Servers by jDrones

Plane 3.8.0 released

(tridge) #21

I have released 3.8.1beta1 with a few small fixes from 3.8.0. If flight testing this weekend goes well I will release 3.8.1 on Sunday.

List of fixes here:

(tridge) #22

The method we've moved to for complex MANUAL mixing setups is to allow a set of channels to be pass-through from RC input in MANUAL. The parameter to control that is MANUAL_RCMASK. That is a mask of channels which are direct pass-thru in MANUAL mode only.
So basically this allows you to setup complex mixing on your transmitter that applies only in MANUAL mode, and set MANUAL_RCMASK to which channels you want to be pass-through.
For your case with differential on ailerons, just set both ailerons in MANUAL_RCMASK and then setup your TX to do whatever mixing you want when your mode switch position is set to MANUAL. That is fairly straightforward for something like a Taranis.
Cheers, Tridge

(wkf94025) #23

Tridge (or other ArduPlane 3.8 devs),

I am building a MyTwinDream for autonomous mapping, using Pixhawk 2.1, Reach RTK GPS, Sony RX100 mapping camera, Dragonlink V3A, Tuffwing hotshoe cable, and SeagullUAV Map2 for firing the camera. My ultimate goal is mapping ~1,400 acres in a single flight with cm-level accuracy in the resultant contours, as well as a gorgeous orthomosaic and 3D mesh of the property. Power system is two Cobra 3420-14 850Kv motors with a pair of Cobra 40A Opto ESCs, powered by a 4S Li-ion for multi-hour endurance. Two questions for you re Arduplane configuration:

1) I plan to go dual elevator servos, for redundancy. How would you configure those in ArduPlane? My goal is not snap roll type aerobatics, but simply the ability to survive and get home should either elevator servo give out, or I should find myself in a steep dive above Vne type speeds, and need a bit more torque/pull on the elevator.

2) I am intrigued by separate motor channels for this twin. Obviously I could go single motor channel, and just Y-cable the motor signal to both ESCs. But it might be fun to play with separate channels, particularly if I put separate Mauch Hall sensors on each power train and have real-time visibility to mAh consumption on each side. That and playing with differential thrust mixed with rudder, perhaps on aggressive "crop duster" missions. :wink: On the other hand, why risk ~$2k in mapping payload to science experiments on relatively new code surrounding multiple motors? I welcome your guidance on this.


(tridge) #24

just set two SERVOn_FUNCTION values to elevator

I think you'll find separate motor outputs works well
Happy flying!

(wkf94025) #25

Thanks for quick reply. I imagine your massive accomplishments on quad planes and tail sitters would make a vanilla twin motor a walk in the park. Will report back soon.


(Trias) #26

A few transition from FBWA to QHOVER and BACK to FBWA with some stick fling in QHOVER in between,

3.8.0 is Solid, thanks Tridge

(tridge) #27

good to hear! thanks for the testing

(Rjiun Raj K) #28

how you tested the landing? where we can find the landing code in ardupilot?
i have Ranger-EX, I compiled Arduplane code. How to do the landing using the code?

(Augusto Iturri) #29

(Rjiun Raj K) #30

can you please tell me, how to implement this on real plane.

(Rjiun Raj K) #31

can you please tell the procedures and code for auto landing.

(David) #32

Admirable goal. Not sure you’ll be able to accomplish this without GCPs. In the past year, I’ve done several tests comparing results with and without GCPs. For the moment, I’d say it’s impossible to get cm level accuracy without GCPs, despite what some vendors claim. We’ve been using PPK data from a Septentrio AstREX with a hot shoe for trigger feedback. We are GEO tagging images with data from the GNSS receiver and not from the pixhawk log files. I don’t know of a way to get better EXIF data to the photos. We’ve never come close to centimeter level accuracy without GCPS. Of the top of my head id say x and y are around a one ft and z various quite a bit more.

Also, if you want to cover 1,400 acres I am assuming you’ll be flying at the maximum legal alltitude, which would typically mean your GSD (ground sampling distance) is greater than 1 cm/px. So your goal may be elusive.

Why do you need cm-level contours?