As this blog still generates a small amount of traffic, I am posting this notice to acknowledge that all of the content in this blog is several years old and the project had now been abandoned.I don’t intend to resume progress on the project in the near or forseeavle future. I will leave the blog up for now as it may have referential value for people pursuing similar projects, even if the information is out of date.
Recently, I was able to take the Bixler to Irvinebank, and take some footage of the burnoff firefront. These images were pulled out with Blender, and stitched together with GIMP to make the panoramic image of the fire.
The landing after taking the shots was less than desirable, resulting in a cracked horizontal stabiliser. Trees surrounding a short strip made the the angle of descent greater, which resulted in a fast and hard landing. Hopefully I’ll be able to use FPV gear next to land and launch from a more appropriate location, and also get a live video feed for observers to watch.
Unfortunately, FPV gear that had recently arrived was faulty, having trouble maintaining a solid connection at any range. Next time…
The final render of the Bixler model.
I decided to make an application showing the orientation of the plane. At this stage, I have only had the option of a stream of rotation data or a box 3d display. Neither of these really gives the picture of what the plane is doing, so I decided to get a model of the bixler. This turned out to be hard. There was not a single sky surfer or similar type airframe. Anywhere.
So initially I was looking around for easy to use modeling software, and surprisingly found some free Autodesk software, 123D Catch. It works by taking in a series of images, finding some common points, and then (guesses and) plots points for a 3d model based on the movement of those points relative to each other. More Info
This GIF shows all the images used to make a model using 123D catch.
After feeding the sequence into Catch, a wonky, disproportioned, holey and automatically textured low effort model was born. Some people have been able to make stunning 3D models with the software, but I was unable to achieve anywhere near the standard they had set. Maybe this could have served the purpose, but really, thats just not professional.
The model produced by 123D catch.
After discounting the use of 123D Catch as a tool to make a realistic and proportional model, I decided to look into some other options, which soon led to Blender. The free, open-source 3D modeling and animation program provided everything needed to create an electronic copy of the Bixler.
After a few helpful tutorials from BlenderGuru, I was ready to start the modeling process. The first take at modeling was really just to get a feel of what I would be doing, and an idea at what I wanted the end product to look like. This resulted in a strange looking curved wingtip plane thing. Great.
The second Go was for serious, using pictures of the plane as a structure to go off. Some techniques I had used didn’t really work that well on the real thing, which resulted in a lot of tiny gashes in the nose of the plane. This model looked quite convincing, though the perspective in the photographs made matching the model to the image nearly impossible.
The second attempt at modeling the bixler, this image is taken so the holes in the canopy can’t be seen
The third attempt began with the capture of a number of photos of the plane. These were then combined (using GIMP, another great free program) to reduce the effects of perspective. Sensing this as my last attempt, I captured a time lapse of the whole thing from here on:
So basically in the video I arrange all the template images, model the left view, top and front, finish up with some tweaks and do a subdivision (make it smooth). After the modeling that shape, I switch to a windows machine, unwrap the model in blender and use GIMP with the stitched images to make a copy of the stripes and text. After adding a quick foamy type texture, and a short animation to showcase the model (kind of got carried away, seen as it was originally for a 3D display) and a high quality static render, the Bixler model was complete.
Anyone wishing to download this FREE model can do so over on TF3DM: http://tf3dm.com/3d-model/hobbyking-bixler-sky-surfer-89447.html
Hopefully helped out some others that were like me, just looking for a good, free, Sky Surfer airframe 3D model.
Beginning work on the pc display, will try to post some periodic updates on progress.
Check out the facebook page: https://www.facebook.com/pages/Arduav/154892691342263
Have a look at the Youtube Channel: http://www.youtube.com/user/Arduav?feature=watch
I have been trying to find a good font to use, just to be recognisable and not so dull, when annotating diagrams and other things like backgrounds and web-related imagery that really needs an easily recognisable font.
Anyway, I have had trouble narrowing it down to one really good one, though I managed to get it down to three. Please chuck a quick vote, on what you would like to see around the project. Thanks!
Carbon Fibre rod, increasing the strength of the fuselage
Recently I crashed the bixler, resulting in a broken tail, damaged battery and cracked nose. The causes have now been determined to be as a result of alterations to the fuselage which weakened the tail section. This has now been strengthened with the addition of a Carbon Fibre rod running under the fuselage.
Modifications to increase airflow internally weakened the tail and caused it to fail, possibly in flight. In addition to the loss of tail control,
Shown are the two modifications that have increased the internal airflow. The perfect cross section (Left) shows the two holes drilled to allow air to flow out of the fuselage, which significantly decreased the strength of the tail.
ailerons were also disabled,after the servo lead disconnected, possibly because of the increased airflow. Connection issues, experienced after the crash also have been resolved. The connection issues, experienced after this crash have been attributed to loose and frayed wiring on the battery.
Bixler motor mount shaved of excess foam to increase airflow immediately surrounding the motor to help cool the motor
Bench tests were recorded, though were uneventful and about 15 minutes long each and therefore will most likely not be posted because of this. A speed build was also recorded which will likely be posted, along with some internal camera footage to monitor the airflow mods.
Look out for some videos and more posts! I have been busy doing things on the project and have spent less time actually documenting.
Over the past few days, I have been testing sensors, particularly ones to be used to log all info I will need from the plane. With some DS18B20s, a DHT11 and the Y-30 SD module, there will be a lot of data gathered.
The Arduino will log to a .csv formatted file on a 4gb SD card, which excel will open up easily, and allow creation of graphs. This is the first release, with some minor improvements in the future. Best to move on to the stabilisation side of things, being more important than a weather station.
Check out the project on GitHub: ArduinoWearherStation
A video, of the best shots trying to get an aerial view on a small fire
A recording micro video camera was used to capture the video, no FPV yet.
The pan jig used was a simple servo and bit of foam, with 6 rubber bands and a bit of wire to hold it all together. Works well, but is interestingly hard to concentrate on where the camera is pointing, as well as what the aircraft is doing.
Got the Uno to send back some live GPS data, this time displayed as a converted KML track. Took this live, but interference from both rc TX and the xBee made for a REALLY bad connection. Seriously had a better connection going through two houses. Shows that the maximum speed achieved was 92.2 km/h. Probably a result of connecting the gap between two disconnected routes, but seemed steady on the graph. The flats are averages calculated by Google Earth to fill in the gaps:
Today finally got Earth Bridge to work, giving a live feed to Google earth. Earth Bridge is a little bit dodgy, but after debugging with full error messages on Google Earth I was able to get it to work for me using these settings with Google Earth 6:
Working settings for Earth Bridge
The weather was cloudy, so had a dodgy gps fix as you can see in the screenshot it appears I went mining before a jump with my laptop out into the paddock. When the system works, it is a great way to display the data though.
Next step to put this through the xBee and have a live stream from the plane!
Arduav 