Missions‎ > ‎

NSL-48

On August 13, 2016, we flew experimental payload NSL-48.   Chris' friend Ben was interested in working a flight firsthand so that he could teach his students overseas.  Paul L. had some new hardware that he wanted to test.  So we tossed together a typical 'box' payload using Paul's worn-out old blue box -- its 12th flight!  In it was:
  • AP510 APRS tracker
  • Spot backup tracker
  • Mobius camera with 2000mAh LiPo
  • Mate808 camera, DIY model, with 1000mAh LiPo
  • Adafruit Feather M0 Adalogger (Arduino microcontroller) with 1200mAh LiPo and external Bluetooth
  • GMC-320+ Geiger counter
  • Remote control tree cut-down system
This payload was attached to a 90cm home-made parachute; hauled by 600g balloon with about 2200g of lift from hydrogen. .

The jet stream was headed westerly, so we had a choice to make:  Launch close to home, around Raleigh, and land somewhere north of Burlington; or drive an hour east to launch, allowing the payload to land closer to home.  The first choice risked landing in a more forested area, far from home.  The second allowed for landing in an agricultural area near home; but to beat the heat, we would have to wake up early and drive an hour to launch.  :(   Logically, it was decided to sleep in a bit and then drive to the east.  And stay in air conditioned cars as much as possible!

    
    
Tie-off of the balloon                    Prep from on-board camera                Click for launch GIF

The launch/chase crew consisted of Chris, Ben and his daughter, Paul and two of his kids.  Prep and launch went off without issue from a public park in Newton Grove, NC.   The flight was very 'nominal' and it had a calm burst at 26,740m.

  

Pamilco river hiding below in the haze (center top)

  
Descent down through the clouds

At burst, the majority of the balloon stayed intact and trailed above the chute.  Thankfully it didn't permanently foul the chute.   Flight timing allowed for the all-important fast food stop and we were able to get under the payload at landing.  One of the chase cars was able to stop and watch the payload descend into a field of soybeans.  We lost Spot upon landing and APRS was only available locally, so we figured the payload had flipped over.  Even though the parachute was bright orange, the chase crew had witnessed a lot of heavy balloon attached.  Sure enough, both the payload an parachute were buried deep under the soybeans.  It took over an hour to gain permission and walk the field to find the payload.
                  
Mate808 image of awaiting chase crew          On-board view upon landing and tumbling over

The landowner hauled the crew out the 200m to the landing site

  
Found it !                        Successful recovery         Time for ice cream


The flight followed extremely close to prediction.  It burst just 94m over predict.  There was a brief increase in descent rate through 16km.  Video seems to indicate that the chute became fouled for about a minute.  This caused the actual landing to fall almost 3km short.
  
Flight visualization -- Actual (blue) vs Predicted (yellow).            Close-up of landing.

  
Pretty typical ascent rates with slight gravity waves.        The payload stayed very warm on the sunny day

          
Geiger counter data



New hardware tests
  • The Mate808 is a new mini camera similar to the 808 #16 and Mobius.  List $25 !   This new model does not perform H.264 compression, so it uses very little power and does not heat up like the Mobius.  But its video files are larger.  It can record in 720 or 1080, and only weighs 6g without battery!  The "DIY" model comes without a case and has a 3cm ribbon cable on the lens (it is designed for easy integration into RC aircraft and such). Due to the lack of video processor, you cannot tweak exposure settings like you can the Mobius, but it uses so little power that it can run a long time on a small LiPo.  And again, list price = $25 !!!
    Flight questions:  How well does it work at altitude considering we can't tweak exposure?  What kind of battery life can we get? Does it interfere with GPS like the Mobius?
Results:  In ground tests, I got 110 minutes of 720 video on a 500mAh battery. In flight, the camera (1000mAh and 2.5hr flight) ran great and still had battery time left.  I set it to 720 so that the 32GB microSD wouldn't fill up.  As it is, I have 18GB of nice video to go through.  I forgot that I program the Mobius at 720 with not only a darker exposure, but that I set the white balance to "fluorescent."  This kills the greens, but makes a BLUE sky.  This looks cool up high, but really mutes the leaves down lower.  Note the different fields of view.  So although it is not apples-to-apples, here are some comparison shots:
  
Raw Mate808 image near burst                                    Raw Mobius image near burst (dark exp, fluorescent balance)
  
Raw Mate808 near landing                                            Raw Mobius near landing

I placed a single layer of aluminum foil on top of both cameras -- between them and the trackers.  This acted as RF shielding and no GPS issues were noted.  But in ground testing, I easily see a decrease in GPS SNR when the unsheilded camera was within 5cm.  The Mate808 may be as noisy as the Mobius.  I suggest always adding a layer of aluminum foil between these cameras and a GPS receiver.  
I was able to spot my chase car on the Mate808, but not on the Mobius video.  I'm very happy with the results.  Oh, did I mention it lists at $25?
  The Mate808 over-saturates with a black dot when looking directly at the sun.
Next we need to get a wide-angle version and see if we run into any focus issues at low air pressures (sometimes seen on 808 #16).



  • The AP510 is a APRS transceiver with serial and Bluetooth.  An Arduino was added to the payload to interact with the AP510 via Bluetooth.  A direct serial attachment was tested, but data was not reliable due to some sort of ground-loopy RF-y issue -- so Bluetooth it was.  The Arduino was programmed to record all APRS packets transmitted AND received by the AP510.  It would also intercept coded messages from the ground and reprogram the AP510 to respond.  Flight test:  What sort of APRS traffic can you hear in the stratosphere?  Is it possible to communicate with a payload using simple APRS messages?

Results:  This worked well (I dig the Adafruit Feather products).  The 1200mAh battery was plenty for the flight.  The Arduino tracked its battery usage and it only went from 4.07V -> 3.93V during the 2.5hr flight.  It recorded 1500 lines of data to its microSD.  So, what can an AP510 with a cheap 2m antenna hear from up there?  It recorded 608 packets from around the region including:
  •     many cars, weather stations, and repeaters around Washington, DC; Charleston, SC; Lynchburg, VA; Columbia, SC; Hagerstown, MD; Chincoteague, VA
  •     two private planes flying around Orangeburg, SC
  •     a car driving around State College, PA  !!
  •     and a repeater way out past the 'quiet zone' west of Green Bank, WV
Interestingly, as it went above 75,000ft it got quiet.  All it heard was a ham in his van headed from Oak Island, NC towards his home in New Jersey.  I copied his traffic even when no repeater was able to pick him up (ie. not on APRS.FI).  The cheap antenna was on the top of the box, so perhaps the box shielded other signals below.  I'll have to contact this ham to see what he has for gear on his van (he apparently runs 50W at home, so I suspect it is something powerful).

Periodically during the flight, I sent APRS messages from the chase car to the payload.  I was using an iPhone app wired up to my Baofeng handy-talkie.  My son just stuck the handy-talkie antenna out the car window.  Eventually the payload heard my signal and responded as expected.  Looking through the logs it appears that my iPhone app was sending malformed packets.  It wasn't until a repeater heard my transmissions and sent a clean copy that the payload actually responded.  So I'll need to work on my transmission rig.  I can also rewrite my code to workaround the bug.


Comments