The Windsond software has matured a lot over the winter. For example, the software can now generate a range of different file formats for the weather data. The settings dialog of this is pictured below. All generated files can be viewed within the application, even while the sounding is still progressing.
Here’s a quick breakdown of the major changes:
- Better curve smoothing algorithm that preserves air layer transitions
- Altitude axes start at the current ground altitude
- Support generating nine different weather data formats
- Select generated data formats in the Settings dialog and inspect the result in the Report panel.
- Dewpoint and lapse rate are plotted
- Plot history of sensor readings and GPS/pressure altitude
- Winds are reported starting at 10m height
- Bugfix in improving cut-down altitude accuracy when using the default cut-down altitude
- Name of launch site can be entered by hand or selected in list
The software update is available without charge to all customers.
In Abu Dhabi, I visited the International Exhibition for Security and National Resilience (ISNR) on April 1. (No, it’s not an April fool.) This is a big trade show held every two years at the grand National Convention Centre in Abu Dhabi, UAE. The 400 exhibitors cover topics such as security, surveillance, terror prevention and counter-measures, emergency services like ambulance and fire-fighting, etc.
At the show I met with several interesting companies where Windsond would serve a valuable role. The small size and ease of deployment of Windsond makes it possible to gather local wind profiles in scenarios where this was previously unpractical. My goal is to continue exploring such alternative use cases in cooperation with these companies. I extend a thank you to exhibitors and fellow visitors from all over the world for the friendly reception!
I’m happy to announce that Kiwi Embedded has sold a Windsond system to Burkan Munitions Systems LLC, based in the United Arab Emirates. Burkan develops various types of ammunition and will use Windsond for gathering wind data during field tests. The company also uses a traditional weather sounding system and find Windsond to be a perfect complement for the lower atmospheric conditions. Advantages include portability, convenience, cost of operation and control over flight path.
A few days ago I conducted an on-site training with the professional and hospitable team. For the occasion I travelled to the UAE for the first time. We spent a day out in the desert for a hands-on walkthough of the system and two short soundings. They had no problem understanding how to operate the system and complimented the ease-of-use and set of features. I was happy to see that Windsond will fit well into their workflow, after I make a small software adaptation.
This marks the start of Kiwi Embedded offering services in on-site training. Please contact me to discuss your case.
Some data formats include a field for the ground wind speed and direction. At first glance, Windsond should be able to report this… but it can’t, since the wind is sensed by the balloon floating with the air movements (i.e. winds) as it rises. Before the balloon starts rising it’s anchored (typically by someone holding it…) and can’t measure winds.
But wait a moment, what exactly is meant by winds on the ground? Due to the physics of fluid mechanics, wind movement very near the ground will be much lower than expected, even approaching zero extremely close to the surface. The ground drag is felt up to 200-400m height, depending on the type of terrain. As it turns out, the standard is to measure ground wind at 10 meters height. Here’s another discussion, by Belfort Instrument. It takes Windsond several seconds to reach 10m height so the GPS has time to sense the change in speed and direction.
Still, the wind close to the ground is more turbulent than at altitude, so a single reading from Windsond cannot capture average conditions like a ground-based weather station can.
By the way, the GPS senses the movement by doppler effect. This is more accurate than calculating heading and distance between successive position reports — GPS position reports may experience sudden jumps and even at the best of times they have a degree of uncertainty that would severely limit the time resolution possible.
At the moment Windsond measures wind once a second and makes an average over the last three seconds when sending data to the ground station. A software extension could transmit all three values to the ground station. A future experiment is to measure five times a second and save this to on-board storage for downloading after the sounding finished.
From now on, Windsond sondes use a mini connector for the battery, and the battery itself is fastened by velcro. To charge the battery, remove it from the sonde and connect it to the USB charger. Previously, the battery was soldered in the sonde and the USB charger connector needed to extend down into the sonde encasing.
This change removes the concern whether a particular sonde is charged and even enables the same sonde to be launched multiple times in succession by changing to a fresh battery. The battery can be removed for repairs, or replaced in case it’s damaged from deep discharge. For different trade-off between weight and running time, other battery sizes can be employed.
Another issue was the strict shipping regulations for LiPo batteries that affect world-wide shipping since Jan 1, 2013 to handle the risk of spontaneous battery combustion that have occurred aboard aircrafts. The rules are complex and cause headaches for many businesses. Even though the Windsond battery is miniscule and hardly constitutes a risk for anything, limitations may that apply — depending on whether you turn to the Swedish Postal Service, DHL, UPS or Fedex. Separating the battery gives the option to dropship the batteries directly from China which bypasses the problem altogether since China has less strict shipping regulations than Europe.
The battery assemby is custom-made for Windsond and I’m happy about how this change makes the system more modular and easier to handle.
The past weekend some friends and I tried attaching a Windsond to a sky lantern. Sky lanterns are made of a light paper bag over a suspended piece of wax that burns to fill the bag with hot air, creating a small hot air balloon. Sky lanterns are very light-weight and don’t carry much fuel to keep them afloat — would it be able to lift a Windsond?
Winter had finally arrived in Sweden and it was about -6 °C with fresh snow covering the landscape. Beautiful as it was, a wind of 5 m/s made the open landscape less pleasant to experience this day. We huddled behind one of the cars to light up the sky lantern. The Windsond was attached with a thin thread.
After a couple of minutes, the sky lantern was filled in spite of the chilling wind and we let it go. It had a hesistant start, then rose by a full 2 m/s as the wind carried it away. It rose to 110 m AGL, then appeared to run out of steam and started to sink again. Unfortunately it never recovered and continued to descend back to ground level by about 1 m/s. The flight was over in 3 minutes but the sonde still travelled 1 km.
Since the flight was over so quickly, we didn’t get a chance to cut-down the sonde and we couldn’t choose the landing location. Knowing the GPS coordinates, with maintained radio contact and the sonde blinking and beeping, it was no problem to find the sonde after walking across a snowy field. The sonde must have been attached to the sky lantern all the way to the ground since the fall was so slow, but we didn’t find the sky lantern and may never learn what happened of it.
The sonde carried the new high-accuracy humidity and temperature sensor which performed perfectly. The sonde was also equipped with an extra memory that logged more data than the radio link has capacity to carry. The firmware was extended to log a number of parameters with timestamps; full GPS readings, different types of temperature readings inside and outside the sonde, pressure, battery voltage and error conditions (although none occurred).
All in all, the results are mixed. Windsond did prove light enough to be carried by a sky lantern. It’s not necessary to bring a helium canister and the price of helium and sky lantern is about the same. On the other hand, sky lanterns are sensitive to winds and high humidity. The open flame presents a risk of starting a fire on the ground. And not least, the altitude of a standard sky lantern is very limited.
A video clip of the launch:
Windsond now has the option for a sensor with especially high guaranteed accuracy and resolution. By placing the sensor 10 cm out on a light-weight boom, the risk of the sonde affecting the temperature and humidity measurements is minimized.
The sensor is Sensirion SHT25, with 1.8% accuracy and 0.04% resolution of relative humidity. Temperature has a typical accuracy of 0.2 degrees C and resolution of 0.01 degrees C.
The high-accuracy sensor is available as an option for all new sondes.
There’s an article about Windsond in the 2013 autumn edition of the Swiss Ballooning magazine Contact. It makes a detailed comparison between Windsond and the method of finding wind directions by manually aiming a measurement device, theodolite, towards a small rising balloon and sending the observed angles to a computer.
The author finds that the theodolite method lacks in reliability since it relies on the rising balloon to maintain a constant rise speed. As Windsond demonstrates, this is not always the case. There’s one disadvantage to Windsond — sudden changes in conditions between two layers of air are diluted by the smoothing of data samples by the Windsond computer software. In response to this, I’ve now developed a more sophisticated algorithm which detects and preserves sudden changes. After testing and optimization, this will be released to all customers.
The magazine is in German. See the Windsond article as pdf. The magazine also mentions how Windsond was used in the European Ballooning Championships 2013 (article as pdf). Or see the full magazine as pdf, from the Swiss Ballooning Association (SBAV).
The Mountain Meteorology Group at the University of Utah is evaluating Windsond for their meteorology research. During one test, the sonde failed to detach from the balloon. They continued to track the balloon to an altitude of 9700 m MSL, at which point the sonde battery was drained. This is close to the altitude of the tropopause, the important point where the temperature reaches a local low point.
The last message before the sonde shut down was transmitted at a distance of 49 km, where the sonde was still rising at 1.5 m/s and travelling at 17 m/s. This is a great result for the basic whip receiver antenna. The reception was becoming increasingly spotty towards the end and the contact wouldn’t be possible without the error-correcting code used in the radio link. The range could easily be improved a lot by using a directional antenna.
To reaching high altitudes on purpose, a bigger balloon would be used to gain a higher vertical speed and to allow the contained helium to expand the balloon further before the balloons bursts. A high-altitude radiosonde commonly use a balloon weighing 100 g, compared to the 8 g balloon commonly used for Windsond. Officially supporting high altitudes would also require proper testing of all components in a climate chamber.
I’m now doing all I can to avoid the mishap from repeating:
- Cut-down is done by heating a thin metal wire until it melts the wire tethering the sonde to the balloon. The duration of the heating will get progressively longer if the detaching fails.
- Before launch, the sonde will detect if the mechanism is broken to alert the user.
- There’s now support for backup cut-down hardware.
- If the sonde still continues to fail to detach it will eventually stop trying. That will save the battery from draining as it did this time. The balloon will eventually pop and with battery to spare it’s still possible to recover the sonde.
EDIT: Windsond has since been used at higher altitude.