Luminous Christmas Star With Old Incandescent Bulbs - Circtuit-Sculpture Style

I wanted to reuse incandescent bulbs scavenged from an old Chrismas garland.

I wanted to make something Christmas-related.

I like making circuit sculptures.

So I went for a luminous Christmas-Star, circuit-sculpture style.

I used:

1. 30 light bulbs scavenged from an old Christmas garland
2. 30 15 Ohms CMS resistors
3. 2 MIC5841 shift registers
4. 3 BS250 Mosfets and 3 100 KOhms CMS resistors (which turned-out not being used, see below)
5. 1 47 KOhms resistor
6. 1 10 µF capacitor
7. 1 micro-switch
8. 1 ATMega32u4 on a breakout board such as this one.
9. some brass wire
10. 1 picture frame (mine is from Ikea)

I drew a template (with Fusion 360 and DXF export of the sketch) of the 3 star rings, making sure that I had enough room in the center to put the two shift-registers.

I then used this template to form the star rings with brass rods.

I made sure that the star rings were not joined where I would later put a lamp since this would make soldering much harder.

Unfortunately I don't have any picture of this step.

The circuit is based upon 2 open-collector shift-registers, daisy-chained.

This provides 16 outputs able to drive the incandescent bulbs.

However the max current supported by the MIC5841 does not allow to have all lights on simultaneously.

Thus the software is done so that at any given time we never have more than half the lights on.

The C3 capacitor and R4 resistor are forming a delayed enable feature (to keep all lights off at power-up until the microcontroller is ready).

The slider switch allows to disable the outputs. This is required when programming the ATMega since the circuit is drawing too much power for a computer's USB port and having the lights on would trigger an auto shut-down of the USB ports.

The P-Channel mosfets drive the high-side of the light bulbs.

In theory, this was supposed to work...

But in practice, when lights of the outer ring were on for instance, lights on the other rings were also slightly on. I did not understand why.

Although I'm quite confident that I was driving the Mosfests correctly (gate to close to 0v, via the shift-register output), they did not seem to be fully-on either.

So this made me give-up the idea of having 3 independant light rings and in practice, the gate and drain of the Mostfets have been shorted and they're not used.

The start is thus working in kind of a degraded mode, with a limited set of animations available, but it's not that bad.

I then soldered the light bulbs, one pin on the star ring, the other one soldered to a 15 Ohms CMS Resistor.

These resistors were then soldered to a common brass rod going towards the center of the star (to the open-collector output of a shift-register).

That's when I realized I would be unable to twist these brass rods to make them go to the intended pins...

So I unsolered them from the CMS Resistors and went the other way around (see below).

The shift registers have some pins connected together, such as the power pins for instance.

This is an easy way to have them as a single block that can then be soldered at the center of the star.

I also soldered the capacitor and resistor that are forming a delayed enable feature (to keep all lights off at power-up until the microfontroller is ready).

The slider switch allows to disable the outputs. This is required when programming the ATMega since the circuit is drawing too much power for a computer's USB port and having the lights on would trigger an auto shut-down of the USB ports.

The output of shift-register one needs to be connected to the input of shift-register two, as well.

I used my star template again to form and place the lines going from the bulbs' CMS resistors to the shift registrs outputs (open-collector outputs).

This was a slow and painful process, but that's the beauty of circuit-sculptures.

As always the idea is to use a couple of plyers to bend the rods such as they adopt the correct path, then solder them.

Extra care needs to be taken not to create a short-circuit when crossing lines.

I then placed the star rings on top of the center part, adjusted the various output lines and soldered them to each ring's CMS Resistor. Once again care must be taken not to create a shortcut, especially at the star tips.

At that time I did not know that I would end-up shorting the Mosfets.

So I soldered their 100K pull-up resistor, soldered them to the rings and connected them via enameled copper wire to their shift-register outputs.

I made sure that the power and data lines going down to the ATMega are positioned such as the ATMega is centered.

After carefully checking the wiring, the absence of short-circuit, I did a first power-up test using an external Arduino.

That's when I noticed something was wrong with the way the lights were lighting-up, although my simple breadboard prototype using 3 light bulbs was working as expected.

And I ended-up giving-up the idea of having 3 independant rings of lights, and shorted the mostfets drain and source pins.

I then soldered the ATMega and checked its power usage.

I printed a case for the ATMega, that I stuck on the back of the picture frame.

I scavenged a female USB connector on a circuit board and used it to connect the ATMega to a cable that is running through the back panel. It's an old USB cable with an USB A male connector that is used to power or re-program the ATMega.

I implemented a set of different animations, taking into account the fact that due to the shift-register max current limit we can't have more than haft the lights on at any given time.

Each animation is randomly selected, and its duration is also randomized.

It's mostly bit-shifting stuff or randomly selecting bits for the twinkling effect.

The demo video gives an idea of the final result.

This is not exactly what I expected and the construction is far from perfect, but nevertheless I'm quite happy with the result.