This morning (UTC) we did it.
3.8.2 beta2 works well on our vectored tailsitter “Belly”.
(Backtranistion at 03:40)
Logfile 00000034.BIN Anmelden – MagentaCLOUD
Regards Rolf
1 Like
@lorbass I’ve never had the tether get caught in a prop, even when I had oscillation about the vehicle Z axis. All you have to do is throttle back if you get in trouble. A lightweight line with an automatic tensioner (very light tension) would probably be an improvement, but I don’t use this for fine tuning, only as a safe way to test for stability after major changes. As soon as it behaves reasonably well on the the tether, I go straight to free flight for final tuning.
Regarding the stability of this tailsitter in hover: it’s great as long as you don’t have any airspeed. But, as you observed, there is not enough pitch authority to counter the wing’s built-in pitch stability (and tendency to enter forward flight) after you pitch over and develop some airspeed. Larger elevons were not a complete solution either. And I was flying PX4 firmware on the Caipirinha, so it would be difficult to compare parameters, unfortunately.
Lastly, it usually takes me more time to find my tools than it does to perform whatever task I had in mind 
But someday I’ll get rid of all the junk and organize everything, maybe.
Hello Friends,
How we can set maximum spin limit of both dual motor for copter mode of dual motor tailsitter.
Thanks Rolf!
Looking at the log, I think that you should increase Q_TAILSIT_ANGLE from 30 to 50. That should help to bring the nose up a bit more before the multicopter controller takes over.
The other problem is that the throttle dropped too much on the transition to QHOVER. You had Q_M_SPIN_MIN=0.22, so it really shouldn’t have dropped that much. I need to look into what caused that.
Cheers, Tridge
I’ve now released 3.8.3beta3 with a fix for the low throttle issue in transition to VTOL from your log
Thanks for the test flight!
Hi Tridge,
We have you to thank for this great firmware.
Unhappily we don’t have much time this weekend. But we will proceed testing beta3 as soon as possible.
Rolf
no worries, it is actually a very small change, so I’m quite confident it is OK.
thanks for the testing you’ve already done!
Hi Tridge,
I am pleased to inform that the new Transitions in FW 3.8.3 Beta 3 works great.
After the good message of the new Transitions, I repaired the Wing in a hurry.
Log : https://drive.google.com/open?id=0Bxldq38bfbqHY1oyakZaVENnd1k
Only a littel unstable when Copter Mode takes over in QHOVER.
But really perfect in FBWA to QSTABILIZE.
Perhaps to improve by modification of a Parameter?
Thank you verry much, great Job.
Kind regards, Otto
1 Like
Thanks, Tridge and Otto!.. Can,t wait to try this …some repairs to do, field flooded, and new laser cutter showed up…working on a foamy TVBS kit…Tridge, Can the same transition be applied to entering FBWA as well?
Thanks for the testing, much appreciated!
looking at your log I think there are two more code changes needed:
- clear the roll/pitch integratators at the end of the FW → MC transition
- add gain adjustment to MC rate controllers based on throttle
The 2nd change may change the tuning a bit, so I’ll think about making it a parameter
Cheers, Tridge
nice! I’d love to buy one
yes, it has new transition logic for both entering multicopter modes and for entering fixed-wing modes (including FBWA). You can see several examples of entering FBWA mode in the video from @lorbass
Cheers, Tridge
[quote=“tridge, post:742, topic:15302”]
nice! I’d love to buy one
[/quote]It,s going to need testers, and your #1, and just now figuring out the laser and design…
will send a prototype as soon as I have one.and a heads up on parts you will need
[quote=“tridge, post:742, topic:15302”]
yes, it has new transition logic
[/quote]so softer entry into FBWA …video looks like still flipping into FBWA with no forward speed?a slower rotation available in tuning?alt gain is ok for me…
Thanks Rick
It can be tuned with the PTCH2SRV_RMAX and Q_A_ACCEL_P_MAX parameters, but given those also affect non-transition control I think we would probably make it obey the Q_TRANSITION_MS parameter as well, and change target angle based on that parameter. That would allow you to adjust the time for transition appropriately for the airframe.
thanks!
With the “old” transitions I tried to tune Q_A_ACCEL_P_MAX from default 110 000 in 4 steps down to 20 000
and couldn’t see an affect whether in the video nor in the log.[ATT(DesPitch,Pitch),PIQP(P,I,D,FF),IMU(Gyr,Acc)]
Log 43 with 75 000 and 40 000
https://drive.google.com/open?id=0Bxldq38bfbqHNXhQRFMzRDNVUGM
Log 44 with 40 000, 30 000, 20 000 and back to 25 000
https://drive.google.com/open?id=0Bxldq38bfbqHY0U3cEU1UHdpVXM
Regards, Otto
right, it wasn’t effective in the old transition code.
I’ve just made a change to address the issue that @tilt raised about having enough airspeed for mc->fw transition. I’m making it do the transition over Q_TRANSITION_MS milliseconds, defaulting to 2000 for tailsitters. That allows it to pick up some speed before it starts flying as a fixed wing. That also controls the rate it leans forward.
For fw->mc I’m still considering if I’ll do a separate rate limit.
I’ve also made changes to scale the surface gains by throttle, which really helps for FBWA->QHOVER transitions, and helps when descending rapidly in hover.
I plan to release 3.8.3beta4 shortly (probably in the next couple of hours) with these changes.
Cheers, Tridge
Thanks for continuing of the improvements.
Don’t hurry, 
we have very bad,cold weather with snow 150m above ground.
Kind regards, Otto
I’ve released 3.8.3beta4 with the changes for throttle surface scaling and timed fwd transitions for tailsitters.
It is now pretty smooth for me in the simulator, and I hope it will work as well for others!
It especially helps with fast descents.
One thing with beta4 is that it is now important that you set your hover throttle correctly, Q_M_THST_HOVER as this is used for scaling of your gains. So if it is too low then make sure you raise it to your normal hover throttle before flying
While Otto is shovelling snow, we had a rain break 
https://vimeo.com/241394758
2 Flightlogs from the filmes tranistions: 00000034.bin and 00000035.bin
https://www.magentacloud.de/share/ualbgv5hup#$/
We have to adjust PIds a little bit, but all is ok and tranistions are perfect
Rolf
1 Like
What command actions are executed during this time (Q_TRANSITION_MS ?)
Regards, Otto
So far it is only used for transition from hover to forward flight. What it does is to calculate the desired rate of pitch as the Q_TAILSIT_ANGLE divided by half of the Q_TRANSITION_MS. The idea behind that is you want to spend half the transition time getting to the transition angle. So what you see in flight is the nose comes down from hover over 1 second to the transition angle (default 45 degrees), then it converts to normal fixed wing flight from there. This gives the aircraft time to gain some airspeed, so it tends to produce a better forward transition as the wing has a bit of lift before the normal fixed wing flight takes over.
You can adjust it by changing Q_TRANSITION_MS up or down to suit your airframe.
Cheers, Tridge