Forever Snowing & Glowing Snow Globe

This holiday season I had decided to make gifts for each of my children (and one of their partners) and fortunately one of them had mentioned their love of snow globes. But the problem with snow globes is that the snowing stops and they quickly become a little boring. So I decided to fix that problem once and for all with a snow globe where it always snows and with a little added magical illumination too.

One of the challenges of a snow globe is that everthing has to be water-tight and so moving parts (and even electrical wiring) going from outside the globe to the inside would be hard to accommodate due to the need to seal the water inside permanently. If overcoming this challenge sounds like fun to you then read on!

Here’s how it works, what I used, and how you can build one too.

Let It Snow, Let It Snow, Let It Snow

Normally you need to shake a snow globe to get it snowing and it only snows for a little while before it needs another shake. Clearly this isn’t high-tech enough for any gift I might give so a little agitation is needed. In chemistry labs, magnetic stirrers are quite common and I found this cheap and cheerful option on Amazon.

Inside there’s a motor that drives a pair of magnets in a dog-bone arrangement and a small circuit board with a switched potentiometer to control the speed of rotation.

A PTFE coated magnet is also supplied with the idea that this is placed in the liquid to be stirred.

The unit was provided with a 9V power adapter and I wanted to use 5V instead as USB power would be more convenient. In addition, the motor was a little on the large side for the snow globe I wanted to use (which has a 100mm diameter) and I decided to replace it with a smaller one which fortunately had the same output shaft diameter. The electronics of the stirrer then needed some work because the smaller motor spun far more quickly than the larger one had. A quick look at the PCB indicated a step-down voltage converter MP1484M and I traced parts of the circuit to reveal the following (this is partially complete as I’d learned what I needed to know).

The datasheet for MP1484M indicates that the voltage is a function of R1 and R2 as follows:

Since R1 is a variable resistor, changing R2 to a higher value is the way to reduce the output voltage range. The value on the PCB was 9.1kΩ so I changed it to 56kΩ (based on the idea that around 15% of the existing range of the potentiometer was usable with 9.1kΩ so a 6x factor might be good.

After a bit of work on a 3D printed base unit (more on that later) I was able to test the assembly with a partially filled globe with some snow (again more later).

A Warm Glow

Some time ago my friend Chris Greening (Atomic14 on YouTube) made a video about wireless LEDs …

The principle of these wireless lights is the same one used in electric induction hobs (or stovetops if you speak USA). The principle of electromagnetic induction was discovered by Michael Faraday almost 200 years ago and says that when two conducting objects are placed close together, turning on an electic current in one induces a current in the other. To make this work well coils of wire are used and the Wireless LEDs product I purchased has a 70cm diameter driven coil along with much smaller (around 5mm diameter) receiving coils.

I had never had a good reason to use these in a project previously but now it seemed appropriate to see if they might work inside a water-filled globe. A few simple experiments (without actually immersing them) seemed to indicate that they still had good range operating through a layer of water.

As Chris describes in his follow-up video you can make your own version of these but I found hacking the bought ones worked well enough for my purposes and with a bit of careful tweeking I was able to make the snow globe base work with the 70cm diameter driver coil so I didn’t have to rewind that either.

I soldered power cables to the marked + and – terminals on the driver PCB and connected these to the USB 5V input on the motor driver PCB.

LED Experiments

To make the globe personal I decided to light up the name of the recipient to appear magically suspended in the water. The letter design was based on a traced and tailored version of an image that ChatGPT created.

I tried 3D printing the letters and uplighting them with one wireless LED each but the effect was very dim and wouldn’t have even been very visible in a darkened room. My first idea to solve this was to put multiple coils inside each letter but the coil orientation wasn’t ideal for this (the coil should be in the same orientation as the driving coil to achieve maximal inductive coupling) and an attempt to increase the power to the coil resulted in a fried driver chip.

There are a number of longer range coils available but these generally use larger driving coils which wouldn’t fit my stand design, so instead I decided to remove the LED from the receiving coils and add a bunch of LEDs inside each letter, placing the coil itself as low down in the globe as possible as this maximises the inductive coupling which results in the highest LED brightness.

While that worked it certainly made for difficult construction and, in the end, I found it simplest to lay out all of the LEDs on kapton tape (reversed so that the sticky side was uppermost) on top of a printed outline of the letters.

Placing all of the LEDs in parallel like this only works if the LEDs are close to identical and I certainly found that a bit of selection was required with the cheapo LEDs I sourced – and even then there are some differences in brightness. Next time I’ll try sourcing better LEDs or introduce some series resistance.

One trick I didn’t think of is documented in Chris’s video (if only I’d rewatched it fully before sealing things up!), which is to place half of the LEDs the opposite way around to take advantage of both cycles of the alternating current. Another placeholder for “next time”!

Floating Letters

Since I wanted to max the size of the letters it wasn’t possible to create the whole insert in one piece (as it would have been impossible to get it into the globe – the old ship in the bottle problem!). So I designed a two part solution where the letters, stems and coil holder are one piece and the base with a cage for the rotating PTFE magnet a second one.

It took many attempts to achieve a suitable result and ultimately I decided I needed a less rigid construction so that it could be coerced into the globe with a bit of judicious squeezing.

In addition, I decided to try to make the contents of the globe appear to float, so I supported the letters on thin, flexible stems (and placed the coils as low as possible), running wires down the stems to the coils. Printing the letters in yellow (Lucy’s favourite colour) PETG (which isn’t so water absorbent as PLA) and the stems in a clear version of the same material results in the letters appearing to float in the globe, especially when the snow is flying around.

Assembling the Letters

To keep the LEDs in place I used double-sided adhesive sheet (cut to shape) and attached near to the front of the letter shell. Then, with a bit of jiggling to seat the letters and a bit of soldering to the coil (with LED removed) the LEDs were securely placed.

I completed the letter assembly by encasing both the letters and the coils in silicone sealant (although maybe I should have used a conformal coating as I think that might have bonded more thoroughly to the electronics).

Final testing that it really does fit!

The bottom of the insert which holds the rotating magnet also has a slot which the tab on the bottom of the letter assembly fits into to hold the two pieces together.

Filling with snow (and fluid)

There are holes in the insert which allow water (and “snow”) to be added after the insert is in place. After researching what people commonly use for the “snow” and liquid in snow globes, I decided to just get a ready-mixed solution and add distilled water. While this isn’t the cheapest solution it worked well from the start and I avoided a lot of experimentation with viscosity and clarity that seem to plague other snow globe creators.

Designing the Stand

An important aspect of the stand design was to ensure that the driver coil nestles as close to the base of the snow globe as possible so that it is approximately in the same horizontal plane as the receiver coils.

The back of the stand has a small panel on which a USB C connector and speed control sit.

The 100mm snow globe comes with a bung which works ok on it’s own but I decided to add sealant around the rim to try to ensure its long-lasting water tightness. I guess only time will tell whether this worked or not.

The Final Cut

Ultimately I’m pretty happy with the end result.

Using Chris’s idea to reverse half of the LEDs in the u and y would probably have ensured that these letters (which contain more LEDs) were a bit brighter – but overall I think a success – and the recipient seemed happy which is all that counts really.

Design Resources

All of the design files are on GitHub including Fusion 360 files and 3D printer files.

Supplies Used

Component	Source
Glass Snow Globe 100mm	https://www.amazon.co.uk/dp/B0DZVC2D2M
Magnetic Stirrer	https://www.amazon.co.uk/dp/B07J59QVGQ
Small electric motor	https://www.amazon.co.uk/dp/B0D2J4HLCJ
Wireless LED lights	https://www.amazon.co.uk/dp/B0DBZ5WC4
Replacement LEDs	https://www.amazon.co.uk/dp/B0BJ8KTXBY
Yellow PETG	https://www.amazon.co.uk/dp/B0D4Z9718K
Clear PETG	https://www.amazon.co.uk/dp/B0D4Z7V8YB
Ready mixed snow solution	https://www.amazon.co.uk/dp/B0DMTDQDP3
Distilled water	https://www.amazon.co.uk/dp/B0DB2YJZWP
Clear sealant	https://www.amazon.co.uk/dp/B0D1VP8J65