Questions and observations from my Quadplane maiden

ArduPlane Plane 3.6
1

There doesn’t appear to be many Quadplane flyers yet, so its still difficult to learn from other’s experiences. I hope the moderator creates a category for Quadplane soon.

I successfully flew my new Quadplane , and I’m very impressed with its functionality. Thank you Tridge for making this possible.

Here is a Landing: https://youtu.be/lERyfchTvqo
Here is a Takeoff: https://youtu.be/f81n0aXdvKc
(these don’t match the logs below)

My Quadplane is a conversion of the 3DR Aero using 1250KV motors and 9x4.5 props on 4S batteries. I tried to keep the props small to minimize drag in forward flight. I’m hovering at about 65% throttle, so I will need to increase the prop size once I add my payload.
I noticed that it is quite difficult to control the heading in gusty wind with the weathervaning effect. At one point while in QHover, the plane almost turned around to face me and I struggled to get it heading back into the prevailing headwind. I reduced Q_WVANE_GAIN from .1 to .05 and that helped.
Question: Why does the rudder not operate in Q modes? Wouldn’t it help to yaw the aircraft using the wind?
I learned that it is better not to fly backward like I would with a copter.

I used the new Transmitter Tuning method which was well thought out.
Question: I don’t understand the Tuning Error messages I kept getting. Mission Planner would say “Tuning Error 3.00” over and over and wouldn’t shut up until I disarmed the plane. What was it trying to tell me? I don’t feel my tuning attempt was very successful as I just can’t get feel for what it is supposed to fly like. She’s a big lumbering giant with sloppy control, but then again there is a skywalker strapped to this quad!. I’d love any advice on my PIDs that anyone can offer.

As expected the extra 1kg of weight and quad parts caused a lot of drag in forward flight. It normally requires about 110W to cruise at 15m/s, and now it takes about 150W, so that’s about a 25% penalty. I wasn’t ever able to find someone to give me an estimate on the penalty before I built it, so there it is. Let me know if your experience differs.

I did some auto flight patterns using VTOL_TAKEOFF and VTOL_LAND. For the most part it did very well. I realize looking at the logs that I forgot to update the Yaw parameters on this plane and the Skywalker is notorious for its horrendous sideslip at high power settings. This is a creating a problem for the Quadplane because the quad motors are working too hard to take out the yaw rather than the rudder. You can see in my short .bin and .tlog that during takeoff the plane is in a 15 deg bank angle, while the rudder is only providing damping. I’ll try to fix that tomorrow. The takeoff went very well. It is a bit disconcerting to see the plane weathervaning around as it climbs vertically, but it did what it was supposed to. The Landing descent was too slow considering I only have 2 minutes of VTOL battery, so I increased Q_WP_SPEED_DN to 250.
Question: Why does it take so long to start descending after it transitions? Is there a parameter that controls this? Does Q_TRANSITION_MS only affect the takeoff transition?

Thanks for any advice you can provide. I’m sure I’ll have many more questions tomorrow.

Quadplane is really incredible!

Here are the .BIN and .TLOGs

1 Like
2

Thanks for your post, very helpful.

3

@iskess,
Good to see your quadplane is flying!

That is a surprise as Q_WVANE_GAIN doesn’t do anything in QHOVER or QSTABILIZE modes. I’ve deliberately kept those modes as simple as possible so they can be used as “recovery modes”. If you want to weathervane while hovering you need to use QLOITER mode.
If you are finding you don’t have enough yaw authority in QHOVER mode then you can increase the yaw rates gains (the Q_A_RAT_YAW_* parameters). In particular increase the P and I gains, and the IMAX.

you are right that we could use rudder in QHOVER and QLOITER, although I don’t think it would be very effective at low speeds. It is probably worth a try though. The trick will be getting the gain right. I’ll see if I can come up with a reasonable approach.

That error means that the ratio of the RMS P term to the RMS D term has dropped below 3.0 while tuning the P or D rate terms for roll and pitch. It is based on the idea that it is bad for the D term to play too big a role in controlling an axis.
In your case you were getting this error even when sitting on the ground in a non-VTOL mode. I’ve just pushed a fix for that to master, so it won’t report an error if the VTOL motors are below 10%.

Interesting, I have heard smaller numbers for the drag, but 25% isn’t that surprising. The interesting thing will be whether you can offset this with extra battery to get the same range.

yes, Q_TRANSITION_MS only applies when going from VTOL to fixed wing, not the other way around.
There are 4 phases to the VTOL landing.

  • initial positioning
  • final positioning
  • first stage descent
  • final descent
    In the example in your landing at the end of “Quadplane tuning attempts and flight.BIN” the stages took:
  • initial positioning took 14.5 second
  • final positioning took 0.8s
  • first stage descent took 18s
  • final descent took 4s
    The two most obvious ways to reduce the time would be to be a bit lower when it starts the landing. The plane didn’t manage to descend while holding the desired airspeed so ended up well above the target of 15m at the waypoint before landing. You could lower TECS_SPDWEIGHT to prioritse height over speed, or start descending earlier.
    Increasing Q_WP_SPEED_DN will also help, but not by as much as starting at a lower altitude.
    Cheers, Tridge
4

Tridge,
Thank you for your detailed response to my numerous question.

  • The weathervaning effect must have been from the gusty wind that day and the airplane’s natural aerodynamic characteristics. I will increase the Q_A_RAT_YAW params to gain some more control. I think activating the rudder will provide some help on a windy day. Thanks for looking into that.

  • I’m still unclear on how to interpret the tuning error. You say its triggered by the ratio between P to D. The default ratio between these is on the magnitude of about 100 with P at .25 and D at .0035 (can’t remember the actual default). So how could I be getting errors about the ratio being 3.00? This would have to mean I misplaced a decimal, but this isn’t possible using Transmitter Tuning. Sorry if my question highlights my tuning ignorance, I’m still quite confused.

  • Regarding the hesitation to initiate the descent, it is the “initial positioning” I need to work on reducing. I will lower the TECS_SPDWEIGHT, especially now that I have less concern with inadvertently stalling with the quad motors available to save the day.

5

it is the ratio of the output from those terms, not the input. So if there is lots of rotational noise in the gyros then the D terms can start to produce a high level of output. If that gets higher than 1/3 of the P output then it alarms, as you don’t want to “fly on D”.

you’d need to increase the navigation speed in VTOL mode (Q_WP_SPEED). Also adjust the positioning of the final waypoint before the land.

6

Thank you Tridge. I got most of my issues worked out today, then I had a terrible crash that will send me back to the workshop for a very long time. I cannot figure out what happened. I’d really appreciate your insight.

Iam

7

Hello Iskess,

I’m interested in converting and E384 (similar to 3D Aero) into a quadplane and am curious if you have any specific recommendations based on your experience. Any suggestions on specific motors, escs, other hardware and parameters that you found were a good fit.

Thank you so much for any assistance you can provide!