A story of the electronics and logistics of launching a high altitude balloon.
For the enclosure you need something light, rigid, with decent thermal insulation. Enter Styrofoam. A 2 x 4 ft sheet from the local hardware store was ~$5. I hate stryofoam. HATE it! If you try to cut it with a knife, you'll end up with tons of statically charged little balls running around. I learned this the hard way building the GPS clock. After getting a flood of emails talking about Nichrome wire, I decided to give it a try. Nichrome wire is the stuff in your toaster. With a little voltage/current, the wire heats up nicely and cuts through styrofoam like butter. For us, it certainly cut quickly, but it was impossible to cut a straight and perpendicular line by hand. This is important as you will need a square box without many holes for the cold air to get in. What we needed was a styrofoam cutting machine, but that would be a project unto itself (and I consider it yak shaving). So instead, I swallowed my OCD and we cut the styrofoam with a regular kitchen knife. Not great, but it worked.
Lessons learned: The box doesn't have to be all that perfect. The metallic duct tape works really well adhering to the styrofoam and will fill in or cover up most of the gaps in the seams. You don't need to go overboard with silicon sealant or making sure the thing is totally air tight - on the contrary, you'll want it to be slightly leaky so that as the air pressure drops to near nothing, the box doesn't explode at 100,000ft. The metallic tape acts as a radar reflector (fulfills one of the FAA requirements).
Building windows: We decided to use small pieces of acrylic to make a window for the still and video cameras to view out. These were a huge hassle that I believe we will avoid on the next build. The windows did help to seal up the box against air flow, but with a properly, tight fit camera, you will get a decent seal. Cutting into the wall to fit the camera also helps keep it in place as the payload starts to spin or shift.
Make sure to label your box! And label it well. Your name, phone number, address, and something about how it is an 'Amateur weather balloon project' should be included. A note stating 'no boy inside' is also acceptable, but the humor may be lost on the person that finds your payload. Laminate the label. Put markings on your PCBs. Include something like a metal tag or identifier that won't degrade over time. My payload is currently lost somewhere out on the eastern plain of Colorado and while I marked the box, I didn't laminate the paper and am worried that even if found, the person will only be able to find me because the PCBs have 'SparkFun.com', 'Load Control', and 'Every Sensor We Know' printed on them. Fingers crossed.
Here you can see inside the box. Nothing really to it. Camera window in the bottom left corner. This was one of the first 'freezer tests' with an Arduino and the BMP085 pressure and temperature sensor, and an OpenLog for datalogging. It's a little bit surprising to throw this box in the freezer for 12 hours, pull the thing out, and discover it's still happily ticking along, cold, but functional.
Here's what the first freezer test showed. The logger worked! The temperature coming off the BMP085 pressure sensor is very accurate. You can see that the freezer takes a around 15 minutes to start to drop the internal box temperature from ambient (~25C) to 0C in 35 minutes! The freezer hits the end of its first cycle around 60 minutes when the box temp pops up a bit, then the freezer kicks on again and starts to drive us all the way down to -18C and holds there. Obviously the box is not sealed, but if this payload is going to 100,000 ft, the enclosure is not going to be able to maintain the starting temperature on the ground and things are going to get REALLY cold. Electronics really do work down at -20C (the test Arduino and OpenLog were still blinking after six hours!). But -50C can be a major problem for batteries, so let's heat the inside of the box...
Electronics will operate in cold conditions, but the outside temperatures can dip to negative 50 Celsius at 100,000 ft up. The enclosure will help protect them a bit, but batteries tend to lose their potential (voltage) and things just generally go bad. I don't need a comfy 20C (70F) in the box, I just need the internals of the box to stay at or near freezing. The batteries and electronics should be quite happy at 0C. So I set off to design a heater system to help generate just a little bit of heat during the flight.
We discussed the load controller in the Emergency Cut-Down segment of the The Balloon, Enclosure, Helium, and Cut-Down tutorial.
I limited myself to a single 2000mAh battery. Assuming the payload would be in the air for 5 hours, this allowed me to source 2000 / 5 = 400mA for heat. 3.7V @ 400mA is (3.7 * 0.4 = 1.48W) nearly 1.5 watts of power! That's not a bad little heater. All I had to do was find a power resistor that could handle that much current and heat. Never forget Ohm's law: V = I * R. Looking for the resistance, 3.7V / 0.4 = 9.25Ohm. Cool. I ordered a couple different 2W power resistors off Digikey (part numbers 13W-2-ND and 7.5W-2-ND). Attaching the 13 Ohm to a LiPo battery, it did indeed heat up. The 7.5 Ohm resistor heated up just past the point of me being able to touch the resistor (hot!) so I went with it. Select the resistor as you see fit.
Here you can see the power resistor attached to the load controller (schematic, eagle files, firmware), inside the box with some metallic tape running around the edge. The load controller has a LM335A temperature sensor on board and the firmware was designed to cycle the relay (turn on/off the heater) to try to maintain a balmy 0C inside the enclosure. There is also the BMP085 pressure sensor (I was using it for temperature only at this point), an Arduino, and an OpenLog to log everything (temp and milliseconds passed).
The resistor is free floating in the picture, but during testing, the resistor was taped down firmly with metallic tape. This allowed for good thermal conduction from the resistor to the tape, and then around the inside of the box. This worked very well to heat the inside of the box. The problem with metallic tape however is that it's conductive! So be sure to watch the bottom side of your PCBs and electronics - if they touch the tape they could short out.
I threw in some freshly charged batteries, sealed everything up and threw it back into the cold corner of the freezer.
Here we can see the heater actually took the inside of the box up over ambient. The heater did work! It staved off cold death for about 200 minutes (3.3 hours) before it looks like the battery kicked the bucket and the freezer started really doing its job. We can see the same slope at 217 minutes as the box crosses the 0C mark and approaches the static temperature of around negative 15C.
So what does this show us? While this was a good exercise, it wasn't productive for me to test at this point. In the end, testing with the full setup, I discovered that the XTend radio heats up like a light bulb - very toasty! It turns out I didn't need a heater at all! A microcontroller is designed to be very low power, but a radio can get very warm. In future launches I will probably use a different radio that could be potentially much lower heat output, so I will need to retest. If you're doing your own launch, be sure to run a number of test with the full system operating from the freezer to make sure you see how the enclosure interacts with the heat coming off the different sub-systems. If you are at all in doubt, I would recommend a heater but it does add a layer of complexity and weight.
Next page - Weights, Measures, and Costs