Propeller Voyage

Update:

HAPB-4 was launched 08/31/2008 at 6:34 am and recovered ~3 hours later. Maximum recorded altitude was 106,384 feet.

The sky was extemely hazy this day but the objective of capturing sunrise was accomplished. Enjoy the Images.

KML file of flight path.

Dr. Thomas Talley and Paul Hubner
 

I started editing my other videos when I noticed the sound wave pattern of the still camera as it was taking pictures. If you look at the sound wave you will notice the it tapers off as the balloon increases in altitude. I noticed this from my last flight when I was editting the videos but I had not really looked at the waveform until tonight. To me this is really cool because it shows how sound travels as you approach the vacuum of space.

Sound Wave

MultiProcessor Flight Computer (MPFC)

HAPB-4 uses Parallax's SPIN Stamp which contains 8 micro processors known as COGs. Each COG can operate independently or in a cooperative environment, leaving me with an unlimited number of ways to use the micro-processor. I also like the fact that it is in the Stamp format. The Spin Stamp module is not intended to be a direct drop-in replacement for any BASIC Stamp module and there are some important points to consider before purchasing the Spin Stamp but it can be a less intimidating method of moving to the Propeller.

I kept the basic functionality of HAPB-3 with a few exceptions. I have pulled the temperature probes since my APRS unit already has that function. I am also replacing the 32K EEPROM with Parallax's Memory Stick Datalogger. Finally, I removed the CRDR reset, SFC Satellite Acquired indicator, and the SFC. It sounds like a lot of chances but its not. The functionality will be incorporated into the SPIN Stamp so I do not need the second processor and associated supporting hardware.

Up until now the propeller was executing code in serial fashion much like the STAMP until I gave the procedures there own memory stack. You can see where the power of the propeller comes into play once you start using multiple COGs running parallel processes. You can run out of COGs very easily and now I understand why people would like to see additional COGs built into the chip. I am using all 8 COGs for this flight. If I need additional parallel processes then I will switch the External Signaling processes back to COG 0 and remove the Debug procedure but for now I will leave it as is because I can be COG-HAPPY until I really need them.

COG 0 - Main program space
COG 1 - LCD operations
COG 2 - Debug operation
COG 3 - GPS operations
COG 4 - GPS Serial operations
COG 5 - USB Drive
COG 6 - USB Serial operations
COG 7 - External Signalling

This flight will be recording GGA data every 4 seconds providing over 3000 data points over a 3 hour period producing a file size of ~300.00KB.   I should be able to get a better idea of maximum altitude this flight.  The Coolpix is setup to take a picture every 15 seconds which should result in over 600 images @ 3264x2448.  The original video recorder will record for over 3 hours and I picked up a 4GB SD card for the HD video recorder that will allow the flight to be recorded for 2 hours 19 minutes @ 1280 x 720 Pixels (H.264) up to 30 fps.  The video recorders are setup at opposite sides of the capsule to increase my chances of capturing sunrise. I set the external signal lights to flash at 40 pulses per minute below 60,000 feet in accordance with FAA rules even though I am flying exempt.

SPIN Code - HAPB-4.SPIN, FullDuplexSerialPlus.SPIN, Simple_Serial.SPIN, Serial_Lcd.SPIN
JPEG/BMP versions of the schematic and layout.

ExpressPCB versions of the schematic and layout.

Telemetry System

I am using the same telemetry system from HAPB-3 with no modifications. I have plans of replacing this unit with own using the propeller but that is for a later flight. For now I will continue to use the Opentracker. It has proven to be a stable and reliable system.

System Testing

I ran a series of tests while going back and forward to work with the GPS recording data every 4 seconds and the results looked good. For that series of test I was using a light toggling on and off to emulate the camera operations and did not have the external lights connected.

After integrating all systems additional tests showed that the external lights were not working properly.  I had written the procedures in a series fashion which resulted in the altitude values toggling between 0 and the actual value.  Adding a temporary variable that could be set to zero and manipulated with each cycle solved the problem.

Assembled the capsule and performed a final 4 hour test showed that all systems were functioning properly.  The finally weight of the balloon train which includes the capsule, emergency radar reflector, and parachute is 5lb 3-3/8 oz. Capsule is ready for launch.

Performed some additional testing this past weekend and found that my altitude procedure was not working properly.  It took several hours but I finally found the problem.  It seems that I was overwriting my altitude value.  Each cycle of the GGA routine would reset altitude to zero and then assign the actual value.  In a serial configuration this would be ok but I am working in parallel fashion now and that caused my anomalies.  The altitude procedure would slow down the external lights but never stop them.  You can see where I am going.  One time it would be zero and the lights would flash, the next time they would be the actual value and they would not flash.  Resolved the problem by reading altitude with a temporary variable and then assigning it to altitude. 

Completed final testing and the capsule is flight ready.  I set the external signal lights to flash at 40 pulses per minute below 60,000 feet in accordance with FAA rules even though I am flying exempt.

The finally weight of the capsule which includes the capsule, emergency radar reflector, and parachute is 5lb 3-3/8 oz

KML File - HAPB-4Test (Right click and "Save Target As", then double click to open in Google Earth)
Excel Spreadsheet used to format KML file

Launch Day (08/31/2008 - 6:34 am)
HAPB-4 was launched and recovered successfully today thanks to the team of Dr. Thomas Talley and Paul Hubner. And a special thanks to my wife who couldn't make this trip but was able to provide us with 0 hour prediction data, monitor the flight from google maps, and assist with remote recovery.

Released at Lake Whitney State Park at 6:34 am
GPGGA,113428,3155.4589,N,09722.1585,W,1,08,1.8,166.8,M,-24.0,M,,*7B

Burst occurred at 8:19 am and reached an altitude of 106,384 feet.
GPGGA,131942,3153.1798,N,09747.9075,W,2,10,0.8,32426.0,M,-23.8,M,,*761

Landed at 8:42 am in a field
GPGGA,134228,3153.2346,N,09752.2611,W,2,09,1.0,341.7,M,-23.8,M,,*7AC01

 
Launch

Before deciding to do a pre dawn launch I reviewed the FAA regulations again to ensure I would not be violating rules governing unmanned balloon flight.  According to 101.1 Subpart A I am not prohibited from performing a pre dawn launch as long as I stay confined to sections 101.1, 101.3, 101.5, and 101.7. Therefore all the rules in Subpart D do not apply and a predawn launch is not in violation of the rules governing unmanned free balloon flight and I am still launching with exemption status. With this information in hand I did not have any concerns or reservations about launching prior to sunrise.

Recovery (08/31/2008 - 8:42 am)

The recovery operations went fairly well during this flight.  We started to drive to the predicted burst area when I received a call from Sandy.  We all started to notice that the balloon was not following the predicted path.  We stopped and parked along CR22 after determining the balloon was not going to predicted landing area and observed the actual path. 

After a ½ hour watching it path we decided the payload was heading west of Meridian State Park.  I was driving us to the intersection of CR4130 and CR1238 where we thought it might land when we lost signal.  The last report had the capsule at 65,889 feet and still falling.  I stopped the vehicle and we scanned the skies in hopes of seeing capsule but it was still too far away. 

Nine minutes had gone by since the last signal report.  I told the crew that recovery would probably depend on the local farmers finding the capsule out in one of their fields.  

Eleven minutes went by and we were still looking for the capsule in the air when the computer announced KE5HDN DASH ONE ONE.  The payload was still descending and was at 18,366 feet still heading our way.  We were within 2 miles of landing zone and now were able to track the payload to the ground. 

We had to backtrack 8 miles to get to the landing zone because of the crazy country roads so the payload had already landed when we arrived near the landing zone.  The road ended 0.2 miles from that last broadcast position.  I parked the car and programmed the rhino110 GPS radios with the last known coordinates.  Even with the GPS unit it still took us about 15 minutes to get to the landing site.  The capsule had landed in a field with 2 foot high grass and was hidden by some bushes.  HAPB-4 was now recovered.

Damage Assessment

The capsule suffered a fair amount of damage. The capsule lost most of the bottom half of the bi-pole antenna at 65,889 feet. It appears this was due to the balloon and parachute whipping around the capsule and it was sheared off from the force. It looked like a clean cut through the plastic jacket, dielectric insulator and centre core. There were a few strands of metallic shielding that were torn and ragged.

The inside of the capsule had extensive damage. The LCD portion of the instrumentation panel snapped off from the impact, however, the LCD was not damaged. The jar of impact damaged the Coolpix L15 camera and the lens is now stuck in the "ON" position and the camera itself will not turn on. It is also possible that the trigger wire was pulled loose from the camera at impact but I speculate that it happened at 99,412 feet. This is the altitude that the camera stopped taking pictures and where the trigger wire probably came loose and pulled away from its contacts. Upon additional inspection of the camera there were over 3 gigs of memory available and almost full battery life remaining.

There were also video issues with this flight. The standard definition camera stopped recording at 81,647 feet due to power failure. The inside capsule temperature was recorded to be 41 F. The batteries had been replaced earlier in the week and may have been previously used. The HD video recording was the most disappointing. It was unable to focus most of the time during the flight and it stopped recording at 56,804 feet. Upon retrieval there was sufficient power to play the video so it was not a power issue with the recorder and had over 2 gigs of available memory. The inside capsule temperature at that altitude was 52 F. The cause of the failure is unknown; however, I had postulated a reasonable theory that fit until I received an email with an alternate theory.

The first theory is based on my testing.  During early HD testing I had problems with the camera recording anything over several minutes.  Through trial and error I isolated the problem to the wireless SD card I was initially using.  After changing to a standard SD card the recorder worked and I was able to capture over 1.5 hours of video at my desk.  During integration testing I choose not to turn on the HAM radio because I do not want to broadcast APRS signals needlessly.  I believe during flight the HAM radio signal had the same effect on the HD recorder as the wireless SD card and the camera stop recording due to frequency interference.

The second theory was brought to my attention by Doug of Seattle, Washington, and is just as plausible.  The HD record was just not designed to work at that altitude. Like many densely package, highly computational computer systems it needs air flow to remove the heat.  As altitude increases air density decreases reducing the efficiency of heat transfer.  The recorder simply got too hot and shut down.

The capsule sustained minor damage.  Several small twigs punctured the bottom of the capsule and the insulation has been compressed a 1/4" by the electronics due to ground impact.  I have calculated the impact was approximately 2092 feet/minute (23.77 MPH) based on my final recorded readings from the flash drive.

The MPFC, telemetry system, and LCD did not suffer any damage and were still intact and working when we located the capsule.

Damage Assessment
MPFC and Telemetry System

Back