Well, it is about time for some more updates on the Mugin. I'll try to go point by point with what has changed since the last post.
The biggest news is that it has flown! We have done a few hours of flight testing including auto flight and everything seems to be working well. With that vibration damped avionics tray seen earlier, there is actually less vibration reaching the IMU on this aircraft than on our electric planes... We have done some early testing on fuel burn and the results have been a little disappointing, but we are re-tuning the fuel injection and may get a little better performance out of it. As it stands, I think we will get ~3-4 hours of flight time. With proper flight planning that shouldn't actually be much of an issue. And, flying longer than that means our 256GB SD cards for the A7ii won't be big enough and we'll have to find another way to store the images.
 |
| Waiting to fly |
 |
| A little photo shoot |
Now, some bad-ish news. A pinhole leak somehow found its way through the plane over last weekend. We discovered it on Tuesday when we pulled the canopy off and found the fuselage filled with about half an inch of fuel and all the printed plastic parts melted into goo. Thankfully, no electronics appear to have been damaged as they were all mounted on their elevated platforms. After a couple days of brainstorming, I ordered a pair of the 100oz tanks we use in Mugin 1.0. It isn't the perfect solution but it will get us flying until I come up with something better. I'm thinking maybe a custom aluminum tank...
 |
| We didn't get any pictures before this point, we were rushing to get everything out of the fuel |
 |
| Goo |
 |
| Stripped bare |
The good thing that came of this is that it forced us to strip the aircraft down to the airframe and start anew. That means we were able to go back and rewire and clean some things up that we probably would not have bothered with otherwise.
 |
| Pretty wiring |
 |
| More pretty wiring |
 |
| A beautifully designed JST power distribution block |
So all in all, it wasn't all that bad. I'm just glad we didn't have that fuel leak down in Peru...
Ok, on to other things. The alternator and power system arrived a couple weeks ago. There were some fitment issues and growing pains but it is all installed now and it works flawlessly. It provides clean 7 and 12V supplies to the aircraft that are backed by the alternator, battery backup, and a shore power connection.
 |
| That alternator sure does look good... It almost looks like a big brushless motor (which, I suppose, is precisely what it is) |
 |
| The first motor test with the alternator installed. |
Also, we have done some testing with the new IP radios, including range testing. With the 23 dBi antenna and the amplifiers we were able to get a mediocre link at 85 km. We have ordered a 30 dBi dish that should give us the mythical 100 km, or at least get us close. Hopefully we will have that soon so we can get back to range testing. In the picture above, you can see the amplifier mounted fore of the avionics stack. And, I'm also about to get to work on the antenna tracker. Version 1 wasn't robust enough so we go a Yaesu G-5500, which is way overkill but will do the job. I just have to build some control circuitry for it and interface it with the Pixhawk.
Finally, a teaser. I've been working on a companion computer to simplify the electronics inside the aircraft and provide us with some computational horsepower if we every have applications that need it. Using an oDroid XU-4, the core functionality is:
- Forward Pixhawk Mavlink stream over the IP network to the ground station(s)
- Pull images from the Micasense RedEdge and save them to an SSD
- Stream video from the nose camera over the IP network
All of these functions have been bench tested and are working. I expect a couple issues will pop up when I install the system in the aircraft but it shouldn't be too bad. The video is great, with only about 0.5 second lag and great quality (especially considering I've limited the bitrate to 1 Mbps). In addition to these 3 functions, we have a few more that will be forthcoming:
- Data logger function for (non-avionics) sensors
- Payload control
- Forwarding of auxiliary telemetry (motor health, fuel level, etc)
- Saving data from RGB camera (maybe Sony, maybe something else)
- More?
I hope to soon be putting together a tutorial on how to implement those 3 core functions on the XU-4 as there isn't a good one yet. Further companion computer discussion will have to be saved until then. I have waxed on long enough.
