Servers by jDrones

Holybro Durandal user experience in Bixler 2

I have been using the Holybro Durandal Flight Controller for over a month and I’m really impressed. My plan with this Durandal / Bixler build was to build a reliable and dependable fixed wing UAV that I could use as a tool to learn ArduPlane functionality and practice autonomous missions. If you don’t have time to read the rest of this blog post, the short version is that based on my experience of using it for this task, I completely recommend it.

Full disclosure; Holybro kindly sent me the unit to test and I’m a member of CanberraUAV.

I’ve been flying ArduCopter enabled Quads since 2013, but I’ve only moved across to flying fixed wing aircraft and ArduPlane this year. This is the third flight controller type that I have tried. The first was one of the cheap single IMU boards; it had a complete failure during my maiden; that led to the Dev team developing the Independent Watch Dog to prevent anyone else from having that problem. Next I tried one of the early Pixhawk clones; I found that temperamental and I didn’t really trust it.

The Durandal on the other hand feels like a high quality piece of equipment compared to the others. It arrives smartly packaged and I was immediately impressed. What was even more pleasing was that it came complete with HolyBro telemetry radios and a power module. As someone that doesn’t yet have a near infinite supply and variety of leads and cables, I find it a huge bonus to get everything that I need from one supplier, so I know that it will all fit together and work. The only other thing that I needed was an RC receiver, so I added one of those.

The Durandal is a snug fit inside the Bixler 2 airframe. I shaved approximately 1mm of foam from both sides of the rib inside the fuselage behind the cockpit. This allowed me to slide the Durandal in easily, and it now fits flat behind the cross-beam on the floor at the rear of the cockpit. As you can see, the fit is quite tight.

One of my bug bares with these flight controller kits is the really short wires for key parts of the system such as radios and GPS. These are often fine for small quads, but they can be difficult for fixed wing applications. I like to keep RF emitters and receivers as far apart as possible, but the length of the included wires forced me to keep them quite tightly packed around the cockpit. If I was a supplier of these flight controllers and ancillaries I would probably add a number of cable length options in my shop. Based on this constraint I mounted the telemetry radio laterally across the cockpit, with the antenna mounted through a hole in the side of the fuselage under the cockpit. I was worried about potential RF interference, but I haven’t had any issues with this set-up.

Imgur

As you can see below, I mounted the GPS unit in the cockpit cover. That allows me easy access to the safety switch that is built into the GPS unit.

Imgur

I have now flown about 30 take-offs and landings with this set-up, most of them fully autonomous. The video below gives you an idea how reliable I am finding it. I don’t have an airspeed sensor fitted, but I am finding that the accuracy of the auto land is really impressive!

Imgur

7 Likes

nice one, thanks Jimmy!

Thanks for share Jimmy

Those are very good landings! It can be really hard to land one of those powered gliders manually as they want to just keep flying until you almost catastrophically kill the airspeed.
That pretty much makes a FC and Ardupilot a requirement in any fixed wing in my books :slight_smile:

Thanks @xfacta . When I was preparing to try auto-land I watched Tom Pittenger’s video on how to auto land. The biggest takeaway for me was the need to understand your glide slope. I spent a couple of missions in FBWA collecting data to do just that. I flew multiple iterations of of ‘climb to at least 60m, zero throttle and glide straight’. Then I could assess the glide slope after the mission from analysing my logs in google earth.

That way I had a good idea what height and range to mission plan my landing cycle. In this case I used a turn onto finals at 200m from the landing point at 60m AGL, followed by another point at 100m / 30m AGL. These landings are the result of that work.

I’m currently testing to see what difference flaps make to the glide slope so I can shorten my landing approach. I’m planning to put all of this in a blog post when I’ve finished.

Nice write up. Can’t wait to have a peek at it next time I’m in Canberra.

1 Like
Servers by jDrones