On April 30th, Cole and I met at my home in Haiku to try and get the Chill Box operational. This was to be our last attempt, as we are down to the last days of The Challenge and need to at least get the project documentation complete so we qualify.
The original water pump was a cheap one completely sealed in plastic. Tests on thursday had shown it gets very warm when running – putting heat INTO the cold system and completely defeating the cooling effect of the peltier. Since we were replacing it with the micro-pump, we decided to fire up the Dremel and see what was inside. A quick bit of surgery (we are hackers after all), showed that it was basically a standard DC motor stuck on an impeller and enclosed in plastic. We exposed most of the metal motor body and then sealed it up with silicone sealant. We didnt test it but perhaps it will still work for some other project – with better heat dissipation characteristics.
The Corsair fan needed a larger inlet vent, so we used the laser cutter to clean up the messy wave side of the acrylic box. We had previously broken off some of the waves that were partially attached. Putting a 4.5″ square worked pretty well. The box was a bit deep, barely fitting in the laser. Even though it was a bit high, it still managed to cut. I put a piece of cardboard over the other side to protect it. Good thing too – although the laser was fairly diffuse by the time it reached it, the cardboard still got burned. On the second pass, I put a piece of scrap wood over the bottom side.
Cole brought his micro pump. This tiny 5vdc pump turns out to be quite powerful, silent and runs very cool. It also has a 3rd fitting for use as a drain. We took out the old pump, rerouted the tubing, replacing much of it with aquarium tubing, and filled the system. The drain valve now exits the bottom front next to the corsair fan. It was MUCH easier to fill and bleed air with this new setup. However, the T-fitting that attaches the (rather bulky but effective) fill valve leaks on the chiller side. It appears the weight of the chiller system is bending the plastic tubing. We stuck a plastic cup under the chiller to support it, and a bit of tubing to align it with the T-fitting. This looks like a major hack, but it stopped the drip. Water circulates quite nicely. We were pleasantly surprised with how effective the micro-pump is at moving water, even with the larger volume tubing.
After we got the whole thing setup again, we decided to give the chiller a full shakedown – run the peltier and see if it can condense water. If it works, then we’re good and will proceed with rigging up a better chiller support and power wiring setup. If it doesnt, well, it was a good learning experience. So we hooked the peltier and Corsair up to the 12v battery, and the little micro-pump to the 5vdc out on Cole’s hacked PC power supply. It starts off looking good. Water circulates, the peltier (or at least the copper block) starts getting cooler. We let it sit – I had to whack Cole a couple times to keep him from tweaking plumping or other parts. The setup is rather fragile and each time he tried to make it better, something got worse. Distract ourselves, we started working on the project documentation. After about 5 minutes we checked the system. The copper block itself seemed to be getting quite cold (condensing a bit where it was exposed). The screws attaching the acrylic water channel to the copper were getting cold. But the pipes were not condensing. They were a bit cooler than ambient, but not much.
At this point, we quickly discussed some modification (try to replace epoxied in peltier, rip out the whole chiller and replace it with an alternative system, turn it into a simple drip irrigation with servo control, etc) and dismissed them all. It was already 6pm on Saturday. We had only 3 days until the challenge was over, and barely any documentation. If we wanted any chance to complete the challenge, we at least needed to document the failure, and lessons learned (things to do different next time). So we shut down development, took a bunch of pictures and called it a night.
Its been a good run. We learned A LOT. We need to sit down and write up a full set of lessons learned for the documentation. I documented some of those lessons in the Acrylic Chill Box Build blog post. One for the chiller system is ‘do not irreversibly connect (epoxy) critical, expensive components‘. Another is to listen when lots of people say an approach (peltier) is highly inefficient.
Here’s the Flicker Set for the April 30 build:
[slickr-flickr tag="ChillBox-April30" descriptions="on" flickr_link="on" sort="title" use_key="y" ]