely an ambitious personal project.

Your alternate is some sort of Arduino widget for the hardware and writing all the code yourself, without the Arduino studio stuff, which is good to get started with, but heavy and somewhat clunky.

[tetsujin]

What you want to do with your microcontroller / LED widget is familiar to me, but you will need to write very much code to make all those things happen.

That's OK, I'm awesome at code. Though I guess I'll have to see how much program memory that code will take up...

If you choose this route, you want a micro that supports USB directly with ports specifically for that purpose, you will need an external crystal, to get the timing of the USB signals accurate enough to meet spec (for most micros)

Yeah, already got that stuff. The 3-D renders I did (linked in the previous post since they can't be inlined) were a complete (but not finalized) design, using ATMega16u2 and a ceramic resonator with spec'd tolerances sufficient for USB operation.

and unless you are already familiar with the complexities of USB enumeration (and they are complex!) get some kind of off the shelf thing to do it.

It's called LUFA. :) I have made USB devices before - a couple gamepads, nothing too fancy - though admittedly, not without the help of someone else's library.

I considered vUSB as a contender briefly (it can do low-speed USB on a microcontroller without special USB hardware) - but in the end there wasn't a lot of benefit. Cost-wise, USB hardware adds about a dollar over similarly-spec'd AVRs, (vUSB costs more than that, if you don't want to use the GPL license) - the only other potential benefit, really, would be getting an AVR that's physically smaller, to save board space. Even then it wouldn't save that much space.

Very many people make the USB widget show up as a virtual COM (serial) port and send text commands to the micro.

Yeah, pretty much Arduino-style. I'm not sure exactly what I'll do if I opt to sell this thing, but going Arduino-style is generally a good bet IMO. I think a lot of people regard Arduino as a lot easier to work with than a bare microcontroller, even though there really isn't much difference. :)

Your alternate is some sort of Arduino widget for the hardware and writing all the code yourself, without the Arduino studio stuff, which is good to get started with, but heavy and somewhat clunky.

None of the off-the-rack Arduinos are small enough. The smallest Arduino-ish board I've seen is the Teensy 2.0 (love that board!), and that's like 30mm long and 18mm wide. The LED units are about 19mm diameter, circular, incorporate a LED (centered to line up with the light guide) and a power supply. Even if I didn't rely on the kit-stock LED socket, it might be hard to find a place to install a board like the Teensy or Arduino Pro Mini. In my recent work on the MG Rick Dias, it was actually surprisingly hard to find a good location for one of my Raven boards.. and those are MUCH smaller - around 9mm x 11mm, smaller when I cut off extra I/O pins I didn't need.

Really, if we're talking just about my own needs... I don't need the "Brigadier". I don't need a USB port built into the board, don't need pushbuttons for changing the LED color, I don't even need to use the LED unit socket. For most projects, my little Raven board (microcontroller + resistors for driving LEDs, in a small package) is enough. But I've been meaning to make a microcontroller board with a coin cell holder on it, and it might be nice to have a board that fits those LED unit sockets. But a lot of what's going into the Brigadier design is for the benefit of other people who would use this thing if I sell it. I don't know if I will do that but it's a fun design challenge, thinking about how to make this thing kind of accessible for non-techies.

[tetsujin]

Damnit... After all the time I spent designing this thing, I finally ordered some boards, then went to Mouser to get the components... They no longer sell the voltage booster IC I used in small quantities. Now I've got to move my whole bill of materials over to another supplier or find a substitute part that I can order from Mouser, and possibly redesign the board again for the new part...

Oh, and their website is cripplingly slow at the moment. That certainly doesn't help matters.

(EDIT): Eh, I dealt with it... Just split the BOM across Mouser and Digi-Key. This is exactly the sort of thing I worry about when I spend hours and hours designing a PCB, though - I worry that suddenly one of the components I chose for the design will be unavailable and I'll have to design the thing all over again. Though hopefully the "non-stocked" items will be stocked again and this is just a temporary thing...

But the good bit is, I've finally ordered some prototypes! Now I just need to figure out how to assemble them. Hot plate maybe?

[Rondie]

Just thought I'd chip in with the smallest arduino-ish board I've seen till now, the
Tinylily mini. Round board which measures in at 14.0mm (0.55 inch) diameter.

[tetsujin]

Thanks! I hadn't seen that one before, that's pretty cool. Some neat stuff to be found in the E-textiles crowd...

I like the Lilypad battery holders, too - one of the smallest "off-the-rack" battery holders I've managed to find, actually...

[Mr. Engineer]

Heh. Happened to me last year when I was designing for the Moebius Pegasus. All of a sudden, the LED design I based on went the way of the Dodo.

Lastly, just to share my experience with you, I did a lot of customisation for one-off Clients and they do not want the 'extra' features much as these will affect my turnaround time and also other unforseen stuff such as component obsolescence and/or their unavailability. I tried to design a lot of features for their 'what-ifs' but at the end of the day, the time and effort put in is just not worth it.

[Madman Lighting]

Welcome to the wonderful world of standard vs custom products.

The only way I've ever found to make real money is to have as much common to all your products as possible, keep your parts count as small as possible, and do all the "customization" in software. Even in software you have to apply those principles too! 90% of my code is the same for all my products, the only part that changes is the data tables that control what the special FX does, and even there I cut and paste lots.

So does it bother anyone that my KBop engine effect is very similar to my Fast Tramp engine effect? Not that I can tell. What about my Wave Motion Gun FX being alot like the KBop torpedo FX? (it is you know....)

Invent once, maybe twice. Sell many many times.

Then go invent something else to sell lots of.

[Mr. Engineer]

Ha ha ha! Thanks for sharing with us on this, Madman!

That was something I learnt last year and so, my latest product would be just that. But then, one thing led to another and I relented and used a jumper system since the models I had in mind was big enough for the design. In the end, it was a more or less a 7 in 1 Lighting system...