The latest release (version 1.3) is 4 layers and weights around 3 grams when printed on a 1.6mm PCB (can further be reduced in the future). It also have some interesting new features: allows self-calibration, can measure its own power supply and has an onboard temperature sensor. Here it is:
On the software side, I might use OrionWspr by Michael Babineau which now supports my board. As an alternative, a simpler version I am developing could also be deployed.
The part I am still wondering about is power. Contrary to my initial understanding, it is, after all, possible to avoid using batteries (or super-capacitors) and power the balloon by solar panels only. That would require somewhat larger panels (probably around 6 or 7 grams) that could generate enough current for the board to run under direct sunlight. An alternative option is to use lighter panel s (1.5 grams or even less) that feed a power charger which in turn charges a 1.5 Farad (5 grams) super-capacitor. Finally, there is the possibility of using lighter panels which directly drive a capacitor without using any chargers. This raw method seems to be preferred by other balloonists but I still don’t have a solution for it.
I recently built a couple of identical WSPR transmitters to be used in direct comparisons between magnetic loops and other antennas, or even between various kinds of magnetic loops. They are now available on Github.
The transmitters are based upon the Si5351 clock generator and their RF section was inspired by the excellent Zachtec’s products. An onboard GPS receiver is used to automatically provide timing and location. In the absence of a GPS fix a button can be used to start transmitting. The same button can also be used to cycle between 5 predefined WSPR frequencies. There are no onboard filters therefore external filters must be used in order to maintain compliance with FCC rules. SMA connectors provide connectivity for both the HF and the GPS antennas.
The firmware is based upon existing libraries and was created around version 1.0 of the board. Version 1.1 introduces an important extra feature: auto-calibration.
I also built another magnetic loop, identical to the one I already have. Together, they will serve as a testbed for my comparison tests.
I really like going QRP from a not so popular location on East River, right between Manhattan and Queens. I normally don’t plan for it. I just go whenever I feel like it and, obviously, I have time to do that.
Yesterday was one of such occasions. It was a particularly interesting day because the WPX CW Contest was going to start at 8PM local time and I had come home relatively early from work.
I wasn’t planning to try to contact CW stations at the very beginning of the contest (too late for me to stay at the river, bands too crowded). Instead, the idea was trying to catch those operators warming up their amplifiers (and their fingers!) in the hours immediately preceding the start of the competition.
Sure enough, those stations were exactly there, as I expected them to be. In one hour I made 9 contacts, all Europeans (plus Russia). Another fun day at the river! Here is the video:
The Pico Balloon concept is simple: you build a transmitter small enough to be carried around by one or two common party balloons. As the transmitter floats around the globe it transmits its current position and altitude so you can follow its path in realtime. If you do it right, the balloon will circumnavigate the globe… repeatedly. Great stuff! I had to do it.
After realizing that it was relatively easy to construct a WSPR beacon based on Arduino and a clock generator, a few months back I decided to try building one.
The first attempt was on a breadboard with an Arduino Pro Mini, a uBlox breakout board and the Etherkit s5351a breakout board. After writing a few hacky lines of code (WSPR libraries are already available) WSPR worked! All I needed was making everything small enough to be carried by a party balloon.
I decided I was not going to use the standard Arduino Pro Mini. Rather, I was going to load the Arduino boot loader on a ATM328P myself and have everything else (GPS + Clock Generator) on a the same board. This unit also worked and, technically, was already small enough to be carried around by a couple of large balloons.
But how was I going to power it? Was a small LiPo battery suitable? After checking the specs I realized that even the smallest ones were too heavy. On top of that, batteries don’t take extremely low temperatures well so it definitely had to be solar power.
In the next revision, which I called 1.1, is where I faced the first serious obstacle. Until that point I was able to successfully load the Arduino boot loader without too much trouble using an ancient Arduino Uno as the ISP programmer. For reasons I still can’t understand, that wasn’t the case anymore with revision 1.1. Fortunately, when I was close to giving up, I remembered that back in the days I had purchased a proper ISP board and tried with that: Bingo! Problem solved. As you can see, on version 1.1 all necessary programming connections have been moved to a section of the PCB which is supposed to be cut out before launch (the wires soldered on the side were part of my desperate attempts to troubleshoot the ISP malfunction…).
At this point I had a transmitter that – together with the solar panels – weighted less than 10 grams! I really believed it was just about time to buy the balloons when I faced the second problem: I had grossly overestimated how much power solar cells can deliver. Yes, tensions can be pretty high but currents… well, currents are tiny: there is absolutely no way the beacon can be carried around by small balloons while being directly powered from solar panels as I originally envisioned. I needed something light that charged a capacitor and then – only then – I could start transmitting.
By searching things like “Tiny Solar Power Charger” I ended up on Jared’s (N7SMI) page. I was stunned: his pico balloon concept and hardware was identical to mine except that he completed it years in advance and, unlike mine, his actually worked and flew already many times. Jared had already solved the power issue with a 1.5F super capacitor charged by an SPV1040 controller and two 0.5V solar cells. Brilliant! Not only that. In order to save power, he provisioned his transmitter to selectively switch off the GPS and/or the clock generator. He also has a temperature sensor and – surprise surprise – a CW beacon too! Clearly, I now have enough material for version 1.2.
On part two I will publish my current schematics, PCBs and source code.