The Raspberry Pi: My First Hardware Experiment

My shiny tiny new Raspberry Pi arrived last month, bringing with it a lot of hoopla and renewed enthusiasm for computing. Now that my first task of getting Slackware ARM installed to it, and configuring it to run “headless” (network only, no video or keyboard connected), it’s time to turn my attention to some hardware hacking. I decided to try out some basic cooling systems for it, to see if they make any appreciable difference in the RPi performance.

Not being able to afford a lab bench, or even access to a laboratory, I opted for a basic setup on my kitchen table. My first experiment focused on a forced-air cooling system (all images are clickable):

The RPi, air-cooled.

This cooling system made no measurable difference in the temperature of the device under test, even under full load. This is not surprising, since the RPi is very easy on the power requirements. Performance likewise remained unaffected.

Next, I tried water-cooling:

Basic water-cooling system.

The results were rather disappointing. The investment in equipment provided no return for system cooling. I’m glad I didn’t go out and buy some water-purification equipment to go with it; I doubt it would have helped.

Next, I tried an oil-cooled system. Since the CO₂ footprint of running the Raspberry Pi is so small, I opted for a plant-based oil instead of the usual mineral oil:

Extra-virgin olive oil, no extra cooling.

Again, no difference. I was really beginning to wonder if my experiment was going wrong, or if I really didn’t know what I was doing.

I remembered reading about some guys who wanted to overclock their system, so they stuck the motherboard in the freezer with bottles of vodka, while they progressively adjusted the jumpers and checked the frame rates in Doom. Well, the strongest thing I drink is beer, and I don’t want my RPi to draw condensation, so I opted for a slightly modified version of their hack:

Yuengling Black & Tan, right out of the fridge.

I don’t know if it cooled the RPi any, or made the performance better. Before I did the test, I took another bottle out and drank it, so if I did run the test, I don’t remember where I put the file with the results. But I know it made me run a lot cooler!

In conclusion, the Raspberry Pi’s own cooling system is sufficient for just about any purpose. Kudos to Raspberry Pi designer Eben Upton for such a *ahem* cool design!