R10 Maiden flights and other videos!

Until this weekend, we have been testing our newer frames with our older Forebrain/Seraphim flight controllers.  But this weekend marks a milestone in R10 development with the first time we were able to fy the full R10 system (the R10 frame with the newer Thalamus flight controller).

Above is a video of a full R10 system flying Thalamus at 800Hz quaternion/attitude update and 140Hz PID loop.  In this video, the old Forebrain/Seraphim units were still attached to allow us to quickly switch over and compare performance.  In the final systems the battery is mounted top-side instead, and the cables are tidied away with the white lycra cable tidies.

And below are videos of the first outdoor test (excellent test flying conditions with no wind) at 800Hz quaternion/attitude update and 400Hz PID and PWM output.  We've not got the tuning just right yet, the yaw is overtuned in this video, and oddly there's some instability during descent which was not observed in the R7 ROFL systems running at much lower rates (possibly something to do with mounting the battery top-side).

At the end of the above video, Henry misjudged the landing and flipped the quad.  He should have trusted the pneumatic landing dampers, which are there specifically for the purpose of allowing safe and smooth landings.

 

On a side note, we also got the prototype static camera mount frames, and are in the process of testing this.

20121101_110023.jpg?width=400

We also bought one of these FPV pan/tilt mechanisms to play around with, with the idea of turning it into a stabilized gimbal for a very small camera (some of our Kickstarter backers suggested the HackHD)

In any case we're developing a gimbal mount specifically for the R10, here's a quick mock-up of what it would look like,  note that this is a mock-up/proof-of-concept, and it was thrown together in about 20 minutes from scrap aluminium frame parts, and there was no design considerations for vibration.  The one that we do design will of course deal with such issues.

Some people have asked whether side-mounting cameras like this is an issue, I think most people here know that a small I-gain in the PID loop will dynamically re-balance the quad for uneven loads like this.  We weren't going to show this following video until we found a pair of giant scissors so that we didn't have to chase the quad around to cut the string, but here it is anyway for your enjoyment.

Not much else to mention at this point, there's plenty more information on the Kickstarter page which will soon be winding down, and we're nearing the 1000% funding goal, we might just make it!

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • no, it's not a kind of DCM, it's a quaternion,  which uses different maths to the DCM.

    Our old code used DCM, when we compared the performance they are very similar, but the quaternion requires fewer calculations.

  • Ok, so it's a kind of DCM fusion algo ? No kalman so ?

    Did you compare your fusion algo to DCM implementation ?

  • Thanks Stephane, we're using quaternions, we actually posted some of the flight code that we're working on here: http://uair.co/content/r10-and-thalamus-full-system-miaden-flights you can take a look if you like.

  • It appears to be very smooth, congrats Max...

    What sensor fusion algo do you use to achieve 800 Hz ?

This reply was deleted.