Balloon Launch
Bryan Klofas - KF6ZEO - November 2003

[ Background | Plans | Construction | Launch | Chase | Recovery ]
[ Theories | What we learned | Weight | Raw APRS packets | Contact ]

Background:
So there we were, Chris and I, with a long weekend ahead of us. Not much homework due, and no big parties, so we were trying to find something to do. Idle minds are dangerous things. I mentioned that I had recently got a TinyTrak3, which is an APRS kit from Byonics. I told him I was planning on fitting it inside the battery bay of my DigiTraveller GPS unit to create a small lightweight APRS unit. He mentioned that he had a small 2meter transciever that we could hook it up to, and ideas started to form in our minds. A friend had just bought some 36inch party balloons, and another friend had a tall tank of helium. We only had 48 hours to start and finish this project, and we were waisting time!

Usually, launching a balloon takes time and money. Lots of planning goes in to who is where when with what equipment. Money is spent on getting the balloons and helium, buying the electronics, and parachutes. Extensive testing is done to make sure the electronics work; lots of planning, careful construction and implementation. Never mind planning; we had to launch that weekend, because both of us were going home for Thanksgiving the next week, and after that it was finals.

Plans:
Weight was our main consideration. The large party balloons that we were going to use could lift a maximum of 250 to 280 grams. Or friend didn't get real weather balloons for his project; he just didn't need them. And we weren't going to wait around for manufacturers to ship them to us. And we didn't have enough helium for big balloons; our other friends tank was only half full. Factor in that we want this balloon to rise, and the maximum weight was about 250 grams. Could we put a complete tracking system, batteries, and parachute in a small container whose mass was less than 250 grams? Chris and I were skeptical, but we decided to make it happen. If push came to shove, we could use two balloons, but that would probably cause more problems than it was worth.

Using two balloons would cause a few big problems. The first question was how to attach the two together. One idea was to stack the two balloons one on top of the other. A good idea, but friction was our enemy there. The string that went up to the top balloon would constantly rub against the lower balloon, potentially causing it to rupture early. We could put two balloons on a spar that separated them by a few feet, and hang the package off the spar. Neither Chris nor I had any suitable spar material. Chris had some carbon-fiber rods, but none of them were long enough. The other problem with the two-balloon design was that since no two balloons are every manufactured exactly the same, one would rupture before the other. This scenario would leave us in limbo; the package would not have enough lift to keep it rising, and too much lift for it to fall rapidly. The package would slowly loose altitude and get blown to the East Coast if it got stuck in a wind. I didn't feel like driving very long distances, so we shelved the two-balloon design. But that meant that we had to make things as light as possible.

We would use the DigiTraveller GPS, the TinyTrack3, and the Alinco, all without their cases, put it all together in a cheap tupperware case, attach a garbage bag for a parachute, and let it go! Fly! Higher!

Construction:
Since we were going to launch that weekend, construction had to be quick. The biggest problem we had was weight (do you see a reoccurring theme here?). As we drove to Chris's house, we talked about all the ways to reduce weight. Everything had to be under 250grams at the very maximum, and that only provided us with about 20 to 30 grams of lift. Now usually, having 5 newtons (about 1 pound) of lift is optimal to get the perfect 1000feet/minute lift. Since our balloon was only going to rise to 10k or 15k feet, then rupture, we just didn't need the tremendous lift. But more than 20g would be nice.

First, I built the TinyTrak3. I just received the kit from a friend only a week earlier, and hadn't had time to build it yet. To save on weight, we decided to leave off the two DB-9 connectors. However, this meant that programming was going to be a hassle, so we put one of the connectors back on. It took about an hour to build the TinyTrak3, but only because I take my time soldering and do a really good job.

While I was assembling the TinyTrak3, Chris was looking at the transmitter. He had a small Alinco DJ-S11 HT that we could use, but he wanted to use a modified FRS radio. Cheap FRS radios can be modded for the high end of the Amateur 70cm band by changing the PLL chip. Chris knew it could be done, but we were running short on time. Also, there is no national APRS frequency for 70-cm, so we would be the only ones listening for our balloon. That was kind of a scary thought; if something went wrong and the balloon went out of our range, we had no way of recovering it. On the other hand, if we went to the 2-meter band, there are many digipeaters, and lots of people listening all the time. I am a big fan of the national APRS frequency, and it didn't take long for me to convince Chris to use the 2-meter HT instead of the modded FRS radio.

I finished the TinyTrak3, and moved on to finding some tupperware for the case. He really didn't have any light tupperware that was the right size. So we decided to have a Tupperware Party! Heading out the local supermarket, we found the perfect size and shape. I brought all the electronics with us to test the fit. The only problem was it wasn't waterproof. None of the tupperware knockoffs they had were waterproof, and we decided that there wasn't going to be any other supermarkets open at 11pm on Friday night. We bought a pack of 4 tupperware cases for $1.07, and headed back to Chris's house.

When we got back home, I started to form aluminum foil into the shape of the tupperware. We would use the aluminum as a ground plane for our upside-down antenna, as well as a RF shield for the unshielded electronics inside. We weren't really concerned about extraneous RF leaking into wires; if there was a problem, our testing would catch it.

For power, we decided to use 3 AA batteries for a total of 4.5 volts. The Alinco HT has no problems with that; it has an integrated 3 AA battery holder. The GPS also has an integrated battery holder that takes 3 AA, so the GPS was fine. But how to power the TinyTrak, which requires 7 to 24 volts? Chris and I thought about it for a while and decided to bypass the onboard 5 volt regulator. Only half a volt less shouldn't kill it, since the 16F841 will work down to 3 volts or so.

What type of batteries to use? Our cheap answer was alkaline, but they don't work too well in cold temperatures. Nicads and NiMHs don't put out enough voltage to keep the system up at 4.5 volts. We were pondering the problem when I remembered that Energizer just came out with a 1.5 volt AA-sized lithium cells for digital cameras and other high-current devices. A quick trip to the local supermarket and $10.25 later we had our batteries. And, they were only XXXX each, XXXX grams less than the alkalines.

Integration of the electronics into the case posed a couple of problems. The radio was just too big with an external mic in the jack. The external mic was needed to connect the TinyTrak with the transmitter. That was solved by soldering the audio from the TinyTrak3 to the board of the radio. We weren't sure what pins to solder to, and after much trial an error we got the right pins. Another problem was that the batteries fit nicely under the radio, but the radio was on the bottom. Since a switch weighed too much, taking out the batteries was our switch. That meant that we had to unpack the package every time we wanted to toggle the power. Not the best plan, but we had no choice.

Testing of parts. Shown here is the Alinco DJ-S11 HT, TinyTrak3, DigiTraveller GPS from RadioShack, and batteries. Everything worked! Integration of parts. Everything fits!
Once the electronics were integrated, we began testing. This is where we found the GPS problem. Its fatal flaw is that it doesn't have a backup battery, so when it loses power it also loses the current satellite almanac. This means that when it loses power, it forgets where it is. The receiver was made near Tokyo, so it resets to the location of the factory outside Tokyo, Japan. Kind of funny, I think. And too cheap to include a backup battery. We could see ourself on findu, but it took about 10 minutes for the GPS to get a local fix.

Aside from the Tokyo problem, we found no problems when the package was integrated. Everything worked as planned.

Then I made the parachute. Built out of a half of a garbage bag, it was shaped like a stop sign with the 8 leads at the points. Chris ran some numbers and found the optimum length for the leads, and we reinforced the holes in our parachute with scotch tape, and tied the leads. It was done! We did a quick drop test in the living room, and saw that the parachute should have been bigger, but we didn't have a bigger garbage bag.

Right before I went home, I ran some winds-aloft predictions. Things were not looking good. If we launched from the CalPoly, the balloon would land near Vandenberg. Since Vandenberg is a high-security military/NASA facility, I thought it was a good idea to stay away from there. On an interesting side note, the NOAA office there launches 2 weather balloons there per day to collect wind data for the upper atmosphere. While this data is great for altitudes above 30k feet, the data is not good for surface winds, and since the was (hopefully) going to burst at a low-altitude. Finally I went to bed at 5am.

Launch:
I got up late Saturday morning, 22 Nov 2003. I called Chris, and he ran the latest numbers from Vandenberg. The winds had shifted considerably since last night. The winds were heading south. We could launch from Cuesta College or Morro Bay. If we launched from Morro Bay, the predicted balloon path took it over the Pacific Ocean near Pismo Beach. Not good, I didn't want it going anywhere near the ocean. Eventually it was decided to launch from Cuesta College because of the wind and also it was close by. We all carpooled out there and drove around the campus a while, trying to find a big parking lot without any lights, power lines, and far enough away from the police station. We found a parking lot at the corner of the college and released a "sounding" balloon. A sounding balloon is an unweighted balloon that tells us the local winds. Not good. The sounding balloon took off east, right for the ocean. We stood outside in the cold wind and had a group meeting about how early to get up the next morning. Earlier is better, and it was decided that we would meet at the hangar to fill another balloon at 7am the next day.

I got up at 6am on Sunday, 23 Nov 2003. An absolutely beautiful day; not a cloud in sight and very calm. I ate a quick breakfast, and arrived at the hangar where we would fill the balloon at about 7am. Cliff was right behind me, and Nash rolled in a few minutes later. Not bad for an early Sunday morning. I called Chris to find out where he was, and woke him up! What a bum. He quickly ran the latest winds-aloft predictions from Vandenberg. The winds had changed significantly overnight; they had almost dissipated. We picked a new launch site just south of town, a little bit farther away than the first.

While we were waiting for Chris, the rest of us filled the balloon we were going to use. We decided not to lug the tall helium tank out to the field. Since we only had a 36inch party balloon that could fit in a car, filling it at school was the easiest. We would then just stuff the balloon in the car and hopefully it wouldn't float away while driving out to the luanch site. When Chris arrived at the hangar around 7:45am, we tested the electronics package and made sure the balloon could lift it. We then stuffed the balloon and package into Nash's car and drove to the launch site.

Right when we got to the launch site, we release a sounding balloon. It went straight up. No wind at all. This is a great sign, because we didn't want to drive for a hundred miles to recover the package. We tied our flight balloon to the rack of the car, and plugged in the batteries. The GPS took longer than usual to find itself today, a full 10 minutes. Chris tested the voltage of the batteries and found it to be 4.3 volts, very low, but what could we do? While the GPS was finding itself, we launched another sounding balloon. It also went straight up, but we didn't see it for too long because it was a purple balloon against a very beautiful blue sky.

Once the GPS found itself, we checked the APRS packets and found that everything was in order except the elevation data. Elevation data was not correct, but maybe it just needed more time. Since we were eager to let the bird fly, we decided that altitude data would be acquired eventually when the GPS reciever found more satellites. We released the balloon at 9:03am.
The complete assembly is ready to launch.  Another sounding balloon is inside the car. This is the final integration before the launch.  The wrapped TinyTrak is being crammed into the case.
Launch! Chris watches it, but it's rising too slowly. Away she goes! By this time the GPS has failed.


The Chase:
The first thing we noticed was how slowly it rose. Maybe half a foot per second. Very slow. Too slow. We waited around the launch area for a while, and then got in the car. Something was wrong. The balloon position had not changed for a few minutes. I ran over to Chris and looked at his laptop, and his laptop had the same data as I did. Had the GPS failed? I guess so. The package was still transmitting every 45 seconds, but the Lat/Long was the same. Doh! We were out of luck, because in our quest to save on weight, we hadn't included a backup system. Well, what could we do? Nothing, except keep a visual on it, follow it wherever it goes, and hopefully recover it. Chris jumped in his car with Nash and headed down the road, and Cliff and I jumped into his car and took off in the other direction.

The first thing I noticed was how low the balloon was, and how slowly it rose. Two minutes of highway-speed driving changed the angle between us and the balloon by more than 90 degrees. It was lazily climbing, not wanting to go high. We turned around when we lost sight of the balloon in some trees by the road. We drove around some more, trying to keep under it. It was headed north, so we went in that direction and found a big parking lot and parked the car. Somewhere in the transition we had lost sight of it. I radioed Chris and Nash and they had lost sight of it as well. We spent a 20 minutes looking for it, and then headed toward the ocean; maybe we would see it there. We rendezvous with Chris and Nash on a tall hill overlooking the whole are. The balloon was not in sight. It was still transmitting the bad packets, and the bad paeckets were being digipeated, but we had no idea where it was.

We decided to head back to the lab and see if it was on the internet, and sure it was. But findu also had the "bad" packets. While we were at the lab trying to figure out what to do, the transmitter stopped working. I guess the batteries died, but there was that weird collision with W6SLZ-9 on the last packet recorded (see the
raw packets). Maybe it landed and was still transmitting but nobody could hear it. Who knows. We decided to drive up to Cuesta Peak (a very tall peak nearby) to see if we could hear it. I brought along my 4 element yagi, but the APRS packets are just too short to do any real DF. We looked around the ridge for a while, then drove back to campus, and talked about our next launch.

Recovery:
Recovery is way overrated.

Theories about failure:
Here is all the information I have about the failure of our package. Every 45 seconds a properly formatted APRS packet was sent from the bird. But the latitude/longitude data was always the same. This leads me to believe that either the link between the GPS and TinyTrak failed or the GPS failed.

The theory that I like the most is that the battery voltage dropped below what was required for the GPS to function. The TinyTrak sensed the loss of the GPS and just packaged the last known position into an APRS packet and sent that packet to the radio. Since the voltage was low to start, it probably would have only gotten lower as time passes. Who knows why the voltage was so low. I'm leaning toward a bad cell, but it could have been bad wiring or a loose cable adding some extra resistance.

Another theory that Cliff brought up was that some lithium primary cells need to be "turned on" or "warmed up" before they work properly. This is done by drawing lots of current thru the battery for a short period of time. The cell then functions normally. I don't know if these Energizer lithium batteries needed to be "turned on." We didn't "turn on" the cells; we just took them out of the package right before flight and put them in the holder. However, the more I think about this the less I like this theory. Total current draw on the whole system is about 350mA when the radio is transmiting. We were transmitting every 45seconds. I woud think that 350mA would be enough to "turn on" the cells.

Another candidate for failure was the TinyTrak. The GPS sends out 12 strings of data every second, and the PIC just might have gotten overloaded, and I just then transmitted the position it had in memory. I think this is very unlikely, though.

What we learned about launching balloons, for next time:
We learned many thing that will help ensure our success next time. First, put somebody on a high hilltop with some very good optics. If anything goes wrong with the electronics, the person on the hill can always keep a visual on it and relay to the chasers where the balloon is. This would have helped us out, and on our visual search for the balloon we came upon a few great spots for this person to be.

Reliable car-to-car communication is a must. For this launch, the two cars communicated on 70cm simplex with HTs inside the car. This was less than ideal, because if the cars were more than a few miles apart we could not communicate with each other. External antennas are needed to keep a reliable link open. Being up on the 70cm band is great because all of that extra RF doesn't interfere with the reception of the APRS packets. The HTs inside the car worked for short distances, but more than a few miles and we had to rely on cell phones (ouch!). Mobile rigs putting out 10 watts with an external antenna connected to a high-level 70cm repeater would be ideal.

We should have included a beacon or other DF-friendly transmitter. We thought about just cooking up a quick low-power beacon, but it all came back to weight. Another piece of electronics would be just too much weight. Next time we are going to put a little beacon on its own power supply with its own antenna, so if the main electronics fail we have a backup.

To get rid of the entire weight issue, next time we are going to fly on a full weather balloon. The 36inch balloon we flew this time just won't cut it. A bigger balloon means we don't need to worry about weight, and can include that DB-9 connector. A bigger balloon also means we can have more battery capacity, a backup beacon, and gain some altitude.

We also learned a lot about the digi path to put in the TinyTrak3. We used a WIDE3-3, which was a mistake. I thought it would be a good idea because if the balloon got caught in a wind we could still see the packets digipeated from far away to where we were. I talked with Bob WB4APR after the launch, and he said that balloons should use a WIDE path. I didn't know it at the time, but when using a WIDE3-3 the digipeaters do not include their callsign in the digipeated packet. When using just a WIDE, they do append their callsign, so we could have seen what digi our balloon was near, and could have helped us with locating it.

The moral of the story:
For less than the cost of two days at an amusement park, you can have lots of fun, and learn more than a class will ever teach you.

Appendix A: Masses:
GPS with case and 3 AAA alkaline batteries:134.6g
only GPS:41.1g
Radio with case and 3 AA alkaline batteries:195.8g
only radio:79.0g
3 AA alkaline batteries:71.2g