@hwurzburg did you ever deal with uncommanded yaw on your boxkite? I don’t remember seeing anything in your blog post about it, but I could’ve missed it.
no never, but I also never flew it close to the thrust speed either…copter tail sitters lose control authority the faster they fly in fixed wing modes…and my vtol yaw was stable, although not rapid
Huh, that’s interesting. I didn’t fly mine near full throttle either, but definitely had the turning tendency. I wonder if the physical build I did was slightly off and that’s throwing things out of whack/trim.
Yeah, 7 and 8 would probably be best since you do have control surfaces you can leave them under the fixed wing controller. Forcing assist means the VTOL rate controller is used with the fixed wing controller slaved to it. In normal VTOL modes the VTOL rate controller also outputs to the control surfaces. The “Motors Only” option means that the VTOL rate controller is only used for the motors with no I term. The allows the control surfaces to do most of the trim hopefully reducing the tendency for motor saturation.
Thanks again @iampete . I am hoping to do a flight test of this today, hopefully I’ll be able to report back with success! 
Success! Thank you for the help! With bits 7 and 8 set from Q_OPTIONS, we have stabilized yaw and can turn in both directions “normally”. Q_OPTIONS_BIT_7_\u0026_8_applied_maiden.bin
The next questions for me are related to QuickTune and also adjusting the params for less of a “ballooning” in the FW to VTOL transition. I sure appreciate all the help with this, I am definitely learning a lot.
In terms of QuickTune, the vehicle is so light and suseptible to breeze that I think it will be somewhat difficult to find a good time/weather to peform Quicktune in QLoiter outdoors. It just gets blown all over. Would there be any reason why I couldn’t perform the Quicktune indoors using QHover (and accounting for lack of GPS fix)?
Additionally, if I’m understanding this section correctly - Tailsitter Planes — Plane documentation related to the FW to VTOL transitions, I can adjust the Q_TAILSIT_THR_VT parameter separately from the Q_M_THRST_HOVER to dial in a more level transition, correct?
The Q_M_THRST_HOVER param seems to be in a scale from 0 to 1, where Q_TAILSIT_THR_VT is in a percentage, so yesterday I did some tests where I adjusted the _Q_TAILSIT_THR_VT around a bit starting at 20% (and then trying 15% and 25%), because my Q_M_THRST_HOVER was about 0.2. I’ve attached those three logs here in the folder linked below, and it would be great if someone could confirm I’m on the right track - I saw a much more level transition yesterday on log Q_TAILSIT_THR_VT_25 but I’m not sure if it was due to me adjusting the value, or just coming into the transition much slower, or both.
You can tune in QHover outdoors, just let the vehicle drift with the wind while it is tunning.
Correct, for less altitude gain you want as low a Q_TAILSIT_THR_VT as you can. There is a bit of a trap here, the percentage is in terms of forward flight throttle. So 15% is actually 0 because Q_M_SPIN_MIN is 0.15. 25% is 7%.
Looking at the 15% log the transition times out because the fixing pitch up fails.
This is also the case in the 25% log:
Although here it does a better job of not timing out.
If the motors have plenty of control authority you can usually the best results with Q_TAILSIT_THR_VT 0 and then increasing Q_A_THR_MIX_MAN and Q_A_THR_MIX_MAX. That is effectively saying “aim for 0 throttle but use as much as you need”. Throttle mix values can go above the max range of 0.9. I have used values of 2 in the past. However, don’t go straight to that change them in increments of 0.25. The downside of too much throttle mix is that the vehicle will think it needs to boost throttle to maintain attitude control all the time and the high average throttle results in a flyaway. Its a good idea to change only one at a time, for example if you change Q_A_THR_MIX_MAN that will effect QStab, so if your seeing issues you can fall back to QHover where Q_A_THR_MIX_MAX is used.
Once you have the pitch following the target nicely and the transition completing without timeouts you can look changing the rate of the pitch up with Q_TAILSIT_RAT_VT.
@iampete Thanks for the reply!
I’m glad to know I can tune in QHover if needed. Managed to find a calm day, so got it done in QLoiter this time.
This is really interesting - did I totally miss that in the documentation somewhere? I had no idea it was in relation to forward flight throttle.
We were doing some testing a week or so ago before I knew about what you were mentioning, and we ended up (between 4 of us) doing multiple flight tests and arriving at or around Q_TAILSIT_THR_VT = 25, Q_TAILSIT_ANG_VT =25 and Q_TAILSIT_RAT_VT = 90, which gave us some transitions like the log I’ve attached here (a similar but not the same aircraft as other logs I’ve posted.)
I’m not sure about
but I think we’re there at least in pitch and roll on this aircraft, so if I understand what you’re saying, we should reduce Q_TAILSIT_THR_VT to 0 - which would mean the aircraft uses “aim for 0, but use as much as you need” throttle in back transitions, not Q_M_THRST_HOVER, which is was what it would use as throttle before we started using Q_TAILSIT_THR_VT at all. Is that correct?
Then we start adjusting Q_A_THR_MIX_MAN or Q_A_THR_MIX_MAX to affect QStab or QHover and try to find a value where it can follow pitch better in back transitions? How should we determine which, MIX_MAN or MIX_MAX to use first? This definition is a bit confusing to me - what determines “active flight”? Some level of autonomy? Q_Hover the least autonomous mode where MIX _ MAX is used?
After writing this out, I think I’m looking for a sequence I can walk through (though I understand that it may be iterative). What I’m coming up with is:
Q_TAILSIT_TH_VT = 0
Q_TAILSIT_MIX_MAX OR _MAN = increase in increments of 0.25 and perform back transitions to see where we start improving pitch target tracking.
Maintain Q_TAILSIT_THR_VT = 0 and
Increase rate of Q_TAILSIT_RAT_VT as needed.f
Thanks for the explanations and for taking the time to respond here. I really appreciate it!
“Normal” tailsitters you would expect it to be forward throttle because its forward flight, when you always have the VTOL motors active its a bit of a edge case. But your right we should document it better.
Yes, that is correct.
“Active flight” means automatic altitude control. Which you should change first depends on how you would prefer to recover. Would you rather do the test in QSTAB and recover to QHOVER if there is a issue or test in QHOVER and recover into QSATB.
Thank you @iampete - that is really helpful. I know a boxkite isn’t exactly a normal airframe, but I’m looking forward to trying to get it to fly as well as I can.
Just for fun here’s a video (complete with goofy music) of us flying these together. They kind of remind me of some type of Star Wars craft.
Well, I’m back for some more learning this semester - we had a crash on first attempted transition earlier this week and I was hoping I could get some help with log analysis, as I’m quite new to VTOLs in general. I compared parameters between the aircraft that crashed and one that I flew on the same day without issue and didn’t see anything really out of place. When watching the flight live, it seemed somewhat like the transition to fixed-wing mode was too slow and then the aircraft perhaps stalled out and got into a spin of some type and couldn’t recover. We never had this occur to my knowledge with our builds last semester. Any insight would be great! Thank you.