The Payload Assemblage

The new components required new arrangement. The HTC was left out this time, but I had a cable with the transmitter sticking out in the middle of the box and a handful of wires and electronics to layout in a serviceable way. The main material I used was polystyrene and the masking tape.

I began by underlaying the NTX2 by a piece of polystyrene. With the antenna pushed inside the incisions, the cable with the transmitter would stick a few centimeters above the box's floor.

The piece would nicely secure the datalogger as well with the probe sticking out of the box about 5 centimeters.

For the batteries, I prepared a whole casing to make them secure and not flying around the box.

The battery box went to one side and another piece of polystyrene, this time underlaying the transmitter from the other side, would secure it in place. Another piece of polystyrene topping at the same height as the battery box covered the datalogger.

Then I added two more pieces of polystyrene to congest the area a little bit more. One went to separate the datalogger from the remaining electronics and the other was placed to push against the previous one and the camera's frame to secure it all in place. This arrangement created a small convenient room for all the electronics. All the heat generated by them would thus be concentrated in this small area. Moreover, the datalogger alongside which some cold air could get in was separated.

All the electronics were connected. At this stage during the actual balloon flight, I would have to turn on the camera and then finish sealing the payload, because I wouldn't have access to the camera later. This meant a few minutes of camera time wasted, so it would be convenient to do it as fast as possible.

And here, all the electronics are in place. Notice that I separated the transmitter from the GPS module and the Arduino by yet another thin polystyrene piece. This was to make sure there wouldn't be any unintended current shortages. I also cut out a little space for the GPS's antenna in one of the polystyrene blocks. To secure the Arduino in place, I used a few of its own pins and simply pierced them through the polystyrene. I also turned the Arduino to face me with the power and ground pins, so they were easily accessible when connecting the battery wires. The camera got covered with a piece of foam, and the remaining block of polystyrene was used to enclose this space from above. After this stage, the payload only waited to be covered by the two main lids.

With the new parts, the whole set had to be re-weighted. After a couple more measures, the final weight settled on 940g. The HTC took some weight off, but primarily the battery pack added it all back.

The time for the last test had come.

I practiced the whole assembly including turning on all the electronics. Then I hung the payload on a nearby tree and periodically checked the signal reception. All went fine for a sufficiently long time, so my payload was finally ready to go.

The Batteries

Since the payload would be exposed to temperatures as low as -50°C, perhaps even -60°C, it was worthwhile to think about the impact on the batteries. Especially because I wasn't completely sure about the payload's capabilities to retain such freezing temperatures. In other HAB projects, I had seen them use Energizer's Ultimate Lithium AA batteries. They are somewhat more expensive, but the manufacturer states their operating temperature as low as -40°C. The specification also states their ability to maintain the capacity of 3000mAh under such conditions provided a constant current discharge of 25mA. All in all, the batteries should allow +10 hours functionality of the transmitter.

The Antennas

I have already described the actual receiving and transmitting, but I haven't mentioned a very important part of the chain without which my solution would work only at little distance. The antennas. Both the radio and the transmitter were equipped with one. As usual, not knowing anything about the subject matter, I turned to Google and a physics textbook. After some time of researching, I had an idea about what sort of an antenna I wanted to use for each device. I started with the one for the transmitter.

This antenna design came from a HAB project I encountered on the internet. It was a ground plane antenna made of just a coaxial cable and a few pins and straws. I had all the components at hand so I gave it a try.

First thing to do was to strip of the plastic cover off the coaxial cable. The cable was 50Ω RG58U. I chose this one because the NTX2's output is 50Ω as well. The length of the stripped part was 17.25cm, a quarter of the wave length at 434.65MHz

Then I uncovered the other end as well and twisted up the RG58U's copper shielding into two cords. The NTX2 had three antenna pins. Two for GND and one for signal output. I fitted the cords with KONPC-SPK pins including the core wire hidden beneath the insulation.

On the antenna end of the cable, I twisted the copper shielding up into four cords this time.

I would fit each cord into a straw and tape them to a foam board to create two perpendicular lines. This would form the ground for the active element of the antenna. The core of the cable would stick perpendicularly to the ground.

After that the antenna was ready for a test. I attached it to the breadboard using pins.

And here is the set at work. The signal viewed using the SDR visibly increased or decreased whenever I either attached or disconnected the antenna from the board.

When I was building the previous antenna, I exchanged several messages with people in a Czech Citizen Band Radio forum who were fairly critical towards this design and used components especially. I, not knowing what was sufficient  for a HAB and what wasn't, decided to take into account their objections and construct a new antenna. This time, I would instead of a coaxial cable use plain copper wires. Also, I came across a neat piece of software that allowed me to design my antenna and provided me with a feedback on its final parameters. The reality is never as ideal as the theoretical model but at least I had a basis. The program was called MMANA-GAL and led me to the design in the image above. It was still a ground plane antenna, but this time, I bent the ground wires at an angle of 35° to the original ground. This would change the impedance of the antenna to 50Ω and match it to the NTX2's impedance.

The parameters of this antenna design.

The radiation characteristic of the antenna.

And the dimensions of the antenna. The length of the active element was changed from 17.25cm to 16.4cm for 434.65MHz frequency.

Since I would have to solder the coaxial cable leading from the antenna to the NTX2 and I would have to securely attach the impractically shaped antenna to the payload box, I decided to redesign the box a little. The antenna and the NTX2 would have to stay in one piece, but I wanted to be able to work with it even without the box. Thus I came up with a solution that would comprise of carving out a hole in the bottom of the box and inserting the coaxial cable inside a squared log of polystyrene. I would then be able to fit this into the hole in the box or take it out again.

I used three copper wires. One for the active element of 16.4cm length and two for the ground of 34.4cm lengths. These two, I would then bend at an angle of 110° inbetween the two ends. I would solder them together and then to the coaxial cable's unwound shielding. To the cable's exposed core, I would solder the short copper wire.

To secure the antenna in position, I carved out four insections in the polystyrene and pushed the wires inside. Though, I should mention that in the end I decided to carve out the insections in the corners, because when positioned like in the picture, the wires were in the way when I wanted to lie the box on its side.

When tested, the antenna worked just fine allowing me to receive the signal with my notebook and SDR without its antenna at the other side of the garden. Something unthinkable with just the NTX2 alone.

Now, it was time to build the receiving antenna. Since the NTX2's manual states its reach to about 500m and I had to count with distances of about +50km, I had to look for directional antennas with as high a gain as possible. The positive side to the HAB story is that since the balloons fly above us, there is nothing in the way for the signal to reach us. The UKHAS website states a record breaking reception distance of 800.3km. And this was achieved with the standard 10mW transmitters. After a traditional research, I decided for a 9 element Yagi antenna.

The parameters of my Yagi. These are sort of reverse engineered, because I didn't manage to cut the elements exactly as the scheme stated, so I made this model to see what exact parameters the program predicts for my actual antenna.

The radiation pattern. The gain of 11.29 dB compared with a dipole sounded good enough to me. 

And the dimensions of my actual antenna.

These were the dimensions for the individual elements as I got them from some Yagi design websites. Unfortunately, I made a terrible decision and bought 2.5 and 3mm steel sticks to use them for the elements as opposed to copper or anything less stiff and more compliant. The piece of paper also contains the proper spacing of the elements.

The list of components for the Yagi contained: two meters of Belden H121A coaxial cable of 75Ω impedance (to match the impedance of the DVB-T dongle), two terminal blocks able to fix a steel stick of 3mm in diameter, one male TV aerial plug, one long enough wooden squared log (2.5 x 2.5cm) and two 1m long  steel sticks 3mm in diameter and one 1m long stick 2.5mm in diameter. To clarify the choice of the material for the antenna elements, I have to say that at first I tried to get hold of some copper sticks, but failed to find any around my town. I came across suggestions to use welding sticks, but as I found out later, they sold them only in packages of 500 sticks. The only thing resembling my vision were these steel sticks sold in an RC model shop. Only later I found out how difficult it was to cut steel. One cut through the 3mm steel stick took me and my handsaw about ten minutes of unremitting sawing. The reason for one of the sticks being only 2.5mm in diameter was that they simply didn't have any more of those 3mm thick. This was the reason for me not being able to get the dimensions of the elements exactly right. The steel would often break off at the end of the sawing leaving crooked endings. Thus I ended up with most of the elements about 1mm shorter or longer.

First, I went about making the dipole element. That meant sawing off two 16.25+1cm long pieces. That one additional centimeter is for the bend at the end as illustrated in the picture above. Each piece would then go to one dual terminal block and would be electrically connected to the coaxial cable at the other side. The end of the coaxial cable was stripped of the plastic coat, the shielding was wound together and put inside one part of the terminal block, and the remaining core wire to the other.

The other end of the coaxial cable, I fitted with the male TV aerial plug. But since the DVB-T dongle had a female MCX connector, I had to find an MCX to aerial plug adapter. With that done, the dipole was ready.

The other elements were simpler. All I had to do was to saw off the right lengths of the individual elements and than just fit them in the dual terminal blocks.With all the elements prepared, I marked the right distances between them onto the wooden log and screwed them down. I also used one strap fastener to secure the cable, so it doesn't rip out of the dipole.

Here is the finished antenna. I only replaced the wooden squared log from the previous pictures with one of appropriate dimensions to seat all nine elements. The whole antenna measured a little over 130cm in length. The picture on the right shows one of the first tests. The transmitter powered by a notebook and the signal being decoded on a PC.

 

With all the antennas and the transmitter ready, it was time for the first proper long distance test. I placed the transmitter on the top of the roof and drove off to one of the hills around my town with a decent line of sight. I set up the antenna and the receiver. You can notice that I had equipped the antenna with a handy tripod. Just a few wooden logs knocked together. The distance to the transmitter was 4.6km and the reception worked without a problem. After successfully finishing this test, I used a layman's logic. If it works at the distance of 4.6km, it will work at the distance of 30-50km as well.

Coincidentally, just today, I noticed a high altitude balloon being released in Budapest. Finally one close enough. The highest number of balloons is released in England, well beyond the reach of my antenna. At about the time the balloon was passing the 26th kilometer in altitude, I took out my antenna and notebook, set it up in the garden and directed it roughly in the right direction. After a couple of moments of fiddling with the settings, I managed to receive and upload 2 telemetry strings right before the balloon's burst. Additionally, I received several more partial strings, but with a lot of errors. Nevertheless, it was my first reception of someone else's payload, and it verified the functionality of SDR and my homemade antenna at the distance of 263km. I only wonder whether the issue with the other telemetry strings was a problem of the notebook not being capable of coping with the overload of data, or if it was simply at the limit of my antenna's reception ability. I did have to aim it through a group of trees after all. I guess only more reception attempts and more work experience with the equipment may answer that question.

The Radio

At this point, we have a working transmitter sending out a signal of 1s and 0s at around 434.650MHz. The question now is how to capture and decode the signal. Amateur radio operators will definitely have an advantage here. I, not being one, was not willing to basically double the price of my project by buying a proper radio just for this purpose, and, unfortunately, none of the people I knew was an amateur radio enthusiast either. Thus I turned to Google once again. Luckily, I eventually came across something called Software Defined Radio.

SDR in general is, as I understand it, a process in which one uses a specific device with an antenna to capture a spectrum of radio frequencies with the demodulation that is otherwise done in the proper radio's hardware, taking place in one's computer's software. Apart from specific SDR intended devices, there are several models of DVB-T dongles that combined with the right software provide comparable functionality for a fraction of the price. The dongle I found on Aukro and ordered contained Realtek's RTL2832U chip and Elonics' E4000 tuner allowing you to tune into frequencies from 60MHz to 1900MHz. These are the important parts distinguishing SDR usable DVB-T dongle from an ordinary one.

The next step after the DVB-T dongle had arrived was to install the software. First thing I needed to do was to install the proper driver. This is done by using a program called Zadig. It lets you choose the right USB port and replaces the driver with the required one. Next piece of software I installed was SDRSharp. This program is the one that allows you to tune across the mediated frequency range. You just select the RTL-SDR/USB option in the left upper corner and then, depending on what you want to do, either WFM for listening to a regular radio station (the image above) or you choose the USB (upper side band) option for listening to your payload (the image below). For the software download links and a manual on how to set it all up visit: http://ukhas.org.uk/guides:sdr_tracker

The signal reception was solved. Next, I needed something to actually decode it. The program to do that is called DL-Fldigi, a modified version of the standard Fldigi software that besides decoding allows you to upload high altitude balloon telemetry data straight to the Habitat webpage that then displays the balloon data on a map. But first, I had to install VB-Cable software that establishes a virtual cable connection between SDRSharp and DL-Fldigi. In SDRSharp, I then selected the 'VB Audio Point' option in Audio/Output and in DL-Fldigi selected the same option in Configure/Sound Card/Audio/Devices/Capture. Now, when I turned the DVB-T dongle on and tuned into the frequency of my payload, a stream of data flowed into the DL-Fldigi.

To decode the data, I first went to Op Mode/RTTY/Custom and selected the options based on the properties of my signal. All I had to do then was to locate the cursor above the two frequencies of received data and click. With the 'Automatic Frequency Control' option selected, the program then stays locked onto the signal on its own even if the frequency drifts. That sometimes happens when the temperature of the transmitter changes. When it heats too much or when it freezes (at high altitudes for example). If there wasn't any problem, the decoded message should appear in the middle window.

This program contains a lot of settings that might give a newbie (like me) a hard time. On my PC, all I had to do to make things work was what I mentioned in the couple of  paragraphs above. But for the actual tracking, I wanted to use my notebook (Asus EEE pc 1000h) for its outstanding battery life. The only problem was, I couldn't get it run. I suspected the EEE's miserable computational power. The both programs would work very slow and the decoder would print out only nonsense characters instead of the telemetry data. A couple of sleepless nights later and some googling of radio terms, I made it work. To summarize for someone dealing with the same problem, I would recommend to fiddle with the following options. In SDRSharp I set the Sample Rate in the Configure option as low as possible and I switched both Automatic Gain Controls off. Instead of that, I operated the RF Gain manually based on the signal strength. I unselected the 'Use AGC' button in the left panel as well and finally I decreased the resolution in the FTT Display option to 2048-4096. The smallest value of 512 didn't work with the DL-Fldigi and delivered no decodable signal. I tried a lot of options in the DL-Fldigi as well, but when I did the aforementioned changes in SDRSharp, DL-Fldigi worked fine even in the standard setup.

The whole setup finally working fine on my notebook as well.

And a picture from the early stages with the signal too strong. Here, it would be useful to turn of the Automatic Gain Controls and adjust the RF Gain manually.