Deep Freeze Blog & Press

< Back to News

The Museum

Product development started in April 2013, and we were rolling out production units in October 2013. 6 months is INCREDIBLY fast for a new product, considering the injection mold alone took about 2 months to make, and we had to go through a lot of iterations in testing.

A trip through memory lane with all of our iterations in design is in order.

Here’s our first design and prototype from April.


The very first rendering. We used a cylindrical window, a cap on top, and a piece of velcro to attach.


This is the initial prototype. The top has tin foil glued to the plastic to act as the reflector.

Velcro wasn’t going to work, though, so we moved on to a different design that used snaps. This would be easier to manufacture and fit on to the poles better.


This 3D printed part was sanded and smoothed and treated with acetone for many hours to make a smooth surface. This was the part used for making molds.

This was the next revision. The side now has a channel for snapping onto a flag, and there’s a groove for the switch, which required a separate part. The groove was too small and we couldn’t get our fingers in to turn on the switch, especially with gloved hands.

This is what it looks like straight out of the 3D printer. It takes a LOT of work to sand it to something usable, and even then there was warpage, it didn’t have the tolerances necessary to snap, and it didn’t have all the features it should have because the resolution wasn’t high.


3D printed parts were extremely rough.



The inside was rough, too, but that wasn’t as important.

After painting to approximate the look, it hardly looked any better. We joked that it’d probably look better painted brown so it would look like wood.


After painting the parts, they didn’t look much better.


Next we tried molding for higher volumes. 3D printing wouldn’t work; it was too slow and too low quality. We took a 3D printed part, sanded and treated it until it looked fantastic, and used it as the positive for a mold. We then tried casting a variety of materials from Smooth-On to get a part we could use. This is where we ran into all kinds of problems, like bubbles, warping after demolding, uneven clamp pressure (which meant that our thin walls in some places were made even thinner, insufficient filling, and overfilling. It was a mess and a disaster.


These parts were cast from the mold and had bubbles and weren’t completely filled.


So many bubbles! We even vacuum degassed it!

Incomplete filling and inconsistent clamp pressure made walls that were too thin.



The opening clogged and wouldn’t let more material in. The other side overcompensated, but the wall was still too thin.

Even if we had figured out the outsides, the insides weren’t turning out well either. The inside has bosses that hold the screws, and they just weren’t filling out well enough, and they certainly weren’t strong enough to hold our product. We were stuck without a way to develop a medium volume product run.


The inside of a cast part. The bosses for the screws just didn’t work.

Fortunately, we were moving so fast, and interest was so high, that we figured we’d move on ahead and make an injection mold. Here’s one of the final 3D printed prototypes of the model. In another post we talk about the method used to put the logos on the part. We had to use this for some nice photos of the product that we would use at a trade show and to show off to potential retailers. This is what we used until we had our first injection molded parts.


A ‘nice’ looking 3D printed part.

Once our injection mold tooling was complete, we got our first parts:


The first enclosure from the injection mold.

The color was wrong, the top color was wrong, but this was just a sample that showed the product that came out of the mold. The plastic was the worst quality possible, and we were extremely nervous when it was easily scratched, the bosses broke, and the feel was strange. Fortunately, this was only a product of the cheap chinese test plastic.

Once our mold was done, we experimented with a few different cap options for coloring.


Choosing between different cap options. Polypropylene on the far right won.

From left to right; 3D printed PLA, ABS, Blue Polycarbonate, Clear Polycarbonate, PolyPropylene. PP was our favorite, but because it has a different shrink rate, we needed to make a change to our mold. Oh well. It was a small change, and the product is so much better for it.