Sunday, September 13, 2015

Silicon casting to fix a Power Wheels

The final result, two new gears.
My nephew's Power Wheels jeep stopped working a while ago. We were able to track the problem down to a stripped gear causing the electronic differential to stop working.

Differentials, both mechanical and electronic, send power to the wheel with least resistance. This was useful in our case because it meant that the stripped gear would just spin and it's counterpart gear would receive no power. A positive transmission would have kept delivering power to the remaining gear and destroyed it as well. By having a good gear to reference, I was able to fabricate an exact copy by using a silicon mold resin casting process.

This was my first attempt at doing something like this. I'll also use this post as a place to follow up and capture lessons learned if this new gear ends up not lasting.

Making the mold was a fairly straightforward process. I hot glued the good gear to a piece of wood and then hot glued the top of a paper cup around it to hold the silicon while it cured. In the foreground is the gear and behind it is the vacuum chamber which is degassing the silicon mold material.

Before and after pouring the silicon mold material, I placed it in a vacuum chamber to degas. Here's a pictures of bubbles evacuating from the silicon. I mixed the binary liquid together using a wooden paddle and this always introduces air into the mix. Surrounding the mixture with a negative pressure causes this air to expand, float to the surface, and eventually leave the mixture. The material expands quite a bit in it's liquid form while this happens, so degassing before pouring ensures that the material doesn't overflow the mold reservoir and spill into the chamber. In the picture below you can see the "high water" mark of the expansion of the silicon before it settled back down to closer to it's original volume. I determined how much silicon mold material I needed by pouring dried rice into my mold form and then pouring it into a cup. I then put a mark on the cup and mixed material to that level.

After pouring the silicon into the mold form, degassing, and then waiting for it to react, this was the result. The form is a near perfect mold of the gear. Thanks to the degassing steps, there were no bubble pockets on the working surface of the mold.

The same degassing process was used for the epoxy resin I used to cast the actual new gear: degas in the mixing cup, and then degass again in the mold. The below picture is of the second degassing step, after I've poured the mixed epoxy into the casting mold. I wrapped duct tape around the edges of the silicon mold to create a reservoir. In the future, I think I'll mold a reservoir lip into the actual mold so I don't have to do this. The tape absolutely refused to stick to the silicon so this is mostly a pressure fit with the tape only sticking to itself.

I did not use any mold release spray for the first gear and it came out of the mold just fine. This gear will be ready to use after sanding the back flat on a belt sander.

Parts list and additional notes:
My vacuum chamber consisted of a 90$ Harbor Freight 2.5cfm automotive vacuum pump, 1/4" OD tubing, a cutoff valve, a wide section of steel pipe stuck to a flat piece of steel with silicon sealer, a piece of craft foam for a gasket, and some plexiglass.

The valve is necessary to sustain a vacuum. If you turn off the pump and leave it sit, the negative pressure in the chamber will start to suck oil from the pump and spray it into the chamber. My initial attempt at using an aluminium pot for the chamber resulted in crushing the pot, so be sure to test your chamber before putting anything important in it.

The blue gear is straight casting epoxy. The red gear I mixed in some sawdust. I'm not sure if the sawdust will add strength, but it did take up some of the volume and saved epoxy resin.

I was able to attain different colors by adding normal food coloring during the mixing step. Adding this color also ensured a uniform mix of the epoxy, once the drop of color was evenly distributed I would know the two transparent liquids were probably mixed thoroughly.