The first attempt at the second launch
The payload takes a snapshot of the launch crew inflating the balloon for the first
attempt at a second launch, as I struggled to fix a troublesome GPS unit
attempt at a second launch, as I struggled to fix a troublesome GPS unit
The first attempt to launch my second high altitude balloon took place on September 5, 2011. I was back at Purdue and managed to recruit a couple AAE undergrads to help out. I was desperately trying to take advantage of the 3-day labor day weekend to launch, since I knew the onset of classes would make it difficult for all of us to find time to work on the project.
An initial launch attempt for Sunday the 4th had been moved back to Monday the 5th in order to allow for more time to prep the payload. Good thing too, on the 4th the Inventek GPS finally died. Perhaps it was because I never reinstalled that inductor leading up to the antenna, and simply jumped the connection. In any case, it was dead.
I spent the entire day integrating the Garmin GPS into the system as well as doing everything else necessary to prep for launch. Unfortunately, the Garmin can be unreliable. Sometimes it throws characters into sentences that don't belong there, and ruins the entire checksum. I was was fully aware of the problem, but given that more than 50% of sentences passed checksum, I deemed it to be an acceptable issue.
The next morning, things changed. Something like 29 in 30 sentences would fail checksum. Very occasionally a good sentence would get through and we would get data, but this was maybe once a minute, which meant that if we missed one sentence due to garbled radio transmissions we would have to wait potentially 2 minutes or longer for the next valid sentence. I decided this was too risky, and called off the launch, despite the fact that we had already inflated the balloon. This led to an important change in the preflight checklist: make sure critical electronics are working BEFORE filling the balloon!
Today's experience vindicated that call - as we got transmissions from the GPS, occasionally one would miss a character, especially as the altitude increased, and missing one single character is all you need to lose an entire sentence's worth of data. Given the other problems we had, if we had been getting data only once a minute (at best) locating the balloon with GPS transmissions alone might not have been possible.
An initial launch attempt for Sunday the 4th had been moved back to Monday the 5th in order to allow for more time to prep the payload. Good thing too, on the 4th the Inventek GPS finally died. Perhaps it was because I never reinstalled that inductor leading up to the antenna, and simply jumped the connection. In any case, it was dead.
I spent the entire day integrating the Garmin GPS into the system as well as doing everything else necessary to prep for launch. Unfortunately, the Garmin can be unreliable. Sometimes it throws characters into sentences that don't belong there, and ruins the entire checksum. I was was fully aware of the problem, but given that more than 50% of sentences passed checksum, I deemed it to be an acceptable issue.
The next morning, things changed. Something like 29 in 30 sentences would fail checksum. Very occasionally a good sentence would get through and we would get data, but this was maybe once a minute, which meant that if we missed one sentence due to garbled radio transmissions we would have to wait potentially 2 minutes or longer for the next valid sentence. I decided this was too risky, and called off the launch, despite the fact that we had already inflated the balloon. This led to an important change in the preflight checklist: make sure critical electronics are working BEFORE filling the balloon!
Today's experience vindicated that call - as we got transmissions from the GPS, occasionally one would miss a character, especially as the altitude increased, and missing one single character is all you need to lose an entire sentence's worth of data. Given the other problems we had, if we had been getting data only once a minute (at best) locating the balloon with GPS transmissions alone might not have been possible.
The Prelude - Will the weather hold?
About 6 weeks later, after having successfully integrated the new GPS (a Lassen IQ, which had it's own set of problems but which functioned beautifully for the launch), I decided it was time for another launch attempt. I was still a bit hesitant to launch, since we would be risking more equipment and even more hard work this time around, but if you don't launch you don't learn. And you learn a lot more by launching than you ever could with just ground testing.
The weather for Sunday, October 23, did not look good throughout the preceding week. I informed my crew that we would be attempting again on the 23rd, but also told them not to expect to actually launch. Poor weather forecast was unfortunate, but I saw no reason not to continue preparations for launch, since it would still be a good exercise.
On Friday before the launch I went to buy hydrogen gas from the local welding supply. After discussing it thoroughly with the team, we decided we could safely handle this potentially dangerous gas, and chose it over helium due to it's greater lift capacity (meaning shorter flight time meaning less distance traveled) and lower cost.
The weather for Sunday, October 23, did not look good throughout the preceding week. I informed my crew that we would be attempting again on the 23rd, but also told them not to expect to actually launch. Poor weather forecast was unfortunate, but I saw no reason not to continue preparations for launch, since it would still be a good exercise.
On Friday before the launch I went to buy hydrogen gas from the local welding supply. After discussing it thoroughly with the team, we decided we could safely handle this potentially dangerous gas, and chose it over helium due to it's greater lift capacity (meaning shorter flight time meaning less distance traveled) and lower cost.
A note on hydrogen as a lifting gas
A NOTE OF CAUTION to those considering hydrogen for their own balloon launches (those who aren't considering handling hydrogen may happily skip this paragraph): Hydrogen is a dangerous and highly flammable gas so be sure to take appropriate safety measures when handling it. Be aware that hydrogen gas will leak through latex, albeit slowly and generally not in large enough quantities to achieve a flammable mixture, but remember that even though the connection between your balloon and the gas cylinder may be secure, doesn't mean there's absolutely no hydrogen floating around nearby. Be sure to inflate in large, well ventilated areas. Due to extremely calm winds we simply inflated outside, though with any sizeable wind this becomes difficult. Choose your launch location wisely.
The Launch
On Sunday morning the weather at the launch site was overcast, but clearing up, so we decided to wait for clearer skies before launching. We used the extra time to run a few more predictions and finally set our target burst altitude at 30.5 km, which would require nearly 3.5 cubic meters of hydrogen.
As we were waiting for the rest of the crew to show up, it started to drizzle, which immediately put me on edge. It started to intensify slightly, but the clouds were still moving away so we held. Eventually, the drizzling stopped, and the sky was clearer than ever.
First order of business was to verify all electronics. Unfortunately, the same computer I had used for a battery test the night before was no longer working - it wasn't getting data from the radio. I suspected it was an issue with the audio cable, as it had given me trouble before, and we decided that someone had to drive to Radioshack immediately and get a new one. So we did, and the same problem happened, but if the pulled the audio cable out just a little bit, it worked fine. This seemed to be an acceptable solution to an unfortunate problem, and so we pressed on.
The next step was filling the balloon. Since we were using hydrogen this time, we were very careful. We researched the issue beforehand and came up with some basic ground rules for filling the balloon. One of which included going over and touching a grounding rod every few minutes. Thankfully no incidents occurred.
As we were waiting for the rest of the crew to show up, it started to drizzle, which immediately put me on edge. It started to intensify slightly, but the clouds were still moving away so we held. Eventually, the drizzling stopped, and the sky was clearer than ever.
First order of business was to verify all electronics. Unfortunately, the same computer I had used for a battery test the night before was no longer working - it wasn't getting data from the radio. I suspected it was an issue with the audio cable, as it had given me trouble before, and we decided that someone had to drive to Radioshack immediately and get a new one. So we did, and the same problem happened, but if the pulled the audio cable out just a little bit, it worked fine. This seemed to be an acceptable solution to an unfortunate problem, and so we pressed on.
The next step was filling the balloon. Since we were using hydrogen this time, we were very careful. We researched the issue beforehand and came up with some basic ground rules for filling the balloon. One of which included going over and touching a grounding rod every few minutes. Thankfully no incidents occurred.
From left: Myself, Jeremy, and Walter fill the balloon. The grounding rod can be seen in the foreground, and the colorful object on the ground is nothing more than a box of zipties.
Slowly, but surely, everything was coming together for launch. Once the balloon was full, I called the airport tower and advised them we would be launching our balloon in the next half hour. They gave us a green light, and I proceeded to activate all electronics in the payload. Finally, we were ready to go for another attempt:
The Chase
I watched the balloon as it floated away. The team yelled at me to get in the car, but I said "Hold on, we don't have to go yet, I can still see it" My best memory from the first launch was watching the balloon go up, so I wanted to solidify this moment.
Finally I got in the car and we drove off to chase the balloon. I must admit, it was getting away from us faster than I had anticipated, and dealing with local traffic and traffic lights made the chase slower than would be ideal. For the first part, everything went smoothly enough. I was calling out the altitude, which seemed to be rising at an incredible rate. Between the time I posted to twitter about the launch and the time I posted the same information to facebook, it had risen 500m!
As we continued on our course pretty much due east, we saw some signs about road construction, but continued through thinking we could maybe hit the next country road and get off. Unfortunately, we hit a closed road, and had to double back. By this point we still weren't caught up with the balloon, which made the doubling back a pain, but it wasn't really something we could have avoided. The new route we were forced to take was very scenic, which meant lots of radio interference.
At this point Jeremy, who had been responsible for launching the balloon, asked to plug his laptop into the power inverter. I was a bit concerned since my laptop and the radio were already plugged in to two out of three outlets, but others on the team didn't seem to express and concern, and his laptop was getting a better internet connection than mine, which we needed in order to be able to plot the payload onto a map.
I was still getting interference from the trees, or maybe it was something else, who knows, but I opened the window to try to stick the antenna out a bit higher to get a better view. At this point we had already switched from the 1/4 wave ground plane antenna to the moxon. The speed of the Jeep made this more difficult than I anticipated, so I quickly brought the antenna back in and closed the window.
Then I looked back onto my computer screen. Nothing. Not even a trace of signal. I immediately unplugged the audio cable from the radio to try to listen to the data qualitatively, but nothing came out. Then I noticed the radio was dead. I quickly looked down at the power inverter, it was dead too. We had a problem.
We took out both laptops from the inverter, but still nothing. I fiddled with the connector, even unplugged it for a few minutes to let it cool down. Nothing. It was burnt. Quickly I remembered reading stilldavid's balloon launch story, and how they hunted for "landline" power outlets in regular stores. Clearly that was what we had to do. We stopped at a McDonalds along the way, and struggled to get everything out of the car. Keep in mind, the computer is connected by the audio cable to the radio, which is connected to the antenna and to the power supply, which is connected to the now-dead inverter. I grabbed the laptop+radio+power supply and handed off the antenna, and walked into McDonalds with this odd array of computers and radios and wires. I searched around frantically for a power outlet, seeing none I quickly asked the employees and they responded there were none. Damn.
We ran back outside, and saw the Starbucks across the street. They HAD to have power outlets. We crossed the 4-lane street, dodging traffic and made it inside the Starbucks. While we were running, Graham and Walter were trying to unscrew the power inverter to try to find a blown fuse we could replace.
Hunting around for a power outlet in Starbucks I quickly found one by the window. I set up the power supply, radio, computer and turned everything on, but left the audio cable unplugged so as to listen for the data. I started waving the antenna around by the window, got something, and plugged the audio cable back in. A couple successful transmissions later, we saw that the balloon was actually behind us, and still climbing at ~24-25km. Meanwhile, Walter and Graham had found that the power inverter did not even have a fuse! We made the call to capitalize on our being ahead of the balloon and go to the Lowe's across the street to buy a new power inverter. The cost might be high, but there wasn't much option at this point.
We packed up and started to scramble back to the car when one of the employees asked us what we were doing. In rush, I quickly explained we were tracking a high altitude balloon, and rattled off the tracker's website. After that odd plug we immediately ran out the door to the car. I was too concerned with our tracking problems to really look around, but the other team members assure me we got some quizzical looks as we blasted serial data through the radio and waved antennas around like crazy. Although apparently some people were not particularly concerned with us, it must have been an amusing sight. Especially since we left just as quickly as we had come.
We got back to the car and headed over to Lowes. I ran in, explained to the greeter that I urgently needed a power inverter. Turns out Lowes doesn't carry power inverters! I ran back outside while attempting to locate a nearby Radioshack. I found one just a couple miles from our present location, and gave them a call. I hurriedly asked about power inverters, and they confirmed that they did indeed carry them. After some minor confusion as to whether it was on the left or the right (confusion which led us to narrowly miss a green and get stuck at a red for what seemed like ages!), we got to the Radioshack, I ran inside and got a $40 inverter with only one outlet (and a fuse on the front mind you).
I ran back outside, and the rest of the team had prepared all the equipment for me to be able to plug-n-play. With everything on, we pointed the antenna in the right direction and started listening. The following two sentences appeared on fl-digi's screen (in parentheses are my comments):
Finally I got in the car and we drove off to chase the balloon. I must admit, it was getting away from us faster than I had anticipated, and dealing with local traffic and traffic lights made the chase slower than would be ideal. For the first part, everything went smoothly enough. I was calling out the altitude, which seemed to be rising at an incredible rate. Between the time I posted to twitter about the launch and the time I posted the same information to facebook, it had risen 500m!
As we continued on our course pretty much due east, we saw some signs about road construction, but continued through thinking we could maybe hit the next country road and get off. Unfortunately, we hit a closed road, and had to double back. By this point we still weren't caught up with the balloon, which made the doubling back a pain, but it wasn't really something we could have avoided. The new route we were forced to take was very scenic, which meant lots of radio interference.
At this point Jeremy, who had been responsible for launching the balloon, asked to plug his laptop into the power inverter. I was a bit concerned since my laptop and the radio were already plugged in to two out of three outlets, but others on the team didn't seem to express and concern, and his laptop was getting a better internet connection than mine, which we needed in order to be able to plot the payload onto a map.
I was still getting interference from the trees, or maybe it was something else, who knows, but I opened the window to try to stick the antenna out a bit higher to get a better view. At this point we had already switched from the 1/4 wave ground plane antenna to the moxon. The speed of the Jeep made this more difficult than I anticipated, so I quickly brought the antenna back in and closed the window.
Then I looked back onto my computer screen. Nothing. Not even a trace of signal. I immediately unplugged the audio cable from the radio to try to listen to the data qualitatively, but nothing came out. Then I noticed the radio was dead. I quickly looked down at the power inverter, it was dead too. We had a problem.
We took out both laptops from the inverter, but still nothing. I fiddled with the connector, even unplugged it for a few minutes to let it cool down. Nothing. It was burnt. Quickly I remembered reading stilldavid's balloon launch story, and how they hunted for "landline" power outlets in regular stores. Clearly that was what we had to do. We stopped at a McDonalds along the way, and struggled to get everything out of the car. Keep in mind, the computer is connected by the audio cable to the radio, which is connected to the antenna and to the power supply, which is connected to the now-dead inverter. I grabbed the laptop+radio+power supply and handed off the antenna, and walked into McDonalds with this odd array of computers and radios and wires. I searched around frantically for a power outlet, seeing none I quickly asked the employees and they responded there were none. Damn.
We ran back outside, and saw the Starbucks across the street. They HAD to have power outlets. We crossed the 4-lane street, dodging traffic and made it inside the Starbucks. While we were running, Graham and Walter were trying to unscrew the power inverter to try to find a blown fuse we could replace.
Hunting around for a power outlet in Starbucks I quickly found one by the window. I set up the power supply, radio, computer and turned everything on, but left the audio cable unplugged so as to listen for the data. I started waving the antenna around by the window, got something, and plugged the audio cable back in. A couple successful transmissions later, we saw that the balloon was actually behind us, and still climbing at ~24-25km. Meanwhile, Walter and Graham had found that the power inverter did not even have a fuse! We made the call to capitalize on our being ahead of the balloon and go to the Lowe's across the street to buy a new power inverter. The cost might be high, but there wasn't much option at this point.
We packed up and started to scramble back to the car when one of the employees asked us what we were doing. In rush, I quickly explained we were tracking a high altitude balloon, and rattled off the tracker's website. After that odd plug we immediately ran out the door to the car. I was too concerned with our tracking problems to really look around, but the other team members assure me we got some quizzical looks as we blasted serial data through the radio and waved antennas around like crazy. Although apparently some people were not particularly concerned with us, it must have been an amusing sight. Especially since we left just as quickly as we had come.
We got back to the car and headed over to Lowes. I ran in, explained to the greeter that I urgently needed a power inverter. Turns out Lowes doesn't carry power inverters! I ran back outside while attempting to locate a nearby Radioshack. I found one just a couple miles from our present location, and gave them a call. I hurriedly asked about power inverters, and they confirmed that they did indeed carry them. After some minor confusion as to whether it was on the left or the right (confusion which led us to narrowly miss a green and get stuck at a red for what seemed like ages!), we got to the Radioshack, I ran inside and got a $40 inverter with only one outlet (and a fuse on the front mind you).
I ran back outside, and the rest of the team had prepared all the equipment for me to be able to plug-n-play. With everything on, we pointed the antenna in the right direction and started listening. The following two sentences appeared on fl-digi's screen (in parentheses are my comments):
$$PRHAL,18:05:07,1,+4027.6771,-08558.3035,01478*5E (<- passed checksum)
$$PRHAL,18:05:1:,!,+402.6975,-08558.2198,01202*56 (<-failed checksum) (numbers in red are altitude in meters)
$$PRHAL,18:05:1:,!,+402.6975,-08558.2198,01202*56 (<-failed checksum) (numbers in red are altitude in meters)
And then nothing. I couldn't believe what I was seeing. 1478m up? Just minutes ago it was at 24km!! I seriously couldn't believe it, but one of the messages had passed checksum which makes it highly unlikely it was wrong. Though the next one didn't pass checksum, the altitude portion seemed to confirm the payload was indeed falling. At this point it only made sense to check instamapper.com, the tracking site for the cell phone. If it had truly landed, there would be a report from the phone. Sure enough there was, and we immediately headed to the location indicated.
It was pretty far, about 30 minutes away from the Radioshack. As were were heading towards it there was one more cell phone transmission, and then the cell phone stopped transmitting. What worried me most was that on the map, the last transmitted location was near a river. And I knew from previous experience the locations transmitted were not precise. They could be off far enough such that the payload was actually in the river.
We parked as close as we could, and fired up the radio again. There was a tone, but no data. Confused, we started heading towards the river. There was a lot of foliage, much more than the area in which my first flight landed. The trees were tall, and skinny too - not many branches. I was worried about the payload being caught up the trees, or at the bottom of the river.... We just kept searching, and it took us some time to get to the reported location, but once we were there, Watler spotted a glimmer from the ground plane on top of the payload. As we approached we saw it was in fact the payload! Hooray!
But where's the parachute?
Everything got throw around a lot, but nothing was damaged. At the time this photo as taken I had already taken out the Zi8 to inspect for damage.
I was ecstatic to see my payload again, and to be 2 for 2 on recovery, but I was confused as to why I couldn't see the parachute. Turns out, there wasn't going to be any parachute to see:
Oops!
The parachute was attached to this ziptie through a swivel from Apogee rockets - meant to let the payload spin freely while not tangling the parachute lines. Given the insanely fast descent rate, the parachute must have broken off at high altitude, perhaps during the Post-Burst Chaos (commonly called PBC in ballooning circles, i.e. #highaltitude), or perhaps before burst at all. The video camera died before there was any evidence of descent (due to burst or otherwise), so I guess we'll never know what really happened.
I later calculated the impact speed to be around 56 mph! And pretty much everything survived intact. The only damage was to the 1/4 wave antenna on the bottom (you can sort of see how it got bent in the photo of me holding the payload in the forest). But that's just a piece of metal coat hanger - easy to fix! Perhaps most fortunately of all, the SD cards were fully intact.
I later calculated the impact speed to be around 56 mph! And pretty much everything survived intact. The only damage was to the 1/4 wave antenna on the bottom (you can sort of see how it got bent in the photo of me holding the payload in the forest). But that's just a piece of metal coat hanger - easy to fix! Perhaps most fortunately of all, the SD cards were fully intact.
The Pictures
Here are a few of my favorite pictures from the launch. The full complement of photos can be found here.
The Data
The system to store data to an onboard SD card worked perfectly, so even though we lost communications with the payload, we have its entire flight history. This is exactly the situation for which the system was designed, and that it worked as expected is pretty awesome!
I found an online NMEA to .kml converter which I used to plot the data from the flight. I had also saved the last track I'd made with the predictor, and overlayed that onto the data from the flight, resulting in the following pretty cool image:
I found an online NMEA to .kml converter which I used to plot the data from the flight. I had also saved the last track I'd made with the predictor, and overlayed that onto the data from the flight, resulting in the following pretty cool image:
Comparison of actual and predicted flight path data. The solid blue is the actual path, and the yellow/green is the predicted path. The areas where the blue dissappears is simply where it goes behind the predicted path. A .kmz file is provided for any interested parties to investigate the differences/similarities on their own.
As described in the caption, the blue is the actual flight path whereas the yellow/green is the predicted path. The file below contains all the data and can be loaded into Google Earth. For the record, the prediction was made using http://habhub.org/predict/, and was for an 11am flight. We actually launched at about 12:27pm EST, so perhaps that accounts for most of the discrepancy in the initial climb portion of the path. The descent discrepancy is obviously due to a much higher than anticipated descent rate :)
| predicted_and_actual_paths.kmz |
Moving forward - Where to go from here?
First, a quick look at what went right and what went wrong on this flight:
What went right:
This flight, like the first once, incurred numerous important lessons learned that will be invaluable for future flights. The experience of launching is absolutely worth more than the cost of the equipment that might be lost. Although the cost may be fairly substantial, it can be kept to a minimum by using fewer components. More importantly, with proper documentation rebuilding lost electronics is trivial.
From here, I've learned a lot about ballooning. These two flights have taught me a lot about what to watch out for when launching a balloon, and more generally how to handle large projects. Given the cost of ballooning (not only does a balloon + lifting gas cost at least $120, an entire weekend is generally necessary for one flight), and the successes of both these flights, I am moving on to learning about rockets, and how to properly fly a rocket. I'll be going for my level 1 certification with a kit built rocket soon, and then I shall test electronics and recovery systems with ground launches, assuring all components work as expected.
What went right:
- Radio transmissions successfully downlinked and stored on board
- Successful operation of the backup transmitter as well, although it might not have been needed
- Camera battery did not die, took lots of photos
- Successfully filled the balloon with hydrogen
- Payload successfully recovered
- No damage to payload container or equipment
- Communications were lost due to inadequate knowledge of power inverter limitations and equipment requirements
- Parachute attachment point failed
- Flight software ceased transmitting after GPS cable disconnected upon impact with the ground
- Items inside payload container were not adequately secured in case of hard landing
This flight, like the first once, incurred numerous important lessons learned that will be invaluable for future flights. The experience of launching is absolutely worth more than the cost of the equipment that might be lost. Although the cost may be fairly substantial, it can be kept to a minimum by using fewer components. More importantly, with proper documentation rebuilding lost electronics is trivial.
From here, I've learned a lot about ballooning. These two flights have taught me a lot about what to watch out for when launching a balloon, and more generally how to handle large projects. Given the cost of ballooning (not only does a balloon + lifting gas cost at least $120, an entire weekend is generally necessary for one flight), and the successes of both these flights, I am moving on to learning about rockets, and how to properly fly a rocket. I'll be going for my level 1 certification with a kit built rocket soon, and then I shall test electronics and recovery systems with ground launches, assuring all components work as expected.
