I have what’s supposed to be an indoor/outdoor temperature/humidity measuring thing I got at Radio Shack. The outdoor part is a smallish white box and it communicates wirelessly with the indoor station. It’s battery powered, of course, and the instructions say to use lithium batteries if you want it to keep working in sub zero (°F) or so temperatures.
Well, I’ve put lithium batteries in it, but it still spends most of the winter not working, even when it’s merely cold (like in the 20s °F). I don’t know why.
It’d be nice to know, though, because I’m planning on sending an electronic device aloft to where it’s about –60°C.
There are several types of batteries based on lithium compounds. According to Wikipedia, lithium thionyl chloride (Li-SOCl2) batteries are suitable “[f]or low temperature applications”. Also, they “are well suited to extremely low-current applications where long life is necessary, such as wireless alarm systems.” Or PRAM backup batteries in iMacs. When my son’s 2003-vintage iMac went belly up last year, I hoped it was just a dead battery, so I bought a replacement… which didn’t cure the problem. The iMac ended up getting recycled, and the (new) battery just kind of sat around. It is a, wait for it, single 1/2 AA lithium thionyl chloride 3.6 volt 1.2 Ah battery. A Saft 14250, to be exact. Data sheet, you ask? Of course.
It’s rated down to –60°C. At –40°C it’s down to 3.1V and 700 mAh; they don’t show curves below –40°C, but they do give –60°C as the lower end of the operational temperature range. And, like I said, it worked for Rusty Nail.
It’s also kind of chunky for this application: 24.5 mm long compared to 40 mm diameter for a standard ping pong ball. That’s before you put it into a holder… which of course you don’t have to do, you could just solder wires to it, and I might end up doing that. But I was browsing battery holders at mouser.com and found one I liked, of which they had two in stock. I figured I’d better buy them while they were there to be bought, so I sent out an order to Mouser, a little sooner than I’d planned to. Said holder is 34 mm long. It’ll fit, but it’s getting kind of tight!
Now, the Saft 14250 and equivalents are not the only Li-SOCl2 batteries around. For instance there’s the Tadiran TLH-2450 wafer cell, whose diameter is the same as the length of the Saft 14250 — one inch, 24.5 mm — but whose height is only 5.8 mm, compared to the Saft’s 14.5 mm diameter. Nearly a centimeter more room to stack things, nothing to sneeze at. Except that the contacts apparently are wires sticking out one face, 5 mm long. The Tadiran’s rated down to –55°C (though once again the data sheet only shows curves down to –40°C). There’s a serious difference, though: Its maximum recommended continuous current is 5 mA. The Saft’s is 35 mA. As for capacity, the Tadiran’s is 550 mAh.
SparkFun doesn’t seem to quote a current draw for their uLog, but the ATtiny24 draws typically about 1 mA and the temperature sensor — see below — will draw about 1.5 mA. But as for the flash memory… am I reading this right? I think so: For writing, it’s typically 15 mA and could be as high as 35 mA(!). Well, that lets the Tadiran out.
Rusty Nail reports a flight duration of 1:25; let’s allow 3 hours. Then we need 120 mAh. The battery capacities quoted above are at 20–25°C, 1 mA, 2.0 V cutoff. The voltage regulator on the uLog needs at least 2.5 V, though, and at lower temperatures and higher currents the battery capacities go down. At –40°C and 40 mA the Saft is down to 200 mAh! Of course it won’t be that cold for the whole flight. They do plot these things out to 60 mA, so at least it appears the battery can still deliver that current at low temperature.
There are other Li-SOCl2 batteries out there, but I’m guessing the Saft is one of the few that’s small enough and yet capable of driving the uLog at low temperatures. Anyway, it’s what I have, and it looks like it’ll probably work.
What about the temperature sensor? In our last episode I was doing calculations for thermistors. Two problems with them: most of them aren’t rated for –60°C, and they’re very sensitive to temperature — meaning it’s hard to cover a range of 80°C with decent precision. Then I stumbled across a writeup of temperature sensors that pointed me toward resistance temperature detectors (RTDs). Their dependence on temperature is close to linear (as opposed to thermistors whose resistance is basically proportional to exp(B/T)), so they’re better able to cover a wide temperature range, and I’ve seen some rated to –200°C costing under $10 each. The one I’ve ordered from Mouser is even cheaper and rated to –70°C. Its nominal resistance is 1 kOhm, and I think the best divider setup is with a fixed resistor of 1 kOhm. Hence, with 3V across the divider, the figure of 1.5 mA above.
Okay, so I think I know pretty well how this PongSat will go together, and I’ll soon have all the parts.
 And non-standard ping pong balls? You can get 50 mm ones. But that’s cheating. Or at least I think so. JP Aerospace might accommodate them, but I’m not asking.