Radiosonde Hunting - A Post-Mortem

Yesterday, I tried to hunt a radiosonde. I was unsuccessful. I'm documenting my attempt here, as well as how I intend to improve for next time.

To start, I should probably explain what a radiosonde is. Radiosondes are launched every 12 hours from a number of stations around the world. They're essentially small devices containing a GPS, radio transmitter, and some sensors. Over an hour or two they're carried by a balloon up to around 30km, then the balloon bursts and they fall to the ground. Radiosondes are disposable, so there's a bit of a hobby around finding and collecting them. Plus, they can be reprogrammed for ham radio purposes. I'd like to get one and port CATS onto it.

It's currently winter here in Atlantic Canada, so radiosonde hunting is more difficult than usual. There's a station in Maine that follows the normal 12-hour launch cadence, and most of their radiosondes land within a few hours of me. The issue is, most of our back-roads are unplowed, and radiosondes tend to land in the middle of nowhere. I live in New Brunswick, and there's a lot more empty space than there are cities!

Yesterday, Friday morning, I did what I normally do, and checked Sondehub to see if any radiosondes were headed my way. To my surprise, the local military base - which follows its own, seemingly-random schedule - had launched one! Not only that, but it was predicted to land about 2 hours from me, very close to a plowed road.

After thinking about it for a bit, I decided to message my manager and let him know I was taking the morning off. I scrambled to get most of my radiosonde tracking gear into my car. The gear consisted of a battery, small computer, and SDR. I used magnets to hold an antenna and GPS to the roof of my vehicle. I also brought an old Android phone with me so I could access the computer from its web UI - I'm still on my Blackberry, after all! I would have brought a handheld transceiver as well, but in my rush, I left it at home.

Once loaded up in the car, I put on some music and started driving. Like I said, it took about 2 hours, and was mainly across back-roads. Aside from some pothole dodging, it was pretty uneventful. I began experiencing GPS problems near the end of the drive, which I believe was from leaving the unprotected GPS module on the roof of my car while it was raining.

Once I got to the predicted landing site of the radiosonde - or more accurately, the road next to it - I initially had trouble picking it up. I didn't write down any details of the radiosonde, so I had to do a bit of MacGyvering to get the sim card out of my Blackberry and into my Android phone. The main issue there was not having a paperclip, but I found a mechanical pencil in my glove box and was able to use the lead to very carefully eject the SIM card holder.

After looking up the radiosonde, I found it was transmitting on 404MHz. I plugged that into my mobile transceiver, and I heard it chirping away. I was really happy about this - it meant I wasn't too late and its batteries hadn't died yet. Still nothing getting detected by my computer, though, so I slowly drove up and down the dirt road, looking for the place where the signal was strongest. Even then, I couldn't decode the radiosonde. In a last-ditch effort, I moved the SDR off of the LNA and onto my car antenna. This worked, and I started decoding the radiosonde transmissions!

Now that I knew its precise reported location, I grabbed my phone and wandered into the forest. It was about kilometer from the road. While the trees were fairly dense, there were also some less dense paths. I made the mistake of running initially - a very exhausting endeavour in deep snow - and was out of breath for the rest of the walk.

I probably got within a hundred feet of it, but absolutely could not find it. It's a small white box, and blends in well to the white snow. I also looked up in case it got caught in a tree, but still, nothing. If I had my handheld transceiver with me, I could have followed the signal strength, but without it, I was lost. I also felt time pressure - I had only taken the morning off, after all - and I could feel the snow melting in my shoes and making my socks wet. In general, I didn't want to be here longer than I had to.

Getting back to the car was a lot harder than I expected. Even with GPS, I found it was super hard to consistently walk in a straight line. Interference from the forestry made the GPS inaccurate and at one point, I found myself walking in the opposite direction!

From there, it was a fairly anti-climactic drive back home. I was a bit disappointed that I hadn't found the radiosonde, but I was also glad I decided to go. I learned a lot and next time I'll be more prepared.

The biggest thing is, I need to bring more stuff. Next time, I'm going to bring the following:

• Handheld transceiver, Yagi antenna, and offset attenuator for direction-finding

• Boots

• An extra pair of socks

• A laptop

  • Although the phone worked fine, I'd feel a bit more comfortable having a proper machine as well

• Weatherproof GPS antenna

  • I'm going to keep the actual GPS inside the car, and just the antenna will go on the roof

I should have taken the full day off. Being crunched for time made the experience a lot less pleasant, and I underestimated how long it would take me anyway.

I'm not entirely sure why I was having problems with the magnetic antenna. I think that might be an issue with the LNA - I may have been overloading the SDR. I wish I tried taking the LNA out of the signal path before changing out the antenna.

Another issue I had was my phone disconnected from the computer, and I didn't notice. Normally you can tell from the GPS data getting stale, but because I was having GPS issues, the data was already old. If I hadn't noticed, I could have spent more time debugging my RF chain. I need to be more conscious of this in the future, and refresh the web page often. I use a tool called chasemapper for the web UI, and while I offered to add some kind of indication when the server link was lost, the maintainer I spoke to seemed uninterested. His recommendation was to just fix my GPS problems. Then, I could use that as a litmus test for server connectivity. That seems fair enough to me.