Altitude: 108,172 feet
Flight: 2 hours, 26 mins
Distance: 45 miles
After a two-year hiatus, we launched yet-another balloon. Not just to use up that tank of helium in the garage, but also to test out a new SurlEE data-logger board, a new MT-1000 APRS tracking transmitter from Byonics, and a three-video-camera setup.
The predicted path from Habhub looked decent, though not with much margin of error for going long (so we were careful to put enough helium into the balloon, which is done by levitating a water jug weighing the desired amount of “neck-lift”).
We launched across the street from “The Domes of Casa Grande,” an odd collection of unfinished structures that were going to be an electronics manufacturing site, as started in the 1980s. Or something like that.
The new SurlEE Balloon System One data logger is on the left, connected via ribbon cable to another board in the Sensor Bay on the side of the payload — this bay provides airflow over temperature, humidity, ozone, and particulate sensors which are protected from direct sunlight heating. There is also a master power slide switch in the bay, which powers both the logger and radio sections from independent battery systems (6x AA Ultimate Lithium cells each). The Byonics radio modules are, from far right: the new 1000 mW MT-1000 APRS tracker, a 15 mW “fox” beacon transmitter, and a 400 mW MT-400 APRS backup transmitter (on the logger power system). Underneath this module are the battery packs and the geiger counter.
We launched 50 minutes late, which put us at the mercy of changing upper-level winds. However, we had run predictions for several target launch times and it did not seem to be much of a concern (at the time).
The kids always have fun at a balloon launch.
That is the shadow of the balloon, of course, not a hole in the ground. Once it lifted off, we could tell that the ascent rate seemed reasonable (we have had some sluggish starts with too-little helium, and a longer chase). We are off again!
1500 feet AGL at 2 minutes. Looking north to Casa Grande.
8000 feet AGL at 9 minutes. Looking down at the motocross tracks at Motoland MX. Must be time to dust off my 87 YZ250.
25,000 feet at 28 minutes. Looking southeast down I-10 towards Newman’s Peak, Picacho Peak, and, in the distance, Tucson.
28,000 feet at 30 minutes. Looking down on Casa Grande.
30,000 feet at 33 minutes. Looking south towards launch site (lower center).
34,000 feet at 35 minutes. Looking down on Sacaton Peak.
57,000 feet at 62 minutes. Looking down on Florence and the Gila River.
58,000 feet at 64 minutes. Looking down on Gila River, Ray Mine, and Pinal Peak.
92,000 feet at 95 minutes. Looking east to the Juniper Fire burning behind Roosevelt Lake. Moving down the Salt River from Roosevelt is Apache Lake, Canyon Lake, Saguaro Lake, and, a little further downstream from this picture, Tempe Town Lake.
108,000 feet at 108 minutes. Looking west past Maricopa towards Gila Bend.
On the far side of Gila Bend is a giant solar power plant: the Solana Generating Station. Abengoa, the Spanish company who built the under-performing Solana, is trying to go bankrupt and leave something like $1.4 billion US taxpayer loan dollars unpaid. This plant has parabolic trough mirrors focused on pipes containing a heat-transfer fluid, which contains nasty stuff like benzene and toluene. Apparently there have been leaks and even a fire there. So why don’t we just snuggle it up right next to farm fields where we grow stuff we eat? The same reason they put houses beside high-voltage power lines and airports, I can only presume.
But I digress: since these parabolic mirrors are curved, we saw a strong reflection from about 35,000 feet all the way to 108,000 feet. Simple flat solar panels would only reflect at one angle (and only be seen at one altitude).
108,100 feet at 109 minutes. Looking down on San Tan Valley. The balloon should be about 28 feet in diameter now, and the jack antenna is in its shadow. We are about to burst.
We burst at 108,172 feet,110 minutes after launch.
Above, you can see how almost all of the balloon detached. We lost more weight than expected — which means descent will be slower than calculated. Hmm, that’s not good.
In 15 seconds, we reach a descent speed over 200 mph!
We fell 10,000 feet in 45 seconds.
Down to 80,000 feet at 112 minutes. Looking down at Mesa and Gilbert.
30,000 feet at 128 minutes, over the landing site.
5500 feet at 141 minutes, over the landing site.
3000 feet at 143 minutes. Note minimal balloon remnants — this is likely why we went long, as we planned for about half of the balloon weight in the parachute load calculation. Weaver’s Needle and Superstition Mountains in the background
The last public (144.390) APRS packet we got was from 3173 feet (MSL), but I got one on my radio in the car (144.920) from about 2500 feet MSL — since that was about the elevation where it went down, we presumed it was pretty, pretty, pretty, pretty darned close to the landing position. So we had good numbers for latitude/longitude/altitude, but sadly we tried to hike in from the wrong side of the mountain and reached sheer cliffs.
I had multiple-packets-a-minute in the car on the private channel with just a vertical mag-mount on the roof and the Kenwood TH-D72 decoding and displaying balloon position the entire trip. Byon iGated the private packets to the APRS internet system, if you wondered why there were so many position points displayed on aprs.fi.
The next day, we hiked back in. It was nasty, rocky terrain, and a killer hike. From the top of the ridge we could hear all three transmitters (even the low-power fox), but the balloon was still hiding 1200 feet down into a ravine.
My brother Kevin was brave enough for the final part of the hike, and found the payload!
Air pressure vs altitude is using a new sensor, the MS5607 from Measurement Specialties. Allen Jordan at NOAA turned me on to this part. It is tiny and much cheaper than the Honeywell unit we used previously. The trick was getting data from it and massaging it to get the actual pressure numbers, as it is expected that you will be using 64-bit integer math. There are lots of spots in the calculations where this gigantic number is subtracted from that gigantic number and then… I had to shoehorn into into a compiler that only had 32-bit math, so it took a bit of trickery to adapt computations to keep from overflowing. My old numerical analysis professor would be proud. A lot of shifting up and down to manage magnitudes, yet keeping enough significant digits for accuracy. I could have changed compilers but I had a lot of this logger code finished two years ago for this PIC processor.
The calculations have a “first-order” result, and are then refined with a “second-order” tweak when the sensor gets cold. It is only a first-order calc until temp drops to 20C (68F), and then step one of the second-order refinement kicks in. When the sensor drops below -15C (5F), there is a slightly different second-order calculation. Since this sensor is inside the payload, and it barely hit freezing inside, this threshold was never reached. Anyway, both calculations are plotted together and seem to be identical.
On our last mission, the external temperature seemed higher than expected during ascent, but reasonable during descent. It was surmised that solar heating of the black plastic sensor in the sunshine was the culprit. So this time we have a sensor bay that directs airflow over the external sensors, but shields them from the sunlight. It seemed to help, though there is still a discrepancy. Probably should have wrapped the compartment in aluminum foil to minimize radiant heating as well. There were two new temperature sensors: TMP-100 (flawless I2C readings — plotted here) and DS18B20 (reading within a degree or two, but had some missed 1-wire conversions leading to glitches in the log data). Internal temperature was a little toastier than our previous payload box. Humidity was low that day, and essentially zero above 20,000 feet or so.
Here we see temperature from launch to recovery. The beginning of the graph is the flight, of course, followed by heating up sitting on the ground, cooling at night, and warming back up the next day. The running electronics kept the inside about 20F warmer than outside at night, until the sun came back out the next day.
Horizontal Speed and Heading is pretty self-explanatory.
Vertical speed gets interesting right around burst time.
We first flew the geiger counter on our last mission, and we see the same interesting results again. There is a definite peak around 50-to-60,000-ish feet, which seems counter-intuitive. I would expect radiation to continue to climb all the way up. It leads me to think something is happening in the circuit — it uses a 400VDC supply and all high-voltage areas need to be sealed so as to not start conducting at low pressures. Is it related to a leakage path? Is the 400V switcher crapping out at low temps? Is it the 1950’s-vintage Soviet geiger-muller tube? Not sure, but it is a fun experiment.
The particulate sensor also flew on our last mission, but we stirred up so much pre-launch dust, and the sensor was in a styrofoam compartment that perhaps shed junk from the walls, so previous results were rather inconclusive.
On this mission, the results actually make some sense. The larger PM10 levels were very high at dusty launch, dropped way down for most of the flight, spiked at burst (of course!), and spiked again at landing. The tiny PM2.5 particles show a similar trend, however, it definitely shows tiny particulates in the atmosphere up to about 20,000 feet. This probably makes sense. This is a simple and uncalibrated sensor, but the graphs are interesting and likely real in a relative sense. I was glad to get good data from this sensor this time.
One other sensor was new this flight: a MiCS-2614 ozone sensor. It is a variable-resistance sensor material that gets heated by a small power resistor. It needed about 35 mA but I deemed the extra power worth the investigation. Now, running a current through a heating resistor is all well and fine but we are going to very low temperatures so how well can the heater work? The unit is rated down to an operating temp of -40C (-40F), so it seemed to be just fine, and there was no mention of temperature compensation in the skimpy datasheet. The sensor current flows through a resistor across which voltage is measured. That resistor is selected to keep the resulting voltages in the range of the ADC. I programmed a lookup table to map the non-linear sensor curve, and to somewhat calibrate it, I looked up ozone that day (60ppb) and was all set to read all the way up to 1000ppb. Sadly, shortly after launch the reading pegged the needle at 1000ppb for the whole flight until just before landing when the levels dropped back down. It was not even an interesting graph to show here. I am not sure what kind of levels to expect at altitude and suspect either my sensor was just not happy, or I needed to measure larger magnitudes and this sensor just cannot do it. Anyway, I consider this experiment inconclusive for now.
So that was our SB-1 mission — thanks for stopping by!
(c) 2016 SurlEE