With the Printed Circuit Board (PCB) designed, the next job was getting it manufactured. Lacking the tools to do so myself, and having no experience ordering a PCB, this was a daunting prospect. Less daunting was sourcing the components for the circuit itself, with electronics components suppliers readily available online. I went with Bitsbox, a UK supplier whom over two orders were very quick to dispatch and deliver the components.

The components ordered include:

• 10x 1K resistors
• 10x 330R resistors
• 10x 5mm Ultra-bright white LEDs
• 10x 12mm tactile push switches
• 2.1mm DC power socket
• 2.1mm DC power plug
• 1x 4xAA battery holder
• 1x 9V battery clip (connects to the battery holder)
• 3m speaker wire (3m additional in second order)
• Heat-shrink sleaving
• Solder

The speaker wire is essentially a pair of wires bound together, which would make managing the cabling easier. The particular wire I ordered was grey with a black stripe down one side, which made it easier to keep track of the circuit polarity (as LEDs, being diodes, only work in one direction).

Trying to find a PCB manufacturer was difficult, as I am a hobbyist rather than someone looking to manufacture large batches of a given PCB. I initially tried to contact a few UK-based manufacturers, but with very limited success (as they seemed more set up for business-to-business interaction rather than making a small batch of custom boards). Looking to hobbyist websites it seemed that going to a Chinese manufacturer was the most cost-effective and easiest way to go, so I tried PCBWay. The interface seemed straight-forward enough to use and the pricing wasn’t astronomical, the only downside was I couldn’t get the extra holes drilled and the minimum batch order was 5. This wasn’t too much of an issue as having spares would be handy, and the PCB layout could work for any generic lighting circuit up to 6 LEDs so could be useful in future projects. Including delivery, I went from order to having the PCBs in hand in just over a week:

Lacking the ability (and skill) to drill the extra holes myself (the manufacturer did drill the holes in the circuit contacts), I figured the best thing to do was to solder the whole circuit together when the gauntlet’s underlayer was ready to have it installed, since once it was secured in place any stresses on the wires should be minimised. The first task however was to solder the resistors into place, which I could do right away.

I used to be pretty good at soldering back in secondary school, I think I’ve gone a bit rusty. Once soldered, I tested the circuit with a multimeter to make sure the solder joints were conducting current adequately.

The rest of the circuit involved measuring the approximate length of each wire needed, based on the planned route from the board to the component’s location on the gauntlet (either on the back of the hand or on the palm), adding a generous extra length (at least 10%) for losses when soldering, as well as a bit of room for placing the wires. (Better to have a bit too much cable than not enough).

The LEDs were relatively easy to solder, with long legs they’re designed to be soldered to wires or directly into boards. The main concern was ensuring I soldered the correct wire to the correct leg (I wanted to keep the wire with the black stripe as the positive/anode).

The tactile switches were a pain, since they are very much designed to be soldered directly onto a board, and not to wires. The tabs are very small and getting the wire to mesh with them enough to take solder was fiddly to say the least. The solder joints are far from the best I’ve ever done. But a few hours of work (with testing the circuit after each pair of LED/switch were added) thankfully yielded a successful result:

Now with the lighting circuit made (and the gauntlet’s underlayer, which was ready by this point but I will document in a later post) it was finally time to bring it together…

The most important components of the “Infinite dGauntlet” are, of course, the Dice themselves. Sourcing the dice was relatively easy, thanks to TheDiceShopOnline where I was easily able to select translucent dice in the appropriate colours:

Coming up with a means of securing these dice to a gauntlet, whilst at the same time incorporating a 5mm LED for illumination purposes would be incredibly difficult to do by hand. Fortunately I had a relatively new hobby to bring to bear: Enter 3D Printing! Before I could use a computer to solve all my problems, however, I had to get the measurements.

Questionably accurate measurements in hand, I downloaded Blender (a pretty good open-source 3D modelling software) and figured out how to kind-of work it in creating-STL-file mode. (The trick is figuring out how to set up the scene units, then how to set the correct scale factor when exporting.) The first step was to create a correctly sized “LED cutout” STL file I could then re-import and subtract in order to have a uniform insert for a 5mm LED.

Working out the external dimensions was a process of (roughly) measuring my left hand to figure out how wide a mount could be and still reasonably fit four across my knuckles. From there it was a case of estimating how large to make each one to fit a 5mm LED in addition to the half-dice, and figuring out the exact sizing within Blender by trying to fit everything together.

The d6 and d12 were the most straightforward, with the cube and dodecahedron being regular solids to use within Blender. A test-print of the d6 mount revealed the dimensions as measured were a bit tight, so I added around 10-15% to the measurements as I would eventually be gluing the dice in, and having them fit a little loose would be less of a problem than them being too tight.

The d20 was similar, with an icosahedron not difficult to produce within Blender. It was slightly more complicated due to being the physically largest dice, but being set in the larger oval mount on the back of the palm:

The d8 and d10 were slightly more complicated, but I reasoned they are both effectively two cones with their bases stuck together, each cone having half the total number of sides. So a 4-sided and 5-sided cone would fit each dice, and also hold them higher giving more room for the LED.

The d4, being a tetrahedron, was the most complex. I had originally thought of mounting it inverted, but that would mean the base of the dice sticking out as opposed to a given value. It would be possible to mount it in the “correct” orientation, but given its concave shape, it would require creating the mount in two halves, and gluing them around the dice:

In reality, I went through the process of designing and printing these mounts one by one, adjusting and reprinting each as necessary. (The d4 required about 4 reprints to get the size right.) My 3D printer, a Monoprice Mini Delta V2, was able to print a mount in around an hour.

With the dice now fitting, I could test the fit for the LED. (The LED cutout seemed to work first time, given the manufacturing tolerances of a 5mm LED I suspect it was my measurements of the dice that were lacking.)

With the d4 mount complete, I focused on the rest:

Final thing to do was add a coat of gold paint, and a bit of silver inside in an attempt to act as a foil to reflect light back through the dice:

When conceiving the idea for what I’m loosely calling the “Infinite dGauntlet” a number of ideas came to mind. The “ideal” functionality and look that I would aim for, and the levels below that I could achieve if resources/time became too much of a constraint. The ideal in this case would be individual illuminations for each dice, and the ability to selectively brighten any/all of the dice by the action of closing the fist. Failing that, illuminating the dice would be a more achievable goal.

5mm Ultra-bright White LEDs seemed an obvious choice, they would likely be small enough to fit inside a 3D-printed mount for each dice yet powerful enough to provide sufficient illumination.
For control, a few options came to mind. Reed switches combined with magnets sewed into the fingertips might have worked, but with a number of the switches in close proximity it may have been difficult to selectively activate them. I eventually decided the simplest option would be some variety of tactile push-to-make switches, generally flatter I figured they would quite easily be sewn into the gauntlet.

So the question arose: How to use a push-to-make switch to make an LED brighter when pushed? A brief bit of research on LED brightness (googling and answer on StackExchange) unearthed the fact that LED brightness is roughly proportional to the current flowing through it. Remembering Ohm’s Law from secondary school:

   

Where V is Voltage, I is current and R is resistance

The plan for the circuit is to run on 4xAA batteries, so the voltage across the LED will be fairly constant. This means Current will be inversely proportional to resistance.

The second piece of the solution comes from how components of a given resistance (namely resistors) combine. In series, resistances combine by simple addition, but in parallel it is their inverses (or conductances) that add together:

   

Essentially this means that two resistors in parallel have a total resistance that is less than either resistor individually. This means I could place a push-to-make switch alongside a second resistor, both connected in parallel with the first. Then when the button is pushed, the resistance of the entire segment is reduced, the current increases and the LED shines more brightly.

Given my knowledge from secondary school electronics, I recall we would always use a 220 Ohm resistor with LEDs, so figured that was a reasonable lowest resistance/highest current for the “active” higher illumination state. I wanted there to be a noticeable difference between the two states, so went with a 1,000 Ohm resistor for the inactive state. Knowing the desired total resistance and value of resistor R₁ figuring out a suitable R₂ was relatively straightforward:

The exact R₂ I would need was 282 Ohms, but 330 Ohms (a more standard fixed resistor value) got me to 248 Ohms, which was close enough. So the basic unit of the circuit would involve an LED and 1 Kilo-ohm resistor, always connected, with a second 330 Ohm resistor and switch in parallel with the 1 Kilo-ohm resistor. Multiply that by six for each light, and the circuit diagram looks something like:

All good in theory, but before I want to the trouble (and expense) of turning this into a circuit board I wanted to be sure it would actually work. It was finally time for me to get myself a breadboard. I found this kit on Amazon for a reasonable price and quick delivery, and before long I had wired up a prototype.

With the circuit figured out, the next task was transferring it to a Printed Circuit Board layout. Once again my experience from school was limited, the software we used (which I cannot remember the name of) was fairly basic. KiCad seemed to be the best open source software for designing PCB layouts, though it has something of a learning curve. (Multilayer PCBs, solder masks and a whole catalogue of component footprints.) However I managed to figure out enough to lay out the circuit I needed. The main challenge was figuring out a design that would be small enough to fit into the gauntlet, but not so small as to be too difficult to solder together. With six LEDs and six tactile switches, as well as a power input, there would be 13 pairs of wires to attach which seemed like it would be too crowded, but this is what I eventually came up with:

The holes were intended as mounting points for the board as a whole, as well as a means by which to secure the wires. In secondary school electronics we would thread a wire through larger drill holes in the PCB before soldering, as doing so meant any forces on the cable wouldn’t be directed into the solder joint, which can be brittle and break easily. I wasn’t sure whether I could find a manufacturer who could drill the holes as specified, or indeed if holes so close together were even feasible for the board material. However if it wasn’t possible I had planned to secure the wires to the gauntlet itself so, once installed, stresses on the solder joints should be minimal.

Now all that was left to do was order in the components and try to get the board manufactured.