Today my Senior Telemaster just passed 2000 minutes of autonomous flight time (now at

Today my Senior Telemaster just passed 2000 minutes of autonomous flight time (now at 2019 minutes total autonomous time for the life of the airplane.) It flies with a hybrid autopilot system consisting of an APM2.5 + a Gumstix linux computer. On an 8000 mah battery, typically I get 40-50 minute flights. The flights have covered 1100 miles (1780 km) total distance (all of that within a tight looping pattern and within controllable line of sight.) This telemaster first flew in 2007 and has seen a few dings and hard landings due to windy conditions, but overall is still in great flying shape.

Congratulations!
To my eternal shame, I bought a gumstix for a drone project, but after struggling to find a way to mount and protect it, I never got around to switching it on. :frowning:

I have plenty of hardware sitting around on my shelves that never really got used for anything, but I still keep it just in case … it just goes with the territory. If we knew everything when we started, we’d buy only what we needed and make perfect use of it, but what fun is knowing everything already? :slight_smile:

Is this a kit? I can’t find kits anymore, I always thought this would be a perfect platform.

This is the older Senior Telemaster ARF (no longer made). It is sold by Hobby Lobby (err Hobby Express is their new name). They have a newer version of the ARF available with some upgrades (plug in wings, flaps, probably lighter…) But this one is a nice big gentle boat. 8’ wing span … flies slow, hauls a lot, very forgiving, pretty rugged. It’s been a great platform for testing and developing. The fuselage is recovered in all red because it came damaged in shipping. I got a replacement fuselage on the way, and then started looking at this one … .turns out it was a quick easy repair, but I couldn’t match the yellow covering, so I pulled the rest of the yellow of the fuselage and recovered it all in red. It’s kind of unique in that it’s big, it’s made out of wood, it’s electric powered, and it’s just slow and gentle and stately in the air.

I have wanted one if these for some time. Slow powerful and flexible. I find myself more relaxed with models like this, elect is even quieter than 4 strikes if yore.

What do you use the gumstick for? Monitoring?

Hi Matt, I stripped the APM2.5 firmware down to the nubbins and just use it as a sensor data collector and servo driver. All the sensor data gets piped up to the gumstix (overo) which runs a 15-state EKF and all the guts of the autopilot there. I’ve been running various evolutions of this basic autopilot system since about 2006-ish, starting back with an original gumstix and a xbow micronav. That’s what original flew this telemaster, but I’ve been upgrading this and that over the years.

What’s the EKF estimating? uvw,pqr,XYZ,ptp… ??

Whoop whoop beer time!

EKF: given gps + imu, estimates position (lon,lat,alt), velocity (NED), attitude (phi,the,psi), plus gyro bias (3 axis) and accelerometer bias (3 axis). Based on a matlab algorithm published in a book contributed to by a professor (Dr. Gebre) at the U of MN.

Ah, Gebre-Egziabher. Yes thanks, I found him. Thanks @clolsonus , I’ll look into that. Sounds like a very useful EKF.
Good luck with your UAV!

On the plus side it doesn’t use a magnetometer so it’s not subject to all the interference, calibration, and noise issues a magnetometer brings to the table. But on the flip side, yaw is a low observable parameter when you have only a gps and imu, and the algorithm requires some amount of acceleration to converge in yaw. In aircraft it really works great, but I haven’t tried it in a copter.

Mm, I can definitely see advantages of eliminating Magnetometer from sensor suite. Do I understand correctly that the acceleration required for convergence is a disturbance from the steady 1g acceleration up in Earth-axis , required to sustain ‘rectilinear’ motion, and also that the algorithm makes no assumptions regarding the vehicle to which the IMU is attached?

Right, it’s my understanding the algorithm needs acceleration in the gps’s frame of reference in order to converge. Typical straight and level flight usually provides enough acceleration for the algorithm to work well. Yaw angle converges relative to true north (not magnetic north) which can have some advantages for certain applications (like camera pointing.)

You are light years farther than me. I just got my first quad flying last weekend. I miss the planes though. Sitting there circling nice and slow.

My first was a “gentle lady”. Made a little power pod with a killer bee 49 on top.

Matt, long ago I had an Ace High Mk II that was 049 powered … good memories! http://gallinazo.flightgear.org/category/model-airplanes/retired/ace-high-mk-ii/

Someday I need to get up to speed with quads … I’ve only done airplanes so far …

Easier than you would imagine. Start off with a kk2.1…easier to get going than the APM (recent experience). It is smooth and quiet, a real pleasure in my back yard.

You already have the radios. I bought three platforms and the x545 from ebay was the easiest. If I were to do it all over again my quad would have cost 200 bucks max. I got a bunch of stuff I didn’t need instead.

I don’t see my self auto piloting the quad. The range seems low, but the APM will go on something soon. It is a really cool platform.