How to Make a Picoballoon




About: I am an eight grade student and I just love making cool (sometimes maybe even useful) devices.

What is a picoballoon and why would I want to build it?! I hear you ask. Let me explain. You all probably know what a HAB (High Altitude Balloon) is. It's a bunch of weird electronics stuff connected to a balloon. There are soooo many tutorials regarding HABs here on Instructables.

BUT, and that's a very big BUT what they don't tell you most times in the tutorial is the cost of the filling gas. Now, you can build a decent HAB tracker under 50€, but if it weighs 200g (which is a pretty optimistic guess with the batteries, cameras, etc.) the helium to fill the balloon can cost you 200€ or more, which is just too much for many makers like me.

So, as you can guess, picoballoons solve this problem by just not being bulky and heavy. Picoballoon is just a word for a light HAB. Light, what do I mean by light? In general, picoballoons are lighter than 20g. Now, just imagine that a processor, transmitter, a PCB, GPS, antennas, a solar panel and also a battery with a mass same as a disposable coffee cup or a spoon. Isn't that just insane?

Another reason (apart from the cost) why you would want to build this is its range and endurance. Classic HAB can fly for up to 4 hours and travel for up to 200km. A Picoballoon on the other hand, can fly for up to a couple of months and travel for up to tens of thousands of kilometres. One Polish guy got his picoballoon to fly around the globe multiple times. This of course also means that you will never see your Picoballoon again after launching it. That's why you want to transmitt all the data needed and of course keep the costs as low as possible.

Note: This project is a collaboration with MatejHantabal. Be sure to check out his profile too.

WARNING: This is a hard-to-do advanced level but also very fun project. Everything from PCB design to SMD to soldering will be explained here. That said, let's get to work.

UPDATE: We had to remove the GPS module last minute because of its big power consumption. It probably can be fixed but we didn't have the time for that. I'll leave it in the instructable but beware that it's untested. You can still get location from TTN metadata so you shouldn't worry about that.

Step 1: The Principle

So, when building a device like this, there are many variations and choices but every tracker needs a transmitter and a power supply. Most of the trackers will likely include these components:

- a solar panel

- a battery (lipo or supercapacitor)

- a processor/microcontroller

- a GPS module

- a sensor/s (temperature, humidity, pressure, UV, solar radiation...)

- a transmitter (433MHz, LoRa, WSPR, APRS, LoRaWAN, Iridium)

As you can see, there are many sensors and transmitters that you can use. What sensors you use is up to you. It doesn't really matter but most common are the sensors of temperature and pressure. Selecting a transmitter is much more difficult though. Every technology has some pros and cons. I won't break it down here because that'd be a very long discussion. What's important is that I chose LoRaWAN and I think that it's the best (because I didn't have a chance to test the others yet). I know that LoRaWAN has probably the best coverage though. You're welcome to correct me in the comments.

Step 2: Needed Parts

So, you'll need these things for this project:

Adafruit Feather 32u4 RFM95

Ublox MAX M8Q (We didn't use this at the end)

BME280 temperature/humidity/pressure sensor

2xSupercapacitor 4.7F 2.7V

Solar panel with output 5V

Custom PCBs

If are launching by yourself, you also need this:

At least 0.1m3 of helium (search: "helium tank for 15 balloons") buyed locally

Qualatex 36" self-sealing foil balloon

Estimated project cost: 80€ (only the tracker) / 100€ (including balloon and helium)

Step 3: Recommended Tools

These tools could come in handy:

wire stripper

soldering iron

SMD soldering iron



glue gun



hot air gun

You'll also need soldering paste.

Step 4: Adafruit Feather 32U4

We had a hard time selecting the right microcontroller for the balloon. The Adafruit Feather turned out the best for the job. It fits all the required criteria:

1) It has all the necessary pins: SDA/SCL, RX/TX, digital, analog

2) It has the RFM95 LoRa transmitter.

3) It's lightweight. It's mass is only 5.5g.

4) It has very low power consumption while in sleep mode (only 30uA).

Because of this, we think that the Adafruit Feather is the best microcontroller for the job.

Step 5: PCB Design and Manufacturing

I'm truly sorry for what I'm going to tell you. We're going to need to make a custom PCB. It's going to be hard and frustrating, but it's necessary, so let's get started. Also, to understand the following text properly, you should read this awesome PCB design class by Instructables.

So, at first you will need to make a schematic. I made both the schematic and the board in EAGLE PCB design software by Autodesk. It's free, so download it!

It was my first time designing a PCB and I can tell you that it's all about getting hang of the Eagle interface. I designed my first board in 6 hours, but my second board took me less than an hour. Here is the result. A pretty nice schematic and a board I would say.

When you have the board file ready, you need to create the gerber files and send them to the manufacturer. I ordered my boards from but you can choose any other manufacturer you like. I set the PCB thickness to 0.8mm instead of the standard 1.6mm because the board needs to be light. You can see my settings for JLC PCB in the screenshot.

If you don't want to download Eagle, you can just download "Ferdinand" and upload it to JLC PCB.

When you order the PCBs, just sit down comfortably in your chair and wait two weeks for them to arrive. Then we can continue.

Note: You can notice that the schematic is a bit different from the actual board. That's because I noticed that the bare BME280 IC is too hard to solder so I changed the schematic for a breakout.

Step 6: SMD Soldering

Another sad announcment: SMD soldering is not easy. Now really, it's fricking hard. May the lord be with you. But this tutorial should help. You can solder either using a soldering iron and a solder wick, or a soldering paste and a hot air gun. Neither of these methods was convenient enough for me. But you should get it done within an hour.

Place the components either according to the silkscreen on the PCB or according to the schematic.

Step 7: Soldering

After the SMD soldering is done, the rest of the solder job is basically a piece of cake. Almost. You've probably soldered before and I hope you'll want to solder again. You just need to solder the Adafruit Feather, antenas, the solar panel and the supercapacitors. Pretty straightforward I would say.

Place the components either according to the silkscreen on the PCB or according to the schematic.

Step 8: Complete Tracker

This is how the complete tracker should look. Weird. Nice. Interesting. Those are the words that come to my mind right away. Now you just need to flash the code and test if it's working.

Step 9: TTN Setup

The Things Network is a global city centered community LoRaWAN network. With more than 6887 gateways (receivers) up and running it's the biggest global IoT network in the world. It uses the LoRa (Long Range) communication protocol which is generaly at frequencies 868 (Europe, Russia) or at 915MHz (USA, India). It's most widely used by IoT devices sending short messages in cities. You can only send up to 51 bytes, but you can easily get a range from 2km to 15km. That is ideal for simple sensors or another IoT devices. And best of all, it's free.

Now, 2-15 certainly isn't enough, but if you get to higher ground, you should have a better connection. And our balloon will be very high. At 10km above sealevel, we should get a connection from 100km. A friend launched a HAB with LoRa 31km up in the air and he got a ping 450km away. So, that's pretty reasonable.

Setting up the TTN should be easy. You just need to create an account with your email and then you need to register the device. At first, you have to create an application. An application is the whole project homepage. From here you can change the decoder code, view the incoming data and add/remove devices. Just pick a name and you're ready to go. After that is done, you'll have to register a device in the application. You need to input the MAC adress of the Adafruit Feather (with the Feather in the packaging). Then you should set the activation method to ABP and you should disable frame counter checks. Your device should now be registered in the application. Copy the Device Adress, the Network Session Key and the App Session key. You'll need them in the next step.

For a more wholesome explanation, visit this tutorial.

Step 10: Coding

The Adafruit Feather 32U4 has an ATmega32U4 AVR processor. That means that it doesn't have a separate chip for USB communication (as Arduino UNO), the chip is included in the processor. That means that uploading to Adafruit Feather can be a bit harder compared to a typical Arduino board, but it works with Arduino IDE so if you follow this tutorial it should be fine.

After you have setup the Arduino IDE and sucessfully uploaded the "blink" sketch you can move onto the actual code. Download "LoRa_Test.ino". Change the Device Adress, the Network Session Key and the App Session key accordingly. Upload the sketch. Go outside. Point the antenna to the city centre or in the direction of the nearest gateway. You should now see data popping up on the TTN console. If not, comment below. I don't want to put everything that could've happened here, I don't know if the Instructables server could handle such ammount of text.

Moving on. If the previous sketch works, you can download "Ferdinand_1.0.ino" and change the things you were supposed to change in the previous sketch. Now test it again.

If you're getting some random HEX data on the TTN console, don't worry, it's supposed to do that. All the values are encoded in HEX. You're going to need a different decoder code. Download "decoder.txt". Copy it's contents. Now go to TTN console. Go to your application/payload formats/decoder. Now remove the original decoder code and paste in yours. You should now see all the readings there.

Step 11: Testing

Now this should be the longest part of the project. Testing. Testing in all kinds of conditions. In extreme heat, stress and with a strong light (or outside on the sun) to mimick the conditions up there. This should take at least a week so there will be no surprises in terms of the tracker behavior. But that is an ideal world and we didn't have that time because the tracker was built for a competition. We did some last minute changes (literally like 40 minutes before launch) so we didn't know what to expect. That is noot good. But you know, we still won the competition.

You'll probably need to do this part outside because the sun isn't shining inside and because the LoRa won't have the best reception in your office.

Step 12: Some Funky Formulas

Picoballoons are very sensitive. You can't just fill them with helium and launch them. They really don't like that. Let me explain. If the buoyant force is too low the balloon will not rise up (obviously). BUT, and this is the catch, if the buoyant force is too high, the balloon will fly too high, the forces on the balloon will be too big and it will pop and fall on the ground. That is the main reason why you really want to do these calculations.

If you know physics a bit, you shouldn't have a problem understanding the formulas above. There are some variables that you need to input into the formula. This includes: filling gas constant, thermodynamic temperature, pressure, mass of the probe and mass of the balloon. If you follow this tutorial and use the same balloon (Qualatex microfoil 36") and the same filling gas (helium) the only thing that will actually differ is the mass of the probe.

These formulas should then give you: the volume of helium needed to fill up the balloon, the speed at which the balloon rises, the altitude at which the balloon flies and also the free lift weight. These are all very useful values. The rising speed is important so the balloon doesn't hit obstacles because it's too slow and it's really nice to know how high will the balloon fly. But the most important of them is probably the free lift. The free lift is required when you will be filling the balloon in step 14.

Thanks to TomasTT7 for help with the formulas. Check out his blog here.

Step 13: Risks

So, your tracker works. That piece of shit you've worked on for two months actually works! Congratulations.

So let's review what risks your probe child can encounter up in the air:

1) There will not be enough sunlight hitting the solar panel. The supercapacitors will drain. The probe will stop working.

2) The probe will get out of range and no data will be received.

3) Strong wind gusts will destroy the probe.

4) The probe will pass through a storm during ascent and rain will short the circuit.

5) An ice coating will form on the solar panel. The supercapacitors will drain. The probe will stop working.

6) Part of the probe will break under mechanical stress.

7) Part of the probe will break under extreme heat and pressure conditions.

8) An electrostatic charge will form between the balloon and air forming a spark, which will damage the probe.

9) The probe will get hit by lightning.

10) The probe will get hit by a plane.

11) The probe will be get hit by a bird.

12) Aliens will hijack your probe. Can happen especially if the balloon will be above area 51.

Step 14: Launch!

So, that's it. It's the D-day and you're gonna launch your beloved picoballoon. It's always good to know the terrain and all possible obstacles. Also you must monitor the weather (mainly the wind speed and direction) constantly. That way, you minimize the chances of your 100€ worth of equipment and 2 months of your time hitting into a tree or into a wall. That'd be sad.

Insert a pipe in the balloon. Tie the balloon to something heavy with nylon. Put the heavy thing on a scale. Reset the scale. Secure the other end of the pipe on your helium tank. Start slowly opening the valve. You should now see negative numbers on the scale. Now is the time to use the free lift value you calculated in step 12. Shut off the valve when the negative number reaches the mass of the balloon + free lift. In my case it was 15g + 2.4g so I shut off the valve at exactly -17.4g on the scale. Remove the pipe. The balloon is self-sealing, it should seal automatically. Untie the heavy object and replace it with the probe. You're now ready to launch.

Just watch the video for all the details.

Step 15: Receiving the Data

Ohh, I remember the feeling we had after the launch. The stress, the frustration, a lot of hormones. Will it work? Will our work be worthless? Did we just spend so much money on something that doesn't work? These are the kind of questions we were asking ourselves after the launch.

Luckily, the probe responded about 20 minutes after the launch. And then we received a packet every 10 minutes. We lost contact with the probe at 17:51:09 GMT. It could have been better, but it's still fine.

Step 16: Further Plans

This was one of our most difficult projects up to date. Not everything was perfect but that's OK, it's always like that. It was still very successful. The tracker worked flawlessly. It could've done that for a lot longer but that doesn't matter. And, we ended up second in the Picoballoon competition. Now you might say that being second in a contest with 17 people is not such a success BUT keep in mind that this an adult engineering/construction contest. We're 14 years old. The ones we were competing with were adults with engineering and possibly even aerospace background and with much more experience. So yeah, overall, I would say that it was a big success. We got 200€, which was approximately double of our expenses.

I'm certainly going to build a version 2.0. It's going to be much better, with smaller components (barebone processor, RFM95) and it's going to be more reliable so stay tuned for the next instructable.

Our main goal now is to win the Epilog X contest. Fellow makers, if you liked this instructable, please consider voting for it. It would really help us. Thank you soooooo much!

Epilog X Contest

Runner Up in the
Epilog X Contest



    • Tape Contest

      Tape Contest
    • Trash to Treasure

      Trash to Treasure
    • Arduino Contest 2019

      Arduino Contest 2019

    14 Discussions

    Michal Choma

    21 days ago

    and good cover (as transmitter) can be using Sigfox. Dont know, how it works in high altitude, but it has good cover at mountains, nature, far from cities.

    1 reply

    Reply 21 days ago

    Thanks. Even the Adafruit board is what I would call an overkill. It has too many unnecessary pins. The Lopy board has both wi-fi and bluetooth. That does not only add weight but also power consumption. So it is possible but not exactly the best option. From the specification I think that the range should be simmilar. Also, range on ground does not equal the range up in the air.

    Michal Choma

    21 days ago

    One more question. It is chance to find electronics? I mean, if ballon will be catched by some tree it can be find. Also it will be dangerous for electric cables (ehm, high voltage wires.)

    1 reply
    Jakub_NagyMichal Choma

    Reply 21 days ago

    There is a chance. But if the flight takes longer than a couple of weeks the chances are that it will fall in the water. Yeah, that's exactly what's written on the balloon by the manufacturer. It could be dangerous for electric cables. But I wouldn't worry about that especially not regarding the descend.


    22 days ago

    You picked an awesome day to launch this project - maybe it's just a coincidence, but today is the first full business day that the UK was supposed to no longer be part of the EU. It turns out we're still part of the UK: but who knows what will happen.

    It has been suggested that if or when the UK does leave the EU, our internet will be cut off and we will no longer be able to send or receive email, ordinary mail, or telephone calls to Europe. The European Space Agency has also threatened to block satellite transmissions so that British people will not be able to watch the Eurovision Song Contest.

    Now, tell me if I am wrong, but surely a whole array of your little PicoBalloons would be able to form a communications network to enable us poor Britons to stay connected to mainland Europe in some way? Even if that was not possible, you could send up PicoBalloons with little cards attached to them with kind messages like, "Thinking of you, love Czech Republic" or "We told you so".

    I think there is tremendous potential here. Perhaps a crowdfunder could be set up to allow you to produce these on a commercial scale. Some funding would also also allow you to hire someone who is good at SMD soldering - I understand from your instructable that you find this part of the build rather challenging.

    Anyhoo, thanks again for an amazing project. I'm not sure if this message will even get through as our internets are starting to get shut down.

    1 reply

    Reply 22 days ago

    Interesting. I didn't know that.

    I don't think that it is really possible to just cut off communication for 60 million people. You have satelites, underwater cables, radio towers and amateur radio towers. So it'd be hard to destroy such a large infrastructure. But if someone did, that would probably classify as a war act possibly starting a war so the Eurovision Contest would be the least of your worries.

    It isn't really possible to send picoballoons from Czech Republic to Britain. This is caused by the planetary wind currents. No matter the weather, the wind in Europe flows more or less to the east.

    An array of picoballoons could be set up. They would have to be anchored though so they don't fly away. Something like this is already in research to be used mainly in developing countries that don't have internet access. Also, solar powered drone arrays are an considered option.

    Lastly since internet is a network of computers, the hard work wouldn't really make sense in the end. In order for this system to work, EU would have to collaborate. If you want to connect with mainland, mainland must want to connect with you. Even if the UK was cut off, internal UK communication systems would still work. Websites with servers in UK would be accessible.

    Your message got through :).

    Michal Choma

    22 days ago

    Zdravim, chcel by som sa spýtať, či ste riešili nejako s letiskom to, či môžete púštat balon do vzduchu? Netreba nato povolenia? Inak velmi sa mi páči vaš projekt a že som tu stretol Slovákov :D Hello, I would like to ask if you have somehow dealt with the airport if you can let the balloon go into the air? No need for permission? Otherwise, I really like your project.

    1 reply
    Jakub_NagyMichal Choma

    Reply 22 days ago

    Super. Ďakujem. Pokračujem po anglicky lebo je to tu medzinárodná komunita.

    We didn't have to deal with that. The balloon is so small that it doesn't mean any harm to a large aircraft. Everything with the balloon and probe weight less than 100 grams. That's the size of a larger party balloon. If a kid accidentally launches a party balloon he doesn't get a penalty. You absolutely don't need any permission at all. I don't know how it works in USA but I think it is the same.


    23 days ago

    Very well done! I especially like that you provided your reasons for why you selected various parts and technologies. You have my vote!

    1 reply

    Reply 23 days ago

    Thanks! Yeah I tried explaining everything in a more relatable and less technical way but I also included the specific info.