My main goal is to perfect a modular high altitude balloon bus: A simple , low-cost, general purpose high altitude platform able to carry a variety of payloads. It will allow more launches for less money.

I will primarily focus on designing and building the bus. I would like another team to work on the payload(s). The latter will be flown free of charge.

I am currently trying to partner up with a high school. I’d love to involve some students, give them the opportunity to fly a small experiment and get excited about space!

The next mission is entitled Furiosity and is scheduled for launch in November.

Update (12/01/11):

I have been able to create a partnership with Valley Christian High School in Dublin, CA. I cannot give all the details yet but some students will work on a couple payload systems for my next launch.  I have pushed the launch back to February to allow them to work on it. More details to come in December.

The main difference with MSP-1 resides in the tracking system.

I got the technical radio operator license in order to get a HAM radio call sign. My call sign is KJ6MUD.
I was then able to use an APRS tracking system. I chose the Byonics Micro-Trak 8000 FA for its simplicity and light weight. The radio transmitter was connected to the Byonics GPS4 which does not have the 60k ft altitude limit.
Tracking the payload after the launch was very easy and accurate. It was set to send its position every minute.

Flight Path from launch to landing

MSP-2 at its highest altitude

All trackers worked very well, each showing the same data.

Flight Data

Despite being a lot heavier than MSP-1 (1kg),  MSP-2 (1.7Kg) reached an altitude of 29,133m (95,581 ft).

Landing site from Launch site: 87km (54 miles)

Flight duration: 2h10min

Average ascent rate: 343 m/min (1125ft/min)

Average descent rate: 687m/min (2254 ft/min)

Max fall rate : 206.4 km/h (128.3 mph) at 23,239.5 m (76,245 ft)

Descent rate at landing: 22.4km/h (14 mph)

Sensors

The computer system did not work as planned. The data from the external temperature sensor and photo resistor does not make any sense. It could be due to a bad mounting (no protection against the wind and the sun…).

The micro-controller had an operating temperature of -40 to +85 °C (-40 to +185 °F) but the internal temperature sensor stopped working at -20°C (-4°F). It had been thoroughly tested at low temperatures. It could be due to hardware failure (wire disconnection?).

The internal temperature sensor stopped working during the descent.

The camera sensors worked the whole time (Temperature in °C):

Battery Voltage (in mV):

This is actually quite surprising. I thought that 3600mV was the hard deck for Canon cameras. But I wrote an intervalometer  script in ubasic to disable the LCD and take pictures every 10 seconds. Disabling the LCD had a substantial impact on the battery life.

The camera battery lasted 4h20min and took 1314 pictures. The standard battery life was 260 pictures!

Pictures

The ascent:

Travis AFB

The San Francisco Bay Area from 95500 ft:

South Bay

San Francisco (You can guess the Golden Gate and the Bay Bridge)

North Bay

Sacramento Valley

Recovery

MSP-2 landed after a 2h10min flight. It landed in a wind farm, east of Livermore, CA.

Despite having the GPS coordinates, when typed into the Google maps application, it did not return the location entered but the location of the closest road, several hundred yards away. We couldn’t find it the first day, I had to come back the next day.

When it landed, the winds got so strong (40+mph), the payload got dragged about 200 yards uphill.

What is left from the balloon. It does not completely burst:

The lines are twisted all the way down and the GoPro Camera is gone:

Despite the scotch tape and the super-lock strips, most of the devices are now moving free.

Conclusion

• Tracking MSP-2 was easy and reliable. (Using APRS was definitely the way to go)
• Recovering it was hard; Next time do not rely on your mobile devices as it’s neither convenient nor precise
• Secure the critical devices better! (trackers, tracker batteries, Gopro…)
• The computer system and sensors need to be soldered all together
• You definitely need to be 2 to launch a high altitude balloon

My most sincere thanks to my friend TJ without whom I would not have been able to make it.

The external temperature sensor failed but I recorded the values from the camera temp (Celcius) and voltage sensors over time (minutes):

Temperatures from the Camera sensors

Voltage from the Camera sensor

The Canon camera and GoPro video-camera stopped working mid-flight because the batteries died. The voltage drops are due to the flash (while still on the ground). It should have been disabled.

The GoPro is still functioning very well. So are all the other trackers and computer systems.
The Canon Camera’s video mode and on/off button are not working anymore. It is still able to take pictures with normal quality.

The communication cutoff is due to the camera module being placed directly above the tracking module. With no satellite acquisition both devices stopped transmitting any position report (even if the phone had good GSM reception).

The bad mounting and the very thin shroud lines induced the payload spin.

All in-flight pictures:

[Show as slideshow]

1234...222324►

The first thing to know is that it is legal. It’s regulated by the FAA under Part 101 Subpart D of their regulations.

If you read the whole Part 101 regulation you’ll notice at the very start that the regulation only applies to unmanned balloons with payloads heavier than 4 pounds. Note that  you should still follow all of the rules in Subpart D to make the experience as smooth as possible.

I was a little worried about talking with the FAA. Especially since I was not sure of the phone number and requirements.

If you were to send a big balloon to Near Space, you would have to contact the right ATC (air traffic control) office. But in our case, the ATC office doesn’t actually have anything to do with small unmanned balloons.

What you need to do is create a NOTAM (Notice To Airmen).

NOTAMs are notifications issued for hazardous reasons, such as:

• Air-shows,
• Parachute jumps
• Closed runways
• Erupting Volcanos
• …

Who do I need to call?

The ATC office I talked to gave me the phone number for Prescott Flight Service Station (877-487-6867).

I believe that they handle all the NOTAM reporting for the West side of the US. Where does West start? No idea, you should call your ATC office to make sure. I know that many people launching  balloons from different states (CA, UT, NV, …) called that specific office.

What does the FAA need to know?

What they ask you differs from one person to another.

What they asked me was:

• Date and time of the launch ?(You need to notify the FAA between 24 hours and 6 hours before the launch)
• Location of the launch? (City, States + GPS coordinates)
• How far it is from any major city, airports, air force base? (It needs to be 5 miles away from everything)
• What is the ascent rate?

That was it for me. It was pretty straightforward.

Be prepared to be asked:

• At what time will it reached 60,000ft up?
• At what time will it reached 60,000ft down?
• At what time will it land?
• Where will it land?

Theoretically, you should also notify the NOTAM station when you actually reach 60k feet up and down, if you lose communication with your balloon, and when your payload has landed. Realistically, no one cared when I called and that was fine by me.

Why 60,000ft? Because most of the air traffic flies below this altitude.

How to choose your launch location and predict the flight’s trajectory?

When looking for your launch site, you need to think about 3 main factors:

1. You must not have any fence, phone pole, electric pole, tall tree around you.
2. You must take into consideration the winds which can carry your payload up to 200 miles away.
3. You must take into consideration the estimated landing area.
4. You must take into consideration the estimated flight path.

1) You really do not want all your work to end up stuck in a tree or attached to a 100ft high electric line.

2) This is critical. There are 2 web tools: This one and that one. Their prediction are, of course, not 100% similar or accurate but they will give you a good idea of the flight direction.

3 & 4) Remember that you need to be able to recover the payload quickly (Batteries might die) and be able to receive signals from the landing location. Avoid highly populated areas or areas with lakes, mountains or dense vegetation.

Note that  you need to make sure your launch site is more than 5 miles away from any airport/AFB.

Against all odds, I recovered my first high altitude balloon (MSP-1) 26 days after its launch.

I launched it on December 31st 2010 and thought I had lost it. But out of the blue, on January 25th 2011 at 7:42am, I received a manual Check/In message from the SPOT messenger sent with MSP-1.

The location reported was: 38.00134 , -120.80861. About 60 miles south east from our launch site. I scrambled a recovery operation the next day to find it.

Once arrived, my assumptions were confirmed: it was a private property used to breed cows. I went to the closest farm to inquire about the balloon and get their permission to look for it. After waiting one hour and a half and seeing no one, I started to hike up towards the location reported. I quickly understood that cows do not like to be disturbed, especially when their calves are around.

As I was walking downhill and reaching the estimated position, I saw no signs of my balloon and started to despair… until I spotted the red fabric of the parachute.

Note the pieces of Styrofoam everywhere. They are too small and spread out to be due to the impact.

MSP-1 used a 6 foot wide parachute for a descent rate inferior than 15ft/s. Even if it landed hard on the only rock of the field, it would have not have created so much damage.

The Spot messenger was hanging out of the tracking module’s remains by a piece of scotch tape. You can also see what was the on-board computer system.

The payload was divided in two parts. The camera module and the tracking module.

The first was almost intact, just covered with the usual dirt and mud after 26 days in a field. The tracking module was however completely destroyed. I found pieces of Styrofoam 20 ft away from each other. Moreover I found the USB battery charger, a USB key and some cables several feet away from each other and they had been pushed into the ground.

So the only explanation I have is that the cows trampled on it a lot. It freed the SPOT messenger which somehow sent me a message.

I retrieved the SD cards and log files and here are the results: The balloon flew 96 minutes before it burst. With a theoretical ascent rate of 320m/min. We reached over 100,000 ft (30km) high.

A few seconds after the launch

Under the clouds

Above the clouds

Facing the sun at max altitude
(Note how low the sun is. It’s 4pm in December)

Facing East getting dark

At 100,000ft #1

At 100,000ft #2

Snapshot #1 from the on-board video camera

Snapshot #2 from the on-board video camera

The balloon burst and is now hanging from the payload

Some pictures are blurry because the payload was severely spinning  the whole time. Despite everything, most of the equipment (cameras, trackers, main PCB) can be re-used, which will be a huge cost relief for MSP-2.

I have a lot of open questions right now:

• Why did the camera stopped working on the way down?
• Why did the computer system stopped recording the external temperature?
• Why was payload spinning so badly?
• Why did the transmissions stop before the launch?
• …

It will take me some time to collect and analyze all the data and understand what exactly happened but this maiden flight brings a lot of knowledge and exciting new opportunities!

I received at 7:42am a manual Check/In message from the SPOT messenger sent with MSP-1.

The location reported is: 38.00134 , -120.80861. About 60 miles south east from our launch site.

The location really is in the middle of nowhere. It’s in the middle of an unworkable field 1.5 mile away from the nearest paved road. There is a dirt road about half a mile away. This area is protected all along by a fence and I have seen cows on Google street view archives which makes me think that it’s a private property used for livestock farming.

Fortunately there are only 3 farms around including one less than 2 miles away.

So my plan is:

1. Go directly to this farm tomorrow morning.
2. Ask if they found the payload
3. If not, whose field it is and ask its owner
4. If the owner has not found it, ask to check out the location
5. If not found at the location, ask at 3rd Farm
6. If the 3rd farm has not found it, post reward poster with contact information

I hope to have more news tomorrow evening.

• Program a system with servo to push the camera’s shutter automatically or remotely
• Program the camera to take pictures automatically

The latter is by far the easiest way to take pictures in Near Space.

The most famous way to program your camera is to use CHDK. CHDK (Canon Hack Development Kit) is basically a firmware enhancement software allowing you to control additional features and options of your camera.

The tool supports a lot of Canon digital cameras. So you really want to use a Canon camera. You can find the supported list here.

Note that it’s an enhancement tool, it is non-permanent and non-destructive. You load it from your SD card and it makes no actual changes to your camera and firmware.

Installing CHDK is pretty straightforward if you have an SD card of 4GB or less. You can find the step by step instructions here.

CHDK was created to run on FAT partitions and SD cards that are higher than 4 GB run on FAT32 partitions. For 8GB SD card or larger, things get more complicated. You can find how to install it here.

Note that despite all my best efforts, I have not been able to make it work on my 8GB SD card. It just would not work. (Maybe due to my SD Card brand…)

Once installed, you can create a script to take pictures with the desired settings every X seconds until the SD card is full or until the battery dies. You can find some intervalometer scripts here. Another interesting feature is that you can save the battery voltage and the different temperatures recorded by the camera sensors (optical elements , battery and charge coupled device (CCD)) in a log file. More info under get_temperature.

I am only describing the ones I have personally tried. This list will be updated as I launch additional high altitude balloons.

GSM GPS Phone with Tracking Software

A GSM Phone will only work up to a couple of thousand feet high. Therefore it’s mainly a way to find your payload once it landed.

The first criterion to think about when choosing the phone is its service network. You need to verify the phone will have a good signal where you plan to launch and retrieve it.

In the Bay Area Verizon is known to have the best network. However Verizon is known to sometimes block the GPS access requests from third party applications. The type of network (GSM, iDEN, CDMA…) doesn’t matter either as long as the carrier provides an unlimited Internet data plan (GPRS, EDGE, EV-DO, UMTS, or HSPA…) for the phone. You’ll need to get a data plan from your carrier for the tracking software to transmit its position over the Internet.

The cheapest, lightest GPS phone I found is currently the Motorola I296. Its previous model which has been used in many high altitude balloon launch, the Motorola I290, has been discontinued. However you can still find on ebay. The I296 costs about $60.

Pros:

• Inexpensive
• Simple and Light

Cons:

• Requires a signal at land site

There are two out of the box tracking software:

• Accutracking
• InstaMapper

I chose Accutracking because it simply offers more features and flexibility than InstaMapper.

• Customizable buffer for store positions while out of coverage (100 max points for InstaMapper)
• More details about the phone at each point (location + GSM Signal strength + GPS signal strength + Battery level)

In this screenshot, the dates in red shows positions that were initially stored because of the lack of coverage

Installing the software on the phone is straightforward. Instructions can be found here.

Satellite Tracker

The cheapest and most convenient is the SPOT Personal Tracker. This tracker does not need cell tower coverage. It communicates its position by satellite. You log on the SPOT website to see the locations it reported.

Note: The Spot Tracker needs to be facing up towards the sky. Make some real tests with your tracking devices activated and inside the payload as it would be during its flight (closed and tapped).

APRS Transmitter (updated 10/02/2011)

To learn the basics about APRS, click here.

An APRS beacon is basically a device (transmitter + GPS) which emits a radio signal containing its current location.

These messages are received by listening radio stations called digipeaters. There are thousands of digipeaters across the US. Digipeaters repeat the message received to other digipeaters until it reaches one radio station that has access to the Internet. These Internet gateways are called an Igates and publish the received messages on the Internet. You can then follow the location of your beacon online at websites such as Google Aprs.

There are plenty of APRS transmitters available. I chose Byonics’ Micro-Trak 8000 FA for its light weight and flexibility. Byonics has other models though.

The Micro-Trak 8000 FA operates on amateur radio frequencies, and requires a valid amateur radio license to use. To get one, take the technician HAM radio license test.  To take it, go here. Most centers propose a one day combo: class + test.
You can just show up in the morning not knowing anything, study in the morning with them and take the test in the afternoon.

You will need a GPS that works above 60,000ft. I bought the Byonics GPS4 (works up to 84km = 275,000 ft).

Pros:

• Sounding balloons
• Cold Weather balloons

Sounding balloons will typically weigh in at 200 grams to a full kilogram and lift 250 grams of payload. They will also go up to 20 – 30 km.

The cold weather balloons can carry a good kilogram and are made to go to the stratosphere (from 10 to 50 kilometers high where temperatures can be lower than -75 degrees C)

The balloon I used for MSP-1 was a Kaymont Cold Weather KCI TX1200 balloon.  It cost about $105 (shipping included).

The balloon has a 12 cm long thick neck used to fill the balloon and attach the load lines.

When you are manipulating the balloon, I highly recommend that you use latex gloves. It’s not to protect you but to prevent you from damaging the balloon with your nails and your sweat.

Remember: The balloon is very fragile. Find an area that’s very protected from the wind otherwise, it will bounce a lot and may burst. Because it’s so fragile, you must launch it from somewhere with no possible contact with fences, trees, phone or electric lines.

How much Helium?

To fill the balloon, you must use Helium and not Hydrogen.

Hydrogen is lighter than Helium but the difference is ridiculous when you think about the risks. Hydrogen is extremely flammable whereas helium is an inert gas. Moreover Helium is easy to obtain and you do not need any special permits if you decide to start transporting several tanks of it around in your truck.

You can find helium tanks at local party stores or welding supply stores.

I found mine at SF Party in San Francisco.

One cubic foot of helium will lift about 28.2 grams. Kaymont recommends a gross lift of 3440g. So I needed 3440/28.2= 122 cf of helium. I took the 244cf tank + a dolly. It cost me $148. Be careful the tank is very heavy.

How to fill the balloon from the Helium Tank?

Helium tanks from party stores usually come with special regulators for party balloons. It’s not the best thing in our case because the flow rate is low, you need to constantly bend the valve  and consequently it slows the filling process down.

The obvious issue is to link the small regulator’s mouth to the 3 cm wide balloon’s neck.  You need to build the joint/adaptor yourself.

You need:

• A plastic tube (3 to 6 feet long)
• A 1 inch PVC 90 degrees Elblow
• A 1 inch female PCV tube
• Some scotch tape

To make sure the dimensions are correct, bring with you the balloon and the regulator when you go shopping. You can find the materials at Lowe’s or Home Depot.

The reason you want to a long plastic tube is to be able to inflate the balloon on a clean plastic cover away from everything.

The balloon’s neck is about 3cm (1.2 inch) wide, a 1inch wide PVC tube is large enough to keep it tight. Make sure to secure it with 2 tie wraps and a solid pair of hands.

Since the inflating process can take up to one hour, the person responsible for the balloon can get very tired. The reason you want an Elbow shape is to attach the inflating balloon to something heavy to prevent it to fly away and to measure the lift.

How much Helium is enough?

The tank regulator may or may not have a volume gauge. To measure the balloon’s lift, I found very hard to use a hanging scale. It’s great to weight your payload but absolutely not convenient to measure the lift. The scale is not accurate if there is any wind. It requires at least 2 person and tends to recalibrate itself each time you switch it on.

The best way to me is to attach the regulator joint to a jug of water. Add the desired amount of water to match the lift (3440g -> 3.5 liters). When the balloon is at equilibrium you know you have reached the desired lift.

Remember: It’s always better to have more lift that not enough. (More Lift = Faster ascent rate = Shorter flight time = Fewer problems)

About sealing the balloon, Dave Mullenix has the best instructions: using tie wraps or heavy twine, tie the balloon neck off tightly above the filling adaptor. Remove the joint. Tie the neck again, four to six inches below the first piece of twine. Tie your payload cord to the neck between these two pieces of twine. Now bend the neck over double and tie it again, twice. You will now have the neck of the balloon bent over double, with the payload suspension cord nestled in the bottom of the bend and the whole thing securely tied.

You’re ready to go over your pre flight checklist.

I wish I  had great news but the truth is the launch failed. I don’t have all the details yet as I need to analyze the little data I have but here is my understanding of what happened:

The main factor of this failure is me.

Bad Timing

I filed a NOTAM to the FAA for a launch at 11am PST near  Davis, CA.
The launch site was 2 hour away from San Francisco. Once on-site we realized that the field I picked thanks to Google Map had a fence. We had to relocate quickly. We found another field that seemed appropriate.

To make a long story short, our 2 wheel drive truck got stuck in a fresh 10-inch deep mud. It took us 2 hours to get out of there. By the time, we got free and had found another spot, it was 12:30pm. We already were 1h30 late.

We started setting up the balloon and payload. The 1200g Kaymont balloon took about 45min to fill.
I enabled the SPOT GPS Messenger and the accutracking software on the GPS Phone.
I saw both signals from my Iphone and another smartphone. Once the balloon was almost ready, I closed the payload lid and added the camera module of top of it. We tapped the whole thing securely, attached it to the parachute and the latter to the balloon.

Checklists are made to avoid stupid mistakes

We were ready to launch at 2:37pm. It had taken 2 hours to set it up. With a minimum of 2h30 flight time and a sunset at 5pm, we were terribly late with no margin of error.

So I committed the unthinkable, I did not go through my preflight checklist.
We launched it as soon as it was ready. The lift had been correctly calculated, within a couple of seconds the balloon was 100ft high. The feeling of seeing your work flying up 15ft/s was absolutely awesome.

“We lost it”

I went back to check the tracking data and realized that both tracking systems had stopped emitting since 2pm-ish.
Before we even launched it! This could only mean one thing, the camera module and/or the tape added prevented the  devices to send any data.

Worse, in the hurry, I had forgotten to put my contact information on the box.
There was literally no way for us to track it and no easy solution for anyone who may found it to contact us. “We lost it” I announced only 5 minutes after the launch.

Experience is simply the name we give our mistakes

I lost a SPOT-1 Messenger, a Motorola i-296, a Parallax circuit board with temperature sensor and data logger, a Canon SD880IS +4Gb SD Card and a GoPro HD Hero + 32Gb SD Card. I failed my family and friends. I am immensely disappointed.

The positive note is that I learned a lot from this project.

-First and foremost, time and action steps are critical. You cannot compromise one for another. Everything has to be done on time and has to be checked. Going over my preflight checklist would have allowed me to identify all issues.

-The second and maybe the most important is to stop reading about everyone else’s projects and feedback but to experiment myself. I spent a ridiculous amount of time trying to understand how to build a joint between the helium tank regulator and the neck of the balloon. I should have just built it and tested it instead of thinking about it for so long.

From the top of my head:

-The launch requires at least 2 people. 3 would be better.

-The payload layout should be simple. In order to reduce the time of the launch sequence,  the shroud lines should be ready to be attached to the parachute.

-Zip ties are your best friend

-The payload experiments and devices should be tested in preflight conditions: ie in a moving, closed and tapped styrofoam box. I suspect my heavy duty scotch tape to contain aluminium blocking signal transmissions.

-If your lift is correct, a truck bed is good enough to launch the balloon. All you need is a clear 75-yard area around you (no fence, no phone or electrical lines)

-To make sure the lift is correct, do not use a hanging scale because it tends to shutdown during the inflation and then re-calibrates itself. Use a jug of water with the appropriate weight. When the lift equals the opposite force, the balloon is in equilibrium, you are good to go.

-People are interested and usually friendly about the launch, it’s okay to be near houses.

Next Steps

I will definitely try again. I need to save some money and will launch a simplified version in a couple of months.

I will detail the architecture of MSP-1 in the next few days.

Blue skies,

Marc